Open Collections

UBC Theses and Dissertations

UBC Theses Logo

UBC Theses and Dissertations

One explanation for why farmers produce cotton collectively in post-Soviet Tajikistan Kassam, Shinan N. 2011

Your browser doesn't seem to have a PDF viewer, please download the PDF to view this item.

Notice for Google Chrome users:
If you are having trouble viewing or searching the PDF with Google Chrome, please download it here instead.

Item Metadata


24-ubc_2011_fall_kassam_shinan.pdf [ 1.78MB ]
JSON: 24-1.0105114.json
JSON-LD: 24-1.0105114-ld.json
RDF/XML (Pretty): 24-1.0105114-rdf.xml
RDF/JSON: 24-1.0105114-rdf.json
Turtle: 24-1.0105114-turtle.txt
N-Triples: 24-1.0105114-rdf-ntriples.txt
Original Record: 24-1.0105114-source.json
Full Text

Full Text

One explanation for why farmers produce cotton collectively in post-Soviet Tajikistan by Shinan N. Kassam  B.Sc.(Agr.), The University of British Columbia, 1989 M.Sc., The University of British Columbia, 1991  A THESIS SUBMITTED IN PARTIAL FULFILMENT OF THE REQUIREMENTS FOR THE DEGREE OF Doctor of Philosophy in The Faculty of Graduate Studies (Resource Management and Environmental Studies)  The University Of British Columbia (Vancouver) July 2011 c Shinan N. Kassam 2011  Abstract In post-Soviet Tajikistan, a heavy concentration on collective cotton production is argued to be based upon vested interests, a cash short economy, historical and political economy reasons. Given challenges in the set up of a functional land cadastral system, access to rural credit can be facilitated through the currency that raw cotton provides as collateral against production loans. While there are other relatively more profitable options for agricultural production, a certain amount of land is placed under cotton, in order to secure financing for non-cotton production. The lender advances credit for cotton production, but a portion of this credit is diverted out of cotton and into non-cotton production. Non-cotton production is not collateralized and results in a private gain for individual members within the “collective”. The lender has full knowledge of this diversion and compensates by pushing a higher level of in kind credit than is needed for the amount of land dedicated to cotton. Accumulating “debt” has become a defining feature of the cotton sector, but I argue that an appropriate definition is non conventional. This is particularly important given that loans are extended over consecutive seasons despite accumulating “debt”. “Debt” can more accurately be defined as the cost of doing business for the lender, who ties together the services of loan provision with that of marketing cotton. Using original copies of farm invoices, state statistics, key informant accounts, as well as secondary survey data, I argue that there is ostensibly little difference in the standard of living between farmers engaged in diversified cropping systems (cotton and non-cotton) and those engaged solely in non-cotton production. In an economy where markets for credit and productive inputs are thin and erratic, the manner in which credit is advanced plays a large role in fostering this indifference. This argument is somewhat different than the prevailing view, which takes the position that cotton “debt” is a constraining factor in the development of the agricultural sector in Tajikistan. One explanation for why Tajik farmers collectively produce cotton at a loss is that it is privately profitable to do so.  ii  Table of Contents Abstract . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .  ii  Table of Contents . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .  iii  List of Tables  . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .  v  List of Figures . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .  vi  Acknowledgements . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . vii Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .  1  1 Relevant Literature . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .  9  2 Background . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 23 2.1  2.2  Stages of cotton financing in Tajikistan . . . . . . . . . . . . . . . . . . . . . . . 28 2.1.1  Stage 1: Sovereign debt agreement . . . . . . . . . . . . . . . . . . . . . . 30  2.1.2  Stage 2: Direct financing . . . . . . . . . . . . . . . . . . . . . . . . . . . 33  2.1.3  Stage 3: Preferential financing . . . . . . . . . . . . . . . . . . . . . . . . 35  Land and Labour markets . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 37 2.2.1  Land . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 37  2.2.2  Labour . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 41  2.3  Productivity on existing cotton operations . . . . . . . . . . . . . . . . . . . . . . 44  2.4  Fertilizer usage and input diversion . . . . . . . . . . . . . . . . . . . . . . . . . . 50  3 Production and profitability . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 55 3.1  Institutions and cotton “debt” . . . . . . . . . . . . . . . . . . . . . . . . . . . . 56 3.1.1  Origin of “debt” . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 57  3.1.2  Cotton contracts and pricing . . . . . . . . . . . . . . . . . . . . . . . . . 58  3.2  Financing of wheat production . . . . . . . . . . . . . . . . . . . . . . . . . . . . 60  3.3  Model . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 63 3.3.1  Farm: Part 1 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 65  iii  3.3.2  Gin: Part 1 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 76  3.3.3  Gin-farm joint profit maximization . . . . . . . . . . . . . . . . . . . . . . 76  3.3.4  Gin: Part 2 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 78  3.3.5  Putting together the pieces . . . . . . . . . . . . . . . . . . . . . . . . . . 79  4 Relative comparison across farm types . . . . . . . . . . . . . . . . . . . . . . . . 87 4.1  Cotton versus mixed production farms . . . . . . . . . . . . . . . . . . . . . . . . 91  4.2  Wheat versus mixed production farms . . . . . . . . . . . . . . . . . . . . . . . . 92  5 Discussion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 101 Bibliography . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 106  iv  List of Tables 1  Export (FOB) statistics for the republic of Tajikistan (1999-2009) . . . . . . . . 26  2  Import statistics for the republic of Tajikistan (1999-2009) . . . . . . . . . . . . . 27  3  National Bank of Tajikistan refinancing rates (%), 1995-2010 . . . . . . . . . . . 32  4  Land under cotton, planned cotton production and actual yields (1998-2007) . . 38  5  Average annual percentage changes in yields of major crops . . . . . . . . . . . . 47  6  Methodica - minimum FOB price calculation for February 22, 2006 . . . . . . . . 59  7  Methodica - minimum ex-works price calculation for February 22, 2006 . . . . . . 59  8  Land under cotton, planned cotton production and actual yields (1998-2007) . . 66  9  Dzerzhinsky farm financial record 2004 . . . . . . . . . . . . . . . . . . . . . . . . 69  10  Dzerzhinsky brigade statistics . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 81  11  National cotton and fertilizer statistics, Tajikistan . . . . . . . . . . . . . . . . . 82  12  Private grain production margins . . . . . . . . . . . . . . . . . . . . . . . . . . . 83  13  Variable names for empirical data set . . . . . . . . . . . . . . . . . . . . . . . . . 88  14  Debt per shareholder among various farm types . . . . . . . . . . . . . . . . . . . 89  15  Retained earnings per shareholder among various farm types . . . . . . . . . . . 89  16  Land per shareholder among various farm types . . . . . . . . . . . . . . . . . . . 90  17  Average cotton yields by farm type . . . . . . . . . . . . . . . . . . . . . . . . . . 91  18  Cotton production per shareholder by farm type . . . . . . . . . . . . . . . . . . 91  19  Imputed cotton prices . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 92  20  Wheat yields between farm types . . . . . . . . . . . . . . . . . . . . . . . . . . . 93  21  Wheat production per farm member . . . . . . . . . . . . . . . . . . . . . . . . . 93  22  Income per shareholder among various farm types  23  Sampled dekhan farm statistics . . . . . . . . . . . . . . . . . . . . . . . . . . . . 96  v  . . . . . . . . . . . . . . . . . 94  List of Figures 1  Cotton credit supply chain . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .  6  2  Map of Tajikistan  3  Average FOB prices for aluminum and cotton fibre exports . . . . . . . . . . . . 28  4  Cotton exports, wheat and flour imports ($000), 1998-2009 . . . . . . . . . . . . 29  5  Real agricultural wages by month, 2000 - 2007 . . . . . . . . . . . . . . . . . . . 42  6  Hectares of land under cotton and grains . . . . . . . . . . . . . . . . . . . . . . . 46  7  Yields per hectare of raw cotton and grains . . . . . . . . . . . . . . . . . . . . . 47  8  Grain yields on independent farms, collectives and dekhan farm associations . . . 49  9  Republican fertilizer use and cotton production . . . . . . . . . . . . . . . . . . . 50  10  Republican fertilizer use v. cotton and grain yields . . . . . . . . . . . . . . . . . 52  11  Dzerzhinsky credit statement 2004 . . . . . . . . . . . . . . . . . . . . . . . . . . 70  12  Dzerzhinsky credit statement 2004 (2) . . . . . . . . . . . . . . . . . . . . . . . . 71  13  A index for cotton fibre (cents/lb) . . . . . . . . . . . . . . . . . . . . . . . . . . 85  . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 24  vi  Acknowledgements Scribing acknowledgements, many years after the initiation of a PhD programme, is challenging. What is not difficult, however, is acknowledging the continued support that I received from my supervisory committee over that period and through the extensions applied for and granted. I have considered myself to be immensely fortunate in having a committee that has stood by me over these years and for the unwavering intensity with which their support was delivered. Les Lavkulich was immensely supportive during the early years of the programme and continued to provide much guidance and mentoring over the course of the programme. Both Ashok Kotwal and Sumeet Gulati were extremely influential in much of the thinking behind this thesis, and very much sharpened a lot of the arguments that are presented herein. I am grateful to both of them as well for accommodating the many Skype calls at seemingly odd hours, given time changes between our various locations. I am particularly indebted to Rick Barichello who, in addition to supervising this thesis, mentored me through my MSc and BSc(Agr) degrees. Rick has always been accommodating and patient in providing critical and valuable input through the many drafts that have been sent his way. I have learned much from him over the many years of association.  My good fortune, however, was not restricted to having a patient, constructive and supportive thesis committee. Richard Pomfret provided valuable comments as the external examiner for this thesis. Both Hugh Neary and James Vercammen, university examiners on the final examination committee, provided comments and input that greatly improved the presentation of the thesis and the arguments contained therein. Hugh Neary strengthened the analysis and discussion of the results, through a helpful set of notes that corrected an error in chapter three. I have much respect and admiration for these examiners, as well as for my thesis committee.  I received invaluable assistance in Tajikistan from Azam Kiyobekov, Vafo Safar, Shodigul Javarieva, Jamshed Lurini, Parviz Shomirzoev, Daler Zurbekov, Ruhafzo Suhrobshoeva, Davlatyor  vii  Jumakhonov, Shermamad Jonmamadov, and Safdar Rajabekov. I am also grateful for helpful discussions, comments and a sharing of levity with both friends and colleagues, and particularly Khaleel Tetlay, Zahur Aman Shah, Dilrabo Jonbekova, Marhabo Jonbekova, Yodgor Faizov, Asanbek Qurbonbekov, Bobo Boboev, Mir Khudobakshov, Orif Zamirov, Rysbek Kurmanov, Soleh Saidmusalamov, Hamid Tursunov, Nicholas Gray, and Dilowarsho Safariev.  I am extremely indebted, as well, to the employees of Dzerzhinsky farm in Leninski rayon of Dushanbe for their time and patience. The information and data obtained from Dzerzhinsky were invaluable in terms of constructing the arguments presented in this thesis.  I thank the University of Central Asia, and particularly the central administration office, for providing me with an office and access to information during my short stay in Bishkek over the summer of 2008. I am particularly grateful to Bohdan Krawchenko, Nasreen Dhanani, Nazir Kassamali, and the staff of UCA, for the courtesies that were extended to me during my stay.  As an employee of Aga Khan Foundation, I acknowledge the flexibility that was accorded to me in order to complete this thesis, while working in both the Kyrgyz Republic and Syria. I reminisce, with fondness, a discussion with Prince Amyn Aga Khan, during my initial employment interview. It may please Prince Amyn to know that his words were strongly influential in my forging ahead and completing the thesis, some five years after this discussion, and despite a number of derailed attempts at moving the research forward. I have learned a lot from the Foundation during the course of my employment. It is my hope that some of the lessons from this thesis will inform the many programmes and projects undertaken by the Aga Khan Development Network in Tajikistan, as well as to inform other institutions and agencies that are active in helping to strengthen the Tajik agricultural sector.  To my fiance, I thank her for her love and support. I apologize for the the odd and long working hours between academic and employment requirements and for all the headaches that she tells me I gave her (I accept...). I hope to make up this loss in the years to come. To my family, and particularly my mother, I owe much for their encouragement and support during the many years that it took to complete this thesis, as well as to my late father, to whom I dedicate this thesis. viii  Introduction The history of Soviet Tajikistan, and Central Asia more generally, can be characterized by a drive for cotton. Little has changed in the post-Soviet era. Russian interests in Central Asia were initially governed by a concern over British plans for the region, coupled with a need to secure a stable source of cotton. For the latter, a need for stability arose from the United States civil war which disrupted cotton shipments to both Russian and British mills. In post-Soviet Tajikistan, a heavy concentration on cotton production is argued to be based upon vested interests, historical political economy reasons, as well as a cash short economy. Access to rural credit is facilitated through the currency that raw cotton provides as collateral against production loans. While there are other relatively more profitable options for agricultural production, a certain amount of land is placed under cotton in order to secure financing for mixed crop farm enterprises. “Debt” is a defining feature of the cotton sector, but i argue that the definition of “debt” is non conventional.  1  Inefficiencies within the sector are perpetuated for historical and  political economy reasons, and given that incentives would seem to be aligned between farmers and those intermediaries who extend credit for agricultural production. This is somewhat different than the prevailing view which takes the position that cotton “debt” is a constraining factor in the development of the agricultural sector in Tajikistan (Stern (2008), World Bank (June 2004), Lerman and Sedik (2009a), Lerman and Sedik (2009b), IMF and World Bank (2009), Bale (2008), International Monetary Fund (2008)). In framing my argument, I utilize original copies of farm invoices, state statistics, key informant accounts, as well as survey data collected by Food and Agriculture Organization (FAO) on 135 farms in Tajikistan. I apply economic principles in the analysis of this data and argue that those collectives characterized by diversified production (cotton and non-cotton production) are where the bulk of “debt” likely exists, and that there is ostensibly little difference in the standard of living between individuals in cotton growing collectives, mixed production collectives (cotton and grains) and non-cotton farm associations. Farmers may, therefore, be indifferent between producing grains privately and producing cotton collectively. Close to 65% of Tajikistan’s population is rural and one out of seven Tajiks, predominantly 1  Throughout this thesis, I place quotation marks around the term debt, in order to distinguish between a conventionally accepted definition of debt, as a financial obligation, and “debt” as part of a credit lending process. I will argue that “debt” is nothing more than the outcome of a credit lending arrangement in which there is an implicit contract between credit lending cotton gins and collective farms, and where the government is a player.  1  young males, is estimated to have emigrated for the purpose of gainful employment. Emigration and remittances, estimated to be anywhere between 20% to 50% of GDP, will continue to be of significant importance as employment and opportunities for reliable private sector engagement continue to be scarce. With only 7% of the land base arable, agriculture contributes 30% to GDP, and provides livelihoods for approximately 60% of the population that remains resident in the mountainous republic. Much of this contribution to GDP stems from the production and sale of raw cotton and cotton fibre. Cotton fibre has, on average, comprised 15% of the total value of exports over the period 2000 to 2007, and is second only to aluminum, which accounts for 60% of the value of total exports. Taken together, remittances, cotton and aluminum are major sources of foreign exchange, and a strong relationship between foreign exchange earnings and factor endowments exists, particularly labour and water resources. Remittances are bolstered by a labour force that is relatively cheap and young; aluminum production is possible given that abundant water resources provide the basis for significant hydropower generation to sustain an energy intensive activity; and previous Soviet investments in irrigation infrastructure, built upon vast water resources, continue to support labour intensive cotton production on a large scale. While cotton is of strategic importance to the republic, for historical reasons, and in terms of foreign exchange acquisition, wheat is of primary importance for farmers within the republic. Cotton and wheat are the two major crops grown within the republic and together, cover approximately 65% of the total cropped area (30% in cotton, 35% in wheat). Wheat is grown by all types of farmers, whereas cotton is predominantly grown on large swaths of land and within institutional structures that mimic the former Soviet kolkhoze. Wheat consumption in Tajikistan has been one of the highest in the world and estimated at close to 200 kg/capita/yr, double the worldwide average of 100 kg/capita/year.2 Meeting this demand has required imports of both wheat and wheat flour and this reliance on imports has been true for much of its history. Therefore, domestic wheat prices in the republic are based upon CIF pricing while cotton is priced on an FOB basis. During the post independence period, the republic has also relied heavily on international aid of wheat and wheat flour, particularly in the years following the brutal civil war that ensued upon independence. As an alternative to cotton, the production of wheat is profitable for private farmers, but thin markets for credit and for material inputs continue to pose a challenge.3 Formal market interest rates for credit range anywhere from 24 to 30%, whereas original copies of farm financial statements reveal that collective cotton producing entities receive credit at a rate of anywhere between 2% and 14%. Working closely with collective type farming systems, the credit lender 2 3  Obtained from Chapter 3 of this thesis provides evidence of this relative profitability in more detail.  2  (typically the cotton gin) funds working capital for cotton production in the form of in kind credit, through the provision of productive inputs and in cash for the purpose of paying monthly wages and picking costs, taxes and cost of utilities. Two important aspects of this system are notable. First, the gins do not engage in credit provision to individual small farmers who do not produce cotton. This is likely due to transaction costs, which are lower when dealing with large farm entities that place a significant amount of land under cotton. Financing cotton only is also reflective of the desire of the gin, which has a territorial monopoly, to capitalize on profits earned from the ginning operation and in potential profit earned through monopsonistic purchase of raw cotton. Secondly, given tight constraints on the availability of working capital, the gin would wish to invest only in those areas where it has the greatest return and security. The legal ability to provide and accept land as collateral within the republic has not yet evolved. In the absence of this facility, it is the forthcoming harvest of the crop for which loans are provided that is the prevailing medium used for collateral. But not all crops are the same, and collateralizing cotton is safer than collateralizing grain. To realize value, cotton must be ginned, and this is a specialized process that requires heavy investment. The need for ginning mitigates the ability of the farm to sell raw cotton into local markets and without the knowledge of the lender. The same cannot be said for grain, which can be processed at the farm level or fed to livestock with limited need for capital investment in equipment. When the lender is interested in output solely for the purpose of loan repayment, he may not be worried about leakage into local markets as long as the loan is extinguished and interest is paid. The gins in Tajikistan are, however, faced with an additional problem and that is a need for a minimum quantity and quality of raw cotton delivery in order to meet contractual obligations with international purveyors of cotton. This requires control over the flow of cotton, as well incentives for farmers to produce raw cotton, through appropriate financing schemes for raw inputs, labour and other production costs. Like credit, local markets for fertilizer and other inputs are thin and erratic. Production of fertilizer in the republic is limited and the bulk of fertilizer is imported, predominantly from neighbouring Uzbekistan. Timing of fertilizer imports is an issue, as Uzbekistan’s industry is highly controlled and regulated. While some fertilizer is leaked into Tajikistan across porous border points and during the sowing season, the majority of fertilizer imported from Uzbekistan occurs after the sowing season. Small independent farmers rely on local markets for fertilizer and many complain of unavailability at the desired time of need or of poor quality and deliberate mislabelling. In pre-purchasing fertilizer, and supplying cotton farms with credit in kind, the gin ensures that its contractual obligations with international purveyors of cotton fibre are not jeopardized at the outset.  3  The literature on trade credit (see for example Burkart and Ellingsen (2004)) would argue that the provision of inputs in kind mitigates the moral hazard problem of borrowers diverting cash away from the intended purpose by ensuring that the transaction is completed in accordance with the purpose of the loan sought. In the case being studied, a moral hazard problem still exists in that the borrowers are able to divert inputs out of intended production and into production of other private goods which are not being held as security on the loan. When loans are extended on the basis of collateralizing the forthcoming crop, diversion of inputs out of cotton production reduces realized productivity relative to predicted productivity. This naturally leads to a paper loss for cotton production, and what is being termed “debt” is argued in this thesis to be (at a minimum), the value of the difference between actual yield and expected yield. The latter is a function of the total inputs provided by the gin for cotton production, while the former is a function of the amount of inputs actually applied in cotton production. Diversion of inputs is, therefore, one key variable in explaining “debt”. Current estimates would suggest that over $550 million of accumulated “debt” has accrued (International Monetary Fund (2008)), but it is unclear what this figure represents, who is required to pay this amount and under what terms will repayment (or forgiveness) be undertaken. While these questions are important, they detract from the reality that the financing mechanism for cotton production has evolved as a response to credit constraints and that this has led to a number of distortions within the sector. When inputs are diverted, the financing of cotton production at rates which are lower than official market rates provides an implicit subsidy for non-cotton crop production. The rationale for this subsidy is that the provision of material inputs, at lower than market rates, acts as a hook for the gin to ensure a reliable supply of raw cotton from the farm; albeit, at volumes which are lower than if the full amount of inputs provided in kind were applied in cotton production. What is being termed “debt” therefore, may not be accurate. Indeed, one could advance an argument that “debt” is nothing more than a level of subsidy, or cost of doing business for the credit lending gin, given its desire to obtain a reliable source of raw cotton. My conjecture that the incentives for the credit providing gin and the farmers are aligned hinges upon four important underpinnings of the current system for cotton production: 1. Access to capital in order to finance production and processing of cotton; 2. Monopsony power in the purchase and processing of raw cotton; 3. A minimum amount of land placed under cotton production; 4. The ability for farmers to divert fungible inputs out of cotton production All have an element of government involvement and all need to exist simultaneously. The republican government directly intervenes in order to facilitate access to capital for the gin, but 4  without guaranteed access to raw cotton, the credit lending gin would find it risky to lend onwards to farmers. With local governmental organs directing raw cotton to gins within their jurisdiction, in order to reap tax revenue, which is linked to the volume of cotton processed, supply of cotton to the gin within its demarcated territory is assured. In addition, when edicts are placed upon minimum amounts of land under cotton production for collective organizations, minimum volumes of raw cotton delivery are assured for the gin. The assurance of supply and volume is further strengthened by a requirement to surrender the forthcoming harvest of cotton to the gin as collateral against the loan. For farmers to willingly participate in the system, private gain from cotton production comes from two sources; (i) the value of non-cotton crops produced with inputs that were earmarked for cotton production and (ii) wage income from the production and picking of raw cotton. The divergence between planned output of cotton, based upon inputs provided, and actual production of cotton, based upon inputs used, is the cost of doing business for the gin. This divergence is the lower repayment rate that the gin is willing to accept in exchange for reaping the benefits of supplying a complete package of input provision, raw cotton purchase, ginning, and sale of cotton fibre. Figure 1 is a general and stylized depiction of the cotton credit supply chain summarizing the above discussion. The figure depicts what I term stage II in the financing regime for cotton production in Tajikistan and which I describe in more detail within chapter 2. I propose and set out to justify the argument that one explanation for why farmers grow cotton collectively in post-Soviet Tajikistan is that it is privately beneficial for each of the individual collective members to do so, even if it is, collectively, a loss making activity. The international purveyor of cotton initially advances a sum of capital (Pa Q1 ) that is determined by an administrative price (set by a domestic regulatory authority) and an agreed upon quantity of cotton fibre to be shipped at a future date. The domestic credit lending intermediary (typically the cotton gin) onward lends this amount, in kind (fertilizer, seed, fuel and operating expenses) as well in cash for the payment of wages. The quantum of inputs and cash is determined by existing norms for input application per unit of land dedicated to the production of cotton. These norms are shown to be higher than technical requirements for current varieties of cotton sown. The “collective” farm, acting on behalf of its members, apportions these productive inputs between cotton (βPa Q1 ) and other crops, typically wheat ((αPa Q1 )). Production of cotton is, therefore, lower than would be expected if the full amount of inputs were utilized in the production of cotton. At harvest, the collective farm delivers an amount of raw cotton (Qc ) to the credit lender, in accordance with the terms of the agreement, where raw cotton was pledged as collateral against the loan advanced. Through the process of ginning, raw cotton is processed into two products (i) fibre, in the amount of (Q2 ) and (ii) joint products in the form of cotton seed, cotton seed oil and cotton meal, all in the amount of (Qj ). Cotton fibre is shipped to the international purveyor (by agreement) while joint products are sold into the  5  Figure 1: Cotton credit supply chain  INTERNATIONAL PURVEYOR PwQ2 - PaQ1 PaQ1 (in cash)  Q2  DOMESTIC INTERMEDIARY (GIN)  INTERNATIONAL ACCOUNT  Qj PaQ1 (in kind)  Qc COLLECTIVE FARM PaQ1  GRAIN PRODUCTION  PRIVATE CONSUMPTION  DOMESTIC SALE OF OIL & MEAL  PaQ1  COTTON PRODUCTION  WAGE INCOME  6  Pa (Q2 – Q1) - C  domestic market for human and livestock consumption (cotton seed and meal). As will be shown in a later chapter, planned output of raw cotton and cotton fibre has been well above realized output since independence. One aspect of this deficiency in meeting the plan relates to the diversion of inputs out of cotton production and into grain production. A second relates to inefficiencies in the conversion between raw cotton and cotton fibre. I provide evidence of the former, while inefficiencies in the ginning sector continue to remain a conjecture due to a lack of reliable firm level data. Notwithstanding these inefficiencies, it must be the case that the gin is able to supply a value of cotton fibre (Pw Q2 ), at the prevailing world price (set on date of shipping), in order to cover the initial amount of working capital paid in advance of the cropping season by the international purveyor (Pa Q1 ). Given lower quantities of shipped cotton fibre (Q2 ) relative to contracted quantity (Q1 ), the world price (Pw ) must, on average over a number of seasons, be higher than the domestically set administrative price (Pa ). Based upon heuristic evidence and key informant accounts, the gain on this transaction is placed into an international account, in the name of the domestic credit lending intermediary. The system, therefore, has a natural and in built mechanism to assist in capital flight.4 In sum, the international purveyor obtains a quantity of cotton fibre that is greater in value than the initial quantum of capital paid in advance of the sowing season. The domestic credit lending agency (gin) benefits from profits that accrue outside of the republic, in the domestic sale of joint products from the ginning process and in profits from the ginning of raw cotton. The cost of ginning, as well as a markup on the provision of capital to the “collective” farm is charged to the collective farm (in the amount of C) and deducted from the value of cotton fibre realized from the ginning of raw cotton shipped. The resulting value of Pa (Q2 − Q1 ) − C paid to the farm is, therefore, negative in most years. The possibility for profitability in cotton production, at the farm level, would be enhanced, were all inputs applied onto land dedicated to cotton production, but there is implicit recognition that a sum of these inputs are diverted into non-cotton production. “Debt”, in this case, is nothing more than a cost of doing business for the gin and remains a paper loss on the account of the collective farm. So why do collective farmers continue to produce cotton collectively? I argue that one explanation is that it is privately profitable to do so and that this profitability has two components; (i) the production of grains and other food crops produced with subsidized inputs and, which are retained for personal collective member consumption, and (ii) wage income, from the 4 Capital flight is argued to exist, in order to evade income tax domestically, as well as due to foreign exchange regulations. Existing rules for transfer of foreign currency out of Tajikistan require a certificate of clearance from the tax office stating that taxes have been paid on the funds being transferred and that these funds were earned legitimately within the country. Most banks have limits on the amount of foreign currency that can be transferred in one single transaction and range from $3000 at smaller banks to $10,000 at larger national banks. Larger amounts require National Bank of Tajikistan approval. While there are creative ways of transferring larger amounts, such as on the names of different individuals, the tax clearance issue comes into play. The value of capital flight can, therefore, be argued to be in evading profit taxes as well as in terms of avoiding “hassles”. The latter is particularly relevant given high levels of corruption.  7  production and picking of cotton. Data analyzed in this study supports the conjecture that the collective sum of these private member benefits is large enough to clear any “debt” at the collective farm level and that for those farms producing both grains and cotton, revenue per member and grain production per member are commensurate with levels realized by private producing grain farmers. The system, while inefficient, would seem to be locked into a low level equilibrium where the interests of various actors within the cotton value chain are aligned. This system persists for historical reasons and for reasons of political economy.  8  Chapter 1  Relevant Literature This thesis aims to answer the question of why farmers in Tajikistan continue to produce cotton collectively, in an environment where farm gate losses in cotton production are accumulating, and where profitable opportunities for producing grain and food crops exist.5 When offered the opportunity to disband from the former Soviet collective, some farmers continue producing in a Soviet style model of collective production, while others choose to farm on privately owned lands. In the case of the latter, these were on lands that were extricated from the dismantled sovkhoze or kolkhoze. Those who remain within collective type production systems accumulate debt and choose to produce both cotton and food crops, while their independent counterparts produce food crops for both household consumption and sale into local markets.6 There are few cases of debt reported by individual independent farmers. Survey evidence (Lerman and Sedik (2009a)) supports a conclusion that cotton cultivation tends to reduce the probability of reaching a higher income in a sample of dekhan farmers when the survey data are analyzed by multivariate regression analysis.7 According to Lerman and Sedik (2009a), adding one hectare to area in cotton reduces the probability of reaching high income (“better than village average”) by 1.8 percent. The authors admit that the result relies on fairly weak data and that the result is not discernible in a simple univariate analysis of family income levels versus cotton growing, when no attempt is made to control for other factors. Nonetheless, the result does fall in line with with relevant literature (Stern (2008), World Bank (June 2004), Lerman and Sedik (2009a), Lerman and Sedik (2009b), IMF and World Bank (2009), Bale (2008), International Monetary Fund (2008)), in terms of notions that the “debt” issue needs to be resolved expeditiously for fear that the sector and the national economy could be irreparably harmed if not attended to with haste. 5 International Monetary Fund (2008) provides the most recent account of the accumulation of what is being termed “debt”. I provide evidence on the profitability of wheat production in chapter 3. 6 I use the term “collective type” throughout this thesis, in order to distinguish between the former Soviet kolkhoze and the new (similar) system of farm associations, which I discuss in an upcoming chapter. The essential difference between the kolkhoze and sovkhoze was the manner in which benefits were disbursed. Within the kolkhoze, all output above the planned target was shared by the members of the kolkhoze who could choose to consume the output or sell it within the local market. Within the sovkhoze, all output was surrendered to the state and a wage provided to all members. In both cases, small plots of land were also provided to members for home production. 7 A dekhan farm is an independent peasant farm as defined by the Law on Dekhan Farms (1992) but has since taken on varying definitions including “family”, “individual” and “collective” dekhan farms. In the case of the latter, these are best described as partnerships of individual or family dekhan farms.  9  There is a prevailing argument that farmers are being forced to grow cotton (see for example Stern (2008)), but on its own, coercion cannot justify the continued cropping of cotton, if farmers are being asked to honour what is being termed “debt”. If coercion is being practiced to its fullest extent, there is no need to appeal to “debt” and an obligation to honour “debt”. It must be the case that other incentives are at play, which appeal to cotton producing farmers, in order to continue the production of cotton without resistance. In uncovering the nature of these incentives, a first look at the history of incentives within Soviet style collective production is helpful. A comment made in (Domar 1966) provides an interesting cue for a review of the literature that is relevant to the subject: Imagine that most of-the obstacles facing Soviet kolkhozes (collective farms) today, such as output and delivery quotas, administrative interference, shortage of strategic inputs (materials, spare parts, fertilizer), depressed prices of outputs, etc., suddenly vanish, and the kolkhozes find themselves in a Lange-Lerner type of a competitive world where everything can be bought and sold at a market price, and where peasants are free to run their own affairs provided the essential structure of the kolkhoz is retained. How would Soviet agriculture, or for that matter any economic sector so organized, fare in such a wonderland? Freed from existing restrictions and abuses, the kolkhoz would presumably revert to its prototype- a producer cooperative which utilizes the labor of its members, purchases other inputs, sells its outputs, pays a rent and/or taxes, and divides all or a part of its net proceeds among its members. The presumed democratic nature of such a co-op and its freedom from capitalist exploitation has made it highly attractive to socialists and social reformers for ages. But its popularity has not prompted its proponents to analyze it with the same loving curiosity that the ”bourgeois” economists have shown toward the capitalist firm. And yet it must have been obvious, at least to some of these proponents, that co-op members are likely to be ordinary human beings bent on maximizing the benefits from their participation in the co-op. A number of criticisms were launched over Domar’s paper8 but given the chain of events since the dissolution of the Soviet Union and the restructuring of the agricultural sector in former soviet republics, his thoughts are perhaps more applicable today than when originally written. Yet, despite the intellectual curiosity raised by the article, there is scant literature available in the public domain to address the issues raised by Domar. Modeling the soviet collective as a producer cooperative, Domar (1966) reworks the model Yugoslavian coop as constructed in (Ward 1958) with a generalized production function and confirmed, in general, the tenor of Ward’s findings. When attached to a supply schedule of labour, a characteristic not present 8  Specifically Robinson (1967) and Oi and Clayton (1968)  10  in the Yugoslav coop model, the paradoxical result obtained in Ward (1958)  9  was reversed.  Subsequent treatment of Ward and Domar’s analysis of producer cooperatives and cooperative enterprises are provided in Sen (1966), Oi and Clayton (1968), Vanek (1970), Bradley (1971), and Cameron (1973). Of particular interest to the topic of this thesis is the analysis provided by Bradley (1971) and the critique provided by Cameron (1973). Unlike the other aforementioned authors, Bradley took exception to the characterization of the soviet kolkhoze as a producer cooperative, or a labour managed firm, in which all production and allocation decisions were made collectively and in which the common goal was maximizing income per worker. In Bradley’s characterization, managerial incentives on the soviet collective were tied to the level of agricultural deliveries to the state and the central plan. These incentives, claimed Bradley, were at odds with the household, or individual on the collective, who maximized utility rather than average income per man day or deliveries of collectively produced crops to the state. A further critique was launched upon previous characterizations such as in Oi and Clayton (1968) that the labour input per member in the soviet collective was invariant to changes in the wage and therefore inelastic. If labour input per member was constant, then the collective must be limited to changes in membership as a device for changing the labour input in collective production, a characteristic of labour managed firms. Bradley correctly suggested that this was an awkward method in which to affect short run adjustments to any policy change or change in planned production and set forth to analyse the short run supply of labour on soviet style collectives in order to highlight the implications of soviet agricultural policies and reforms for member incentives and agricultural performance. While all of the aforementioned works have relevance to this thesis, the critique of Bradley’s article and rejoinder by Cameron (1973) uncovers an important facet of the wage structure in the soviet kolkhoze. In particular, any labour allocation that is made within the growing season and which is paid from the residual income, after delivery of the state mandated plan, there is little certainty of this wage prior to harvest. In short, anything that affects net income has a direct impact upon the residual wage provided to the individual collective member. In a residual wage system, the rewards that accrue from collective labour are therefore contingent upon the quality and the quantity of collective labour inputs provided by all members. This, as Bradley suggests, is a major source of uncertainty to the collective membership leading to the possibility of shirking and disincentive. The soviet sovkhoze, in which members earned a wage and no residual income, was one response for dealing with the moral hazard problem. The manner in which inputs are provided to the collective and the disposition of inputs which enhance private 9 In Ward’s analysis, the means of production are nationalized and the factories turned over to a general committee of workers who are free to set price and output policy in their own material self interest. The objective is to maximize income per worker and the decision variable is the size of membership. The firm, it turns out, has a negatively sloped supply schedule and reacts to a price change by altering the rate of output in the opposite direction.  11  gain is a more modern solution to deal with the moral hazard problem in present day Tajikistan. In following the analysis of Bradley (1971) and in considering the debate contained within Bradley (1971) and Cameron (1973), Bonin (1977) analyzes a collective farm household allocating time between two distinct labour pursuits, work on both the collective and labour plot, and leisure. Contrary to the results of Bradley (1971) and Cameron (1973), Bonin shows that the Stalinist policy of increasing taxes against the collective as well as crop quotas, generate positive work incentives in both private and collective plots and at the expense of leisure. Moreover, a result that the collective household’s response to price changes for both collective output and private output is ambiguous may suggest a basis for the Soviet penchant to choose quantity type control instruments over price controls in order to meet state plans. While in the past, central plans were almost always met on paper10 , they were often unrealistic and untenable. In present day Tajikistan, the covering up of low production levels is not practiced and state statistical agencies officially report that between 1998 and 2008, the state plan for cotton was largely unfilled. One aspect of this current situation is that, while the national plan for cotton is set on the basis of volume, the implementation of the plan and the directives in the field are based upon number of hectares devoted to cotton production. The conversion between planned volume and devoted hectares is based upon an overestimated norm for yield and with this knowledge, underachievement of the plan is predictable. Incomplete privatization and land reform have reopened some of the above discussions, given the observation that a not so insignificant number of farmers remain within a collective type production system in the post-Soviet era. In this thesis, I investigate one reason for why members may be indifferent between farming collectively and farming independently. The choice is influenced by a lack of effective markets and marketing channels for both (micro) credit and farm production. I specifically argue that the production of a farm product which requires further processing provides an avenue for a lender to tie together the provision of in kind credit with monopolistic services of a processor and marketing agent. How is this unique? In rural environments where credit markets are weak, production of basic food crops uncertain (vagaries of weather, government policy, etc.) and where employment in non farm activities are limited, the production of farm products which require processing offer the potential for obtaining credit that can be diverted to the production of food crops or in the acquisition of cash which can be used to support household basic needs. When the credit lender ties together the provision of credit with the requirement to surrender final product of the processable product, he is essentially guaranteeing the return of capital for the inputs that he provides, the margin 10 Most commentators on the Soviet period would suggest that plans were often not met and that ingenious schemes were developed to meet audit requirements which often led to double counting of production. In the case of cotton, there are accounts of how cotton was moved across farms on the Tajik and Uzbek borders, which enabled each side to provide evidence of meeting the plan with relative ease.  12  obtained on the supply of these inputs, as well as any profit earned in the processing operation and sale into the market of the processed product. The moral hazard problem of diverting the final product directly into the market by the farmer is limited by the need for processing prior to sale. For raw cotton, this processing through a ginning operating is expensive and prohibitive for small farmers, even when they are collectively organized. But, if this was the end of the story, there would be nothing unique to the process. The farmer would produce only if there was a profit and the lender would only provide credit if he obtained product that was of a quality and a quantity that remunerated his services as a credit lender, a processor and a marketing agent. In the post-Soviet era, the persistence of collectivist forms of agricultural production, continues to remain a puzzle, and all former Soviet republics would seem to have been faced with this problem in varying degrees. Amelina (2000) reports that the number of loss making collective enterprises in Russia increased from 5% of the total in 1994 to 82% in 1997, while the number of collective enterprises (in various forms) had remained relatively stable. In 1997, collective enterprises in Russia still produced 50% of the agricultural output and operated upon 80% of the agricultural land. A number of possible reasons have been offered for the persistence of collective and cooperative agricultural systems: 1. The notion of a “safety umbrella” and of joint action in which the exposure to risk under collective or cooperative production is lowered relative to private farming. (Lerman and Feder (2004)), (Machnes and Schnytzer (1993)) In addition, social capital within the group may serve as a mechanism for sharing risks with others and as a means for obtaining cheaper or more secure access to the provision of services. (Ruben and Lerman (2005)) 2. Where there is a perceived lack of expertise and skill in independent private farming, pensioners and non agricultural workers or those with little interest in tilling the land may prefer to leave their land in the collective rather than leasing their land to struggling individual farmers. Opting to leave the land in the collective is argued to provide greater security for future streams of payment. (Lerman and Feder (2004)) 3. The Sandinista regime in Nicaragua, as an example, favoured the creation of agricultural production cooperatives as an alternative means for poor and landless households to obtain access to land, credit and extension services as well as a device to maintain rural stability. (Ruben and Lerman (2005)) Many of these agricultural production cooperatives are based upon familial ties (Carter and Luz (1993)) and production on collective fields are usually reserved for cash crop production while subsidiary plots are maintained for home consumption  13  4. Former collective farm managers have developed significant respect as local community leaders and figureheads. In a quest to maintain power, the fight for retention of the collective structure is a reasonable proposition. Social norms, as well as political, social and economic networks may also provide enough “muscle” to maintain the collective farming system when former collective farm managers privately gain from the persistence of the collective farm. (Lerman and Feder (2004)) 5. The involvement of regional authorities and local governments, particularly where agriculture plays an important role in the regional economy may foster the continued existence of the collective. Where local regional budgets are dependent upon taxes or other forms of payments from collective structures, the process of decollectivization may be slow to proceed. (Lerman and Feder (2004)) 6. The preference of foreign agencies and international organizations to finance collective activities and the neglect of domestic agencies to provide support for private farming activities in remote areas reinforces a degree of coordination and collective action in order to capture the benefits of services and programmes offered by international agencies. (Lerman and Feder (2004)) All of the preceding are reasonable conjectures, but ones that can be dealt with through appropriate policy measures. The accumulation of collective debt on cotton farms and through monopsonistic practices, coupled with input diversion and private gain is somewhat more difficult to tackle, particularly when the incentive structures of all stakeholders would seem to be aligned. Cotton is strategic for the republic, in large part because of the foreign exchange that it garners, while the financing of cotton provides access to materials, inputs and cash for farmers to produce non cotton crops. Yet, some farmers have broken away to produce noncotton crops, and seemingly obtain production credit from primarily informal credit markets, but also from formal (micro)credit institutions. Are independent farmers any different in terms of accessibility to credit or are there other distinguishing aspects of this subset of farmers that prompted them to seek independence from a collective production entity? In present day Tajikistan, and after cessation of the civil war, the movement toward individual (private) farming has been arduously slow. Many of the former Tajik collective farms have been reorganized, but most commentators would agree that there is little difference in the structure of operations between the current form and the previous soviet kolkhoze. Decollectivization and “land reform” has, for the most part, resulted in the distribution of paper shares as opposed to land title certificates.11 In a number of cases, farmers were unaware or unclear as to which strip of land they have inherited from the decollectivization process. Those 11  The nature of the decollectivization varies by region. Generally, however, workers are allocated land on the basis of number of years of service on the collective or state farm. Stature and position also plays a role as farm directors received larger allocations of land.  14  wishing to exit the collective and to remove their land share were faced with the real risk of obtaining marginal land. In the early years of independence and shortly after the end of the civil war (1992-1997), many of the kolkhoze and sovkhoze structures were also indebted to the government and were behind on payments to the state for material purchases and taxes. In some cases, farmers wishing to break off from the kolkhoze or sovkhoze were asked to pay for prorated debt that had accrued on the farm based upon the land holding that had been assigned to them on the basis of the decollectivization scheme. These amounts were relatively low, but in the absence of any formal credit facilities for small farmers at the end of the civil war, many farmers were weary of a move toward independent farming without access to reliable sources of financing working capital. Those that moved early were predominantly ones who had access to capital from non formal credit institutions and largely from family members working outside of Tajikistan and sending back remittances.12 In addition, this subset of farmers was also likely from previous kolkhoze and sovkhoze farms which had high labour to land ratios and therefore, the amount of land extricated, and consequently the prorated amount of debt, was relatively low. In other cases, the choice to leave or to stay, was also predicated upon the structure of the farm household and particularly upon the gender balance of the household. A large proportion of those who emigrated (approximately 1 out of 7 Tajiks) were largely young males. Female headed households were likely not in a position to take on the challenge of independent farming and could find support in collective production processes. Again, no official data exists to confirm this notion, but while driving through many of the rural areas in Tajikistan, one is struck by the enormous number of females working on cotton fields, while private farm holdings are typically tended to by males. Discussions with government and industry officials suggest that those collective or state farms in which irrigation systems were still functional, and where machinery was still in relatively good state, were prime targets for local government organs (hukumats) to exercise their mandates in fulfilling their district quotas of the national plan for cotton production. Given that cotton is a water intensive crop and that heavy machinery is required for efficient land preparation and sowing, the hukumats may have been willing to let poorly endowed collectives follow the prescribed course of land reform. In the case of farms where hukumats have made it difficult for the land reform process to evolve, “freedom to farm” (Lerman and Sedik (2009a)) has been severely hampered by state interference in intra land use planning, either by unofficial coercion or through legislative acts. This does not mean that individual farmers could not extricate their land from the kolkhoze or sovkhoze, but the point being made is that in the well endowed farms, the process may have involved more bureaucratic hurdles and wrangling. Lerman and Sedik (2009a) conclude that while land reform had significant positive achievements in Tajikistan, 12  There is no data to support this contention and this perspective is largely based upon conversations with farmers and industry personnel and government officials.  15  much less had been accomplished relative to other CIS countries that had more robust land reform policies. Among cotton growing farms, only 14 percent of respondents indicate that the individual farmers have freedom in their decision of whether to plant cotton and how much, 56 percent of the farmers interviewed indicated that it was the farm director who decided upon the cropping mix and allocation, while 28 percent responded that it was the local government (hukumat) who directly intervened in the planting decisions.13 There are no decrees on the stipulation of minimum land to be planted in cotton, but a lot of anecdotal evidence and opinions expressed that there are demands by the local hukumats to plant 70% of land under cotton. It is further argued that these edicts serve the purpose of attempting to meet the indicative cotton production plans set by the national government and to maximize tax revenue for local governments which are stimulated by the volume of cotton produced within their district. Yet, enforcement of crop mix plans by command have a clear impact upon efficiency and optimality of input use on farms and would seem to be counterproductive in meeting the goals of the hukumat. Indeed, for most of the last decade, the cotton plan has not been met in a majority of the districts and for 8 of the 10 years between 1998 and 2008, the plan was not met nationally. Edicts on land under collective production are not new in post-Soviet republics and follow similar patterns during the Soviet era. Khrushchev is reported to have instituted a plan to put pressure on collective farms during the 1950’s for reducing the size of private plot areas and it is argued that one intention of this policy was to induce an increase in the amount of labour devoted to collective crop production and out of private crop production (Ireland and Law (1980)). Studies on the Soviet collective farm have generally been upon an analysis of labour input and modeled upon the general theory of the producer cooperative. The tie in to the literature in terms of labour input on ante bellum cotton production is also relevant in this regard as the economic essence of slavery involved the ability of the owner to control the allocation of labour time between market and non market activity (Wright (1978)). None of the studies surveyed, however, have looked at the allocation of material inputs for collective production and private production. In other words, while the allocation of time between marketable and consumable goods is important, this study is concerned with the allocation of material inputs between a collectively marketable crop (cotton) and privately consumable crops (eg. grain) which creates a tension between optimality, as viewed from the collective farm level, and optimality as viewed by the individual collective member. In an environment where regional governments and other agents have a stake in maintaining the collective farm structure in order to foster cotton production, a first mover advantage in leaving the collective may not exist. Given a conscious choice to work within a cooperative 13  USAID and World Bank (2007) as cited and discussed in Lerman and Sedik (2009a)  16  or collective system, and with land use rights vested to the collective, the current system possesses elements that are common within the wide and varied literature on land tenancy and contractual structures. The literature on land tenancy and contractual structure in agriculture begins with the observation of an individual (landless) tenant and an individual landlord who voluntarily agree upon a right of tenancy. In the case at hand, we are interested in a group of farmers, who have the right to farm a piece of land, but not the ability to sell or encumber the land14 , and who are willing to vest that right to a farm manager who coordinates the planting scheme and the marketing of the agricultural commodity. Within the existing literature on agricultural tenancy, the puzzle has been one of providing a convincing explanation for why differing contractual forms exist, both spatially and temporally. The focus has largely been upon the optimal provision of inputs (land, labour, skill, fertilizer, seed, et cetera) as well upon issues of effort and shirking. In providing pieces of the solution to the puzzle, the conjectures made within the existing and vast body of literature include (i) non functioning or highly imperfect markets for farm inputs (Reid (1976), Bliss and Stern (1982), Pant (1983), Jaynes (1982)), (ii) moral hazard and the screening of employees (Hallagan (1978), Newberry and Stiglitz (1979)), (iii) risk sharing and the tradeoff between risk and higher transaction costs (Cheung (1969)) (iv) the contribution of unmarketed resources by both landlord and tenant in which both agents share an incentive to self monitor (Eswaran and Kotwal (1985)). An appreciation of interlinkages or interlocking factor markets is clearly present within most of the studies related to agrarian development. The observation of interlinkages in rural markets is now well documented and this thesis adds further examples of interlinkages in factor markets to the rich literature on this topic. In a predominant number of cases, the observation has been the interlinkage between rural land and rural credit markets. An explanation for the observance of both “high” and “low” rural interest rates has been offered through the notion of interlinkage and differing attitudes toward risk (Mitra (1983), Braverman and Stiglitz (1982)). Basu (1983) for example, has also maintained that “potential risk” in rural credit markets generates an inherent tendency for them to get interlocked with other markets and that there are natural reasons for the emergence of isolation and interlinkage in less developed economies. One serious drawback according to Gangopadhyay and Sengupta (1986) is the assumption that interlinkage is assumed to begin with rather than being formed as a utility maximizing decision of the tenant and that loans can be acquired for both production purposes and consumption purposes with each type of loan exhibiting a 14 On the basis of the constitution, all land in Tajikistan remains in the exclusive ownership of the state and landholders are only accorded land use rights. The inability to collateralize land (or land use rights) is one constraint in the effective functioning of credit markets in the republic. This is a conscious choice that the republic has made, but the reasons for this decision are not documented.  17  unique interest rate which may be higher or lower than the market rate. In the case studied within this thesis, the assumption of interlinkage at the outset is reasonable given (i) governmental interference in cropping patterns, (ii) weak or thin credit markets and (iii) that the lender fixes the amount of the loan on the basis of standards or norms applicable to a specific amount of input use per unit of land put under cotton. Data analyzed within this study provides evidence that loans advanced for cotton are well above any reasonable norms for input use per unit of land, and particularly in relation to fertilizer use in cotton production.15 The farmer, in this case, has little flexibility in negotiating the amount of loan needed, but given that in kind loans (fertilizer, cash, fuel, etc.) are advanced in an amount greater than needed for the production of cotton, there would seem to be implicit knowledge on the part of the credit lending gin that a portion of the inputs are diverted to non-cotton production. The combination of “excessive” loans advanced for cotton production, along with monopsony power in the purchase of raw cotton has seemingly been the root cause of non-profitability in the production of cotton at the farm gate.16 (Lerman and Sedik 2009b) claim that the accumulation of “debt” in the Tajik cotton sector can be traceable to pervasive government intervention in financing and production.17 When coupled with a poor enabling environment for restructuring of Soviet farming systems, improvements in profitability and efficiency have been impeded. The authors state: The debt crisis in Tajikistan’s agriculture has been caused by a combination of two factors typical of such situations in many countries: (a) the inability of the farms to make a profit under current conditions and (b) continued lending by the banks to cotton producers regardless of reduced payment capacity and lack of credit-worthiness. Their contention is that the farm debt issue, as present in post-Soviet Tajikistan, is not strictly a transition economy phenomenon and that similar situations can occur within market economies - Israel as an case study within their paper.18 The common feature in their discussion was that the problem encompassed a whole economic-social sector in each of the countries studied. Their recommendation was that standard debt resolution measures involving insolvency and traditional approaches would lead to an unacceptable social cost for the rural populations as whole. There is little that one can contest within the arguments made by Lerman and Sedik (2009b). 15  Fertilizer is, for instance, provided as in kind credit to the farms and data collected for this study indicates that it is provided at levels which are at least 40% higher than standard industry norms. 16 I will provide evidence in chapter 3 to support a conjecture that private gain from the diversion of these inputs into private production more than offsets this loss in cotton production. 17 I discuss in detail, the various means by which the government has intervened in the financing of cotton in the next chapter. 18 The authors did not include a third reason, which could be that the gins are creating a dependency on the part of the cotton producers which can be exploited, much as in the case of the post bellum US cotton farmers which I discuss shortly.  18  Indeed their arguments echo what many international and multilateral organizations are stating and particularly the World Bank, the Asian Development Bank and the European Commission. There have also been calls for a moratorium on “debt” within the cotton sector and most discussion on the cotton “debt” are couched in the language of a “crisis”. Bolton and Rosenthal (2002) investigate the question of whether there are net benefits to having political institutions that permit ex post intervention in private debt contracts within a democracy and as a correction mechanism for incomplete contracts. Their paper argues that the simple threat of moratoria enhances efficiency. As the likelihood of bad economic shocks increases, state contingent debt moratoria always improves efficiency ex post and may also improve ex ante efficiency. Moreover, if the likelihood of macro shocks are not strong, state contingent moratoria has the benefit of improving ex ante welfare by completing incomplete debt contracts. In varying degrees and interspersed within the document, this thesis delves into the question of state contingent moratoria and the possible benefits or dangers that such moratoria may have in a post-Soviet economy such as Tajikistan. The topic is somewhat relevant, given the various calls by multilateral organizations for the Government of Tajikistan to resolve the cotton “debt” issue and various announcements by the government to forgive a portion of the current outstanding debt estimated in 2008 at $520 million (International Monetary Fund (2008)). In surveying the literature on cotton debt, one is struck by the similarities in the problems faced by small cotton farmers in the post bellum American south and the cotton farmers in Tajikistan. Ransom and Sutch (1972), Ransom and Sutch (1975), Wright (1978), Reid (1976), as well as a host of other historical books have documented how the southern farmer’s dependence upon external financing had made him vulnerable to exploitation by the same granters of credit. In the postbellum south, the granter of credit was the local country store, who prevented farmers from producing surplus’ of food crops so as to maximize his volume of business, as a supplier of credit for cotton, a seller of foodstuff, and a purchaser for the cotton. There is debate over the role of the country store and whether tenants were coerced into “too much” cotton and whether usury and the mechanics of the lending arrangement were bonding tenants into a system of debt peonage. One argument is that, given that the country store tied together the requirement to purchase household consumption items (tea, sugar, tobacco, foodstuff) which had low price elasticities of demand with the extension of credit, store owners would want to maximize tenant income. If cotton was the most profitable crop and if most of these consumption items could not be grown on many farms in the area, then the country store strived to have the most profit maximizing crop mix planted in order to extract profits by charging monopoly prices on the goods that the tenants purchased (Alston (1990)). While the local merchant is not the supplier of credit to farmers in Tajikistan, the similarity lies in the manner in which credit is advanced. In both cases, the creditor, seeking security for the loan places a lien on the forthcoming crop. In the postbellum south many of the farmers  19  were tenant farmers who did not have the ability to mortgage their lands. Tajikistan farmers are also unable to mortgage lands and the legal implications of collateralizing land use certificates are unclear. By insisting on the forthcoming crop, the creditor essentially limits the options available to the farmer or the collective with respect to the sale of the cotton. Production options become limited over time and the farmers are in effect “locked in” to the production of cotton. But, there is a question about how long this system can be maintained. The possibility of moratoria and political intervention would certainly be one explanation for persistence on the part of the indebted farmers but it would not explain why the lenders continue the process of lending in an environment of accumulating debt. The literature on corporate finance would argue that optimizing individuals or firms would not repeatedly lend in an environment where there is continuous renegotiation unless a mechanism to protect the lender is at work and that there exist some optimal debt contracts which would persuade a debtor to pay out cash flows rather than divert them (Hart and Moore (1998), Aghion and Bolton (1992)). Both Hart and Moore (1998) and Aghion and Bolton (1992) show how control shifts from the debtor to the creditor in certain states of the world. When the debtor has residual control rights, actions will be taken to increase private benefits but at the expense of the lender(investor). In contrast when the lender(investor) has control, actions are taken that do not respect the debtors private interests. A state contingent optimal allocation arises where the debtor has residual control rights in those states where his private benefits are relatively high and the lender should have control in those states where the debtors private benefits are relatively low. The fundamental difference between Hart and Moore (1998) and Aghion and Bolton (1992) is that in the former, the shift in control is endogenous and occurs because the debtor fails to make a promised repayment as opposed to to an assumption that control shifts are triggered by a verifiable state of the world. Further insight is gained from Akerlof and Romer (1993) and in an analysis that depicts how bankruptcy for profit can occur when firms have an incentive to go broke and to profit at the expense of society (loot) rather than to go for broke (to gamble on success). Bankruptcy for profit, according to Akerlof and Romer (1993) occurs most commonly when a government guarantees a firms debt obligations and their paper sets out to analyze a number of important financial crisis including the S&L crisis. The analysis and models contained in both Akerlof and Romer (1993) as well as Bolton and Rosenthal (2002) are found to be particularly relevant to this thesis. An equally relevant area in the literature that ties in to theory of bankruptcy for profit is the topic of trade credit and input diversion.19 Burkart and Ellingsen (2004) claim that it is typically less profitable for an opportunistic borrower to divert inputs than to divert cash and that illiquidity of inputs is what facilitates trade credit. Diversion is interpreted within their 19  The manner in which credit is advanced to the Tajik collective cannot strictly be classified as trade credit wherein a stipulated grace period is provided for repayment of the goods provided. There are, however, a number of lessons and insights offered to the study at hand from this relevant literature.  20  study as “any use of resources which does not maximize the lenders” “expected return”. Their claim is staked upon the assumption that an input supplier has a monitoring advantage over banks in that the supplier automatically knows that an input transaction has been completed, whereas a bank or other lenders incur monitoring costs in ascertaining whether inputs have been purchased with the cash advanced on loan. An implicit assumption is that the borrower will not liquidate the supply of inputs through sale as it is not in the interest of the borrower to do so given the need to discount the value within the market. Where the receivables of the borrower are not encumbered as part of the bundled exchange of goods for credit, the assumption is valid. In contrast to Burkart and Ellingsen (2004), diversion in this study is defined as the employment of resources out of pledged production of a particular commodity and into an alternative occupation which cannot be directly monitored by the lender. The assumption is that diversion lowers the lenders “full” return and increases the borrowers current assets. 20  “Full”, in this case is a state where the borrower produces only that product for which the  loan is extended and does not practice any diversion of inputs. Under such a scenario, there may be little if any monitoring advantage to the lender of inputs on trade credit over a lender in cash. The only advantage gained is in the ability of the lender in trade credit to adjust his prices accordingly, in order to anticipate diversion as defined in this study. When the final production is collateralized and recovered by the lender, he may also not be overly concerned with the monitoring of use of inputs as the incentive for the farmer in obtaining a subsequent loan is contingent upon providing “reasonable” and fair production through which the credit lender is effectively remunerated for his services. The system has naturally developed a set of prices and processes that keeps all parties happy. The fundamental pillars within this system comprise of: 1. Access to capital in order to finance production and processing of cotton; 2. Monopsony power in the purchase and processing of raw cotton; 3. A minimum amount of land placed under cotton production; 4. Collateral of fixed assets is not required; 5. The ability for farmers to divert fungible inputs out of cotton production and these are explained in more detail within the following chapters. I begin first with an overview of the Tajik economy and a description of land, labour and input markets. I then turn to the important role that foreign exchange and domestic credit constraints have played in the evolution of crop mix patterns within the republic and particularly the continued importance of cotton. An analysis of incentive structures on the collective farm in post-Soviet Tajikistan, 20  Current assets need not be in the form of fixed capital but may be consumed within the period accumulated or in a later period.  21  and particularly the incentive structures for growing both cotton and wheat is subsequently undertaken. This is followed by a comparative analysis of returns between cotton growing entities and private entities that produce wheat and food crops. The role of local and national governments in maintaining this system, and all five of the pillars listed above are detailed throughout. Specific emphasis is given to how the government has fostered the ability of the gins to ensure a steady and reliable access to raw cotton while allowing collective farms to maintain a level of living standards that are relatively comparable to private farmers. These analyses will support the contention that collective farmers are indifferent between working on collective cotton farms and working on their own private lands where a variety of grains and food crops are cultivated. With much to learn from the literature on the post bellum south, the conjecture fits well within the literature on corporate finance, trade credit, and particularly within the literature on bankruptcy for profit.  22  Chapter 2  Background In this chapter, I provide the background and history on input markets within Tajikistan’s agricultural sector. My specific focus is upon credit, land and labour markets. In the case of credit, I outline the history of the various stages in which the government has intervened, and with respect to ensuring that credit is provided for the production and processing of raw cotton. I continue with an argument that the regulations and the characteristics of land and labour markets are intimately tied to the manner in which credit is extended and to the incentive structures that exist within collective farm households within the republic. I also undertake an analysis of the constraints related to productivity enhancement on existing cotton operations and initiate a first look at the issue of fertilizer use and the diversion of fertilizer from cotton production to grain production. Tajikistan has been portrayed as the poorest of the Central Asian republics at the dissolution of the Soviet Union. Bordered by Afghanistan to the south, China on the east, the Kyrgyz Republic to the North and Uzbekistan to the west, the landlocked republic boasts some of the most impressive and highest mountain ranges including the Trans-Alay range in the north and the Pamirs in the southeast. Close to 65% of its resident population, estimated at 7 million, is rural. With elevations ranging from a low of 300 metres to a peak of 7495 metres, 93% of its land area is mountainous, and of the 7% that is arable, cotton dominates the agricultural landscape. The country’s economy relies heavily upon the export of cotton fibre and aluminum, and together, they have comprised close to 75% of the total value of exports since 2001. Light industry is limited and those that do function are in need of significant capital investment. A large aluminum plant built during the Soviet Union continues to function in the Turzunzoda region. With a smeltering capacity of over 500,000 tonnes per year, the plant is one of the largest in the world and is purported to consume 40% of the country’s total electricity production. In addition to providing significant export revenue to the republic, the plant also employs a relatively large number of employees, in an economy where employment is limited, and where emigration is prominent.21 21 In 2003, 14,000 employees (10% of the industrial labour force) were hired by the aluminum plant (United Nations Economic and Social Commision for Asia and the Pacific (2003)). Very little has changed since 2003 in terms of employment numbers and given the importance of the plant for local employment, restructuring of this state owned enterprise has been delayed due to fear of social unrest in the event that a significant amount of labour is shed within the local community in which the plant is located.  23  Figure 2: Map of Tajikistan  Source:  One in seven Tajiks is estimated to be residing in other former Soviet republics for the purpose of gainful employment and in order to provide support to families in Tajikistan through remittances. A recent relaxation of rules relating to taxation of remittances has permitted transfer through official bank channels, but a large amount of remittances still flow into the republic informally. Estimates suggest that these remittances may range from 12% to 50% of GDP. Given the large foreign exchange earnings from cotton and aluminum exports, as well as remittances, (unofficial) dollarization has been on the rise in the republic. A large percentage of bank funding is skewed towards the dollar, and concerns have been raised about the vulnerability of banks, given a possibility for exchange rate induced credit risk.22 Like other republics of the former Soviet Union, Tajikstan faced a number of economic and political problems at dissolution. However, in addition to the challenge of economic transition, 22  Loan dollarization increased from 57% in 2004 to 68% in 2007 (International Monetary Fund (2008)). Much of this lending, in dollar terms, has been for construction and housing, but cotton loans are also priced in dollars and then disbursed in local currency given strict restrictions on foreign exchange movement within the country.  24  the republic was also faced with the challenge of a civil war and transformation of a divided republic into a stable post conflict republic. The vestiges of the civil war that ensued upon independence in 1992 and ended in 1997 are still visible. More than 50,000 are claimed to have lost their lives and over 600,000 persons displaced during the war which was fought along geographic and ethnic lines. A large number of ethnic Russians also fled during the civil war, along with many professionals. Infrastructure was badly damaged during the conflict and funds for reconstruction and maintenance were scarce. Where donor funds were available, the perception among the layman is that widespread corruption diverted funds from their intended purpose. In view of the civil war, there was a significant delay in attending to pressing matters related to the economic transition, and most linkages (both forward and backward) had been completely severed. The effects of this severing were felt most acutely within the agricultural sector. By the end of the war in 1997, Tajikistan’s agricultural sector was still dominated by a Soviet legacy of central planning in the production of cotton, but without the support of the centre to finance production. Factor markets for labour and capital did not exist and, in addition to other important sectors, the government did not possess the liquidity to finance production of cotton within the then existing kolkhoze and sovkhoze enterprises. Given limited options for earning foreign exchange, the government sought options for financing the production of cotton, a historical crop within the republic, and one that took on more strategic options after independence and the civil war. I discuss the various attempts made at financing cotton production shortly, but turn first to the strategic importance of cotton in post-Soviet Tajikistan and particularly in the acquisition of foreign exchange. Export statistics for major categories are shown in table 1 while table 2 shows major categories of imports for the same period. Together, cotton and aluminum had, for some time, averaged approximately 75% of the export revenue for the republic and both sectors are under consistently heavy criticism from international organizations and other stakeholders for environmental degradation and poor labour standards. Table 1 indicates that cotton fibre exports have, in terms of percentage, contributed less to total export earnings, over the past four years. Given that both aluminum and electricity exports have been relatively constant in terms of volume, the relative percentage decrease is entirely due to lower world prices for cotton and a spike in aluminum prices in 2006, as depicted in figure 3.23 From a strategic perspective, lower international prices for cotton fibre have also been problematic for the republic in terms of addressing the tradeoffs between cotton production and 23 Prices were obtained by dividing total export revenue by total export quantity. The prices are, therefore, annual averages. While cotton prices fluctuate by month, the cycle of production and ginning are such that most cotton is shipped at the same time each year and therefore the prices are relatively comparable on the basis of annual averages.  25  Table 1: Export (FOB) statistics for the republic of Tajikistan (1999-2009)  26  Exports ($000) Aluminum Cotton Fibre Electricity Other  1999 688686 308097 91185 174680 114724  2000 838927 433555 91829 181601 131942  2001 651571 398397 71500 78548 103125  2002 697079 398555 127971 67414 103138  2003 797899 429554 192742 55141 120462  2004 914924 563354 161597 57580 132392  2005 908717 563016 143912 52555 149234  2006 1399023 1049510 128667 49015 171831  2007 1468170 1082983 137845 59619 187724  2008 1406350 1012992 108218 59748 225392  2009 1009968 589463 99683 63188 257634  Aluminum (%) Cotton Fibre (%)  45 13  52 11  61 11  57 18  54 24  62 18  62 16  75 9  74 9  72 8  58 10  224282 92235 3831  229181 85889 3973  253730 85701 4037  253585 84082 3831  273541 78796 3909  275893 89913 3927  278443 90249 3849  279971 86590 3985  286994 75383 4047  289711 84639 4066  295812 101334 3966  Exports: Aluminum (tons) Cotton fibre (tons) Electricity (million Kwh)  Source: National Bank of Tajikistan website (  Table 2: Import statistics for the republic of Tajikistan (1999-2009)  27  Imports ($000) Alumina/aluminum oxide Natural Gas Oil and Petroleum Electricity Wheat Flour Other Wheat and Flour (%) Imports: Alumina/aluminum oxide Natural Gas(million m3) Oil and petroleum (tonnes) Electricity(million kw hours) Wheat (tonnes) Flour (tonnes)  1999 663136 128675 36023 51738 179094 32775 13043 221788  2000 859445 198542 65268 68009 249375 36323 8342 233587  2001 687529 185030 26668 73218 98319 30371 7449 266474  2002 687492 178009 22358 69575 82219 23332 12361 299638  2003 881304 235813 24306 73009 61539 12416 18772 455448  2004 1375228 284790 27909 106834 65033 15041 33343 842278  2005 1329799 361823 26600 126129 58137 30690 45434 680986  2006 1722614 386670 34927 191367 66834 28026 48953 965838  2007 2455444 376583 64469 275292 65715 40121 95015 1538249  2008 3269803 377709 74348 412510 87518 58864 148632 2110222  2009 2568730 290056 52003 322633 76119 73566 100996 1653356  6.91  5.20  5.50  5.19  3.54  3.52  5.72  4.47  5.50  6.35  6.80  446956 750 304800 3641 278346 82200  545980 729 202234 5242 321278 55406  537095 572 276054 5396 264232 54697  601373 486 278380 4659 291623 123958  682251 532 305066 4618 144008 189126  614489 616 356636 4807 125472 215950  798691 629 308069 4506 288577 346334  839796 635 379235 4840 276337 376082  836856 645 518807 4361 282733 511287  817127 513 539410 5146 254662 520404  643429 217 635453 4276 403757 368146  Source: National Bank of Tajikistan website (  Figure 3: Average FOB prices for aluminum and cotton fibre exports 4000  3500  3000  2500  2000 Aluminum ($/ton) Cotton Fibre ($/ton) 1500  1000  500  0 1998  1999  2000  2001  2002  2003  2004  2005  2006  2007  2008  2009  Year  Source: Constructed from data in table 1 wheat imports.24 Wheat, flour and bread products have always been of primary importance to households in Tajikistan. In the first quarter of 2008, 35% of all household expenditures on food was on bread and bread products and the consumption of these bread and bread products exceeded medical norms by a reported 17% (State Statistical Committee, Government of Tajikistan (2008)).  2.1 Stages of cotton financing in Tajikistan This section provides an account of the three stages that the republic has undergone in order to finance the production of cotton. Admittedly, the discussion is descriptive and with significant detail. The literature on cotton credit in Tajikistan is often erratic or written at a 24 While world prices for cotton have been erratic over the period covered in figure 3, I will show in chapter 3 that average prices for cotton in Tajikistan have been falling relative to world prices since 2004. A number of reasons explain this difference in trends and include issues related to grading standards, as well as the institutional structure for registering cotton exports within the republic. In the case of the latter, there are loopholes within the current structure that provide the ability to enhance both capital flight and tax evasion through the undervaluing of cotton exports.  28  specific point in time. In contrast, this section provides a detailed and historical account of the various systems that have been used to finance cotton production in post-Soviet Tajikistan, and the prominent role of the government in the finance of cotton. The overt preoccupation with financing cotton production has at least two potentially detrimental effects. First is the importance of cotton as a main export crop and a source of foreign exchange, which leads to political pressure and interference from local governmental organs to enforce national edicts on the production of cotton. This clearly has an impact upon efficiency within the agricultural sector and influences “optimal” crop mixes. Secondly, the pressure to finance cotton, in an environment where default rates are high25 , potentially induces instability within the banking and financial system, and leads to lower available credit for all sectors, not just for cotton. Figure 4 depicts the value of cotton exports against wheat and flour imports and the trade balance between these products. Figure 4: Cotton exports, wheat and flour imports ($000), 1998-2009 250000  200000  150000  100000 Cotton Fibre exports ($000) Wheat Imports ($000)  50000  Flour Imports ($000) Residual ($000)  0 1998  1999  2000  2001  2002  2003  2004  2005  2006  2007  2008  2009  ‐50000  ‐100000 100000  ‐150000  Source: Constructed from data in table 2 25  This claim that default rates are high is deduced from the various reports and claims that the sector is heavily indebted. While this study looks at “debt” in a different light, formal lending agencies may be accepting “debt” and the incidence of debt in a more traditional sense.  29  In constructing the residual curve in figure 4, there is no intention to suggest that cotton exports directly and solely finance the import of wheat. Given the fungibility of export revenue, any commodity or export item can be used to finance the import of wheat. The intention for constructing the residual curve was a a noticeable link between cotton financing schemes and the divergence between cotton export revenue and wheat import expenses. Three stages are noticeable within this graph and each one of these corresponds neatly to a strategy for financing cotton production within the republic. Stage one from 1996 to 2003 corresponds to a period where cotton financing was on the basis of a sovereign agreement between international purchasers and the Tajik government and where the trade balance between cotton fibre exports and wheat imports was relatively stable. Stage two (2003-2007) moved to a financing regime in which private contracts between Tajik intermediaries and international purveyors of cotton fibre provided much needed capital to domestic producers without a sovereign debt agreement, but where the national bank of Tajikistan played a significant role in financing cotton production. During this period, the trade balance between cotton fibre exports and wheat and flour imports was decreasing and a transition in financing structures was initiated at a point where there was no trade balance in these imports and exports. Stage three, the current stage (starting from 2007) is one where the government of Tajikistan itself is financing a portion of domestic production, through the Ministry of Finance, while private agreements between intermediaries and foreign purveyors of cotton continue but within a newly regulated structure aimed at preventing capital flight. With the residual curve showing an upward trend at the initiation of this current stage, it would seem too coincidental if the divergence between cotton export revenue and wheat import expenses did not influence the decision for each of these policy changes related to cotton financing. Each of these three stages is detailed below.  2.1.1 Stage 1: Sovereign debt agreement At the end of the civil war, and in a move ostensibly aimed at backstopping the production of agricultural commodities, the Tajik government entered into a sovereign debt agreement with the Swiss cotton purveyor Paul Reinhart. The loan was channeled through the state owned Agroinvest bank in Tajikistan for onward lending to agricultural enterprises which, in 1996, were still in the form of kolkhozes and sovkhozes. The effective interest rate on the sovereign loan was 10% per annum and the National bank of Tajikistan channeled the loan to Agroinvest bank at a markup of 12%.26 The combined cost of the loan (22%) was significantly lower than the national bank’s internal refinancing rate to commercial banks which, in 1996, averaged 138% per annum as shown in table 3. 26  The loan was undertaken by Reinhardt, through Credit Suisse First Boston (CSFB), and guaranteed by the sovereign debt agreement. The 10% interest on the loan accrued to CSFB and thus this initial transaction cannot strictly be considered as an advance by Reinhardt. The risk of fluctuating world prices were, therefore, to be borne by the government of Tajikistan and ultimately the farmers to whom the loan was channeled. As will be discussed shortly, future rounds of financing have, however, been more closer to advances rather than loans.  30  Historical accounts differ at this point, but what is clear is that Agroinvest bank entertained proposals from a number of intermediaries to disburse loans to agricultural enterprises which would reduce the transaction costs for the bank and which would presumably reduce the risk of default. In the case of the latter, the intermediaries focussed their attention entirely upon the production of cotton, and without the benefit of being able to collateralize on land, accepted the forthcoming harvest of cotton as collateral. Three points deserve particular mention here. The first is that the restructuring of state farms and a process of land privatization had not yet been initiated and, therefore, there was no legal basis for the collateralization of land, as all land belonged to the state. The second, was that there was no legal basis for the collateralization of the harvest and therefore no basis for ensuring that the farmers would relinquish all of their yields at the end of the cycle, particularly in a year when prices were low. Given that the kolkhoze and sovkhoze structures were still under state control, the ability to collateralize the forthcoming crop was, therefore, contingent upon a partnership between the local governmental structures (hukumats) and the directors of the kolkhozes and sovkhozes. Thirdly, in view of the inability to collateralize land and given the state structures for agricultural production, the intermediaries were reluctant and unwilling to advance any loans for the production of grains or vegetables. Quite apart from the risks inherent in the collateral of perishable produce, and thin local markets, the privatization of industrial enterprises had been initiated and provided an added incentive for the sole financing of cotton production. A number of strategic gins had been acquired by the intermediaries in the process of industrial privatization.27 The intermediaries now controlled not only the financing of cotton production, but also the critically important area of processing raw cotton into its primary components of (i) fibre and (ii) cotton seed. In the absence of processing, raw cotton has limited value. To the intermediary28 , the value of the collateral (in this case raw cotton) was only convertible when the raw cotton was processed into its marketable products. The gin was a critical component in the conversion of the collateral. To the intermediary, the gin also provided another important cog in the mechanism for controlling the cotton sector. One of the primary inputs in cotton production is seed, a joint product 27  Agroinvest bank may not have been inclined to ensure that the loans taken out by the intermediaries were actually disbursed as loans to the farmers. If so, the question of what collateral Agroinvest bank took from the intermediaries is of interest and one for which no immediate answer is forthcoming or even contemplated in public discussions. 28 In this section, I use the terms intermediary and gin interchangeably given that the intermediary is also the owner of the gin. The nature of gin ownership is not clear and there are varying arguments. The most common is that the gins were commandeered during the privatization process by government officials and other influential persons aligned to the government. There is little evidence to corroborate these claims, but little to also refute the claims. The gins continue to remain secretive operations, limit visitors to their plants, and are unwilling to engage in any interviews. World Bank (May 2007) claim that the government owns a 25% stake in some ginneries.  31  Table 3: National Bank of Tajikistan refinancing rates (%), 1995-2010 1995  32  January February March April May June July August September October November December per quarter 1 2 3 4 per annum  100.00 100.00 120.00 250.00 250.00 250.00 250.00  1996 250.00 250.00 180.00 180.00 180.00 142.00 120.00 72.00 72.00 72.00 72.00 72.00  1997 72.00 72.00 72.00 72.00 72.00 72.00 72.00 72.00 72.00 72.00 72.00 75.77  1998 78.10 75.36 71.10 44.33 36.94 36.83 36.13 34.37 33.20 38.10 41.00 38.77  1999 29.92 20.55 22.44 17.70 17.93 14.04 16.43 14.39 11.77 10.92 16.48 14.53  2000 16.16 20.00 19.91 20.20 21.92 21.92 21.60 21.95 20.74 20.32 20.60 20.60  2001 20.60 20.60 20.60 20.60 20.29 20.00 20.00 20.00 20.00 20.00 20.00 20.00  2002 20.00 20.00 20.00 20.00 20.00 20.00 20.00 20.00 20.00 20.00 20.00 20.00  2003 20.00 20.00 20.00 20.00 20.00 20.00 20.00 17.75 16.77 16.63 11.43 9.51  2004 15.00 11.92 11.40 11.17 10.00 10.00 10.00 10.00 10.00 10.00 10.00 10.00  2005 10.00 9.46 9.00 9.00 9.00 9.00 9.00 9.00 9.00 9.00 9.00 9.00  2006 9.00 8.72 8.50 8.50 8.50 8.50 8.50 8.68 9.00 9.00 9.93 11.29  2007 12.19 13.00 13.00 13.00 13.00 13.00 13.00 13.00 13.00 13.00 13.00 13.19  2008 15.00 15.76 16.00 14.83 14.75 14.75 14.07 14.00 14.00 14.00 13.79 13.50  2009 13.31 12.00 12.00 11.07 10.00 10.00 8.71 8.00 8.00 8.00 8.00 8.00  100.00 155.65 250.00 188.41  226.15 167.47 88.17 72.00 138.13  72.00 72.00 72.00 73.27 72.32  74.83 39.34 34.58 39.27 46.88  24.42 16.57 14.22 13.03 17.02  18.66 21.35 21.44 20.50 20.49  20.60 20.30 20.00 20.00 20.22  20.00 20.00 20.00 20.00 20.00  20.00 20.00 18.19 12.54 17.66  12.79 10.38 10.00 10.00 10.79  9.49 9.00 9.00 9.00 9.12  8.74 8.50 8.72 10.08 9.01  12.72 13.00 13.00 13.07 12.95  15.58 14.78 14.02 13.76 14.54  12.45 10.35 8.24 8.00 9.74  Source: National Bank of Tajikistan website (  2010 8.00 8.00  obtained during the ginning process. With ownership of the ginning process, the intermediaries could control the financing of cotton production, have control over the stock of seed29 , and equally important, they could control the stocks and sale of cotton fibre in order to maximize returns in a sector characterized by fluctuating world prices. Left to third party investors, the gins and the ginning process could have potentially undermined the intermediary and the ability to control the sector, particularly if the third party owners of the gin would provide credit to the farms at any point in the future. The amount paid for the gins in the process of privatization is not a matter of public record and therefore there is much speculation as to where the capital was raised for the purchase of the gins. Some have (quietly) argued that a portion of the initial capital of $138 million dollars provided under the sovereign debt agreement may have been used by the intermediaries to finance the purchase of the gins, thereby limiting the amount of capital that was actually disbursed to the farms for the production of cotton. In early 2000, Agroinvest engaged in a debt-equity swap with foreign purveyors of cotton (and intermediaries), where a portion of the sovereign debt was swapped with minor equity assets in the bank. While this was of strategic importance to the national government, the continual accumulation of non performing domestic loans continued to pose a problem for the republic. By 2003, it was clear that the institution was a going concern. In addition to rent seeking activities on the part of the intermediaries (gins), falling world cotton prices in the 1990’s, crop failures in 1999 and 2001, as well as weak fiscal management, the majority of loans extended for cotton production were in default. With pressure from multilateral organizations, the government split Agroinvest bank into two separate entities in 2003. An asset management company (Kreditinvest) and a commercial bank (Agroinvest). At the end of 2004, all liabilities (largely non performing cotton loans) were transferred to Kreditinvest, along with a private capital injection of $300,000. International Monetary Fund (2008) opine that the private injection came from politically well connected individuals as well as the the intermediaries and their international creditors. No further discussion is had on the motivation for investing in Kreditinvest, but it seems reasonable to argue that the intermediaries and the individual investors had much to gain in keeping Kreditinvest solvent, either through (i) continued and sustained benefits from engaging in the cotton trade, (ii) from delaying the payment of previous loans that had been advanced for cotton production or (iii) both.  2.1.2 Stage 2: Direct financing The splitting up of Agroinvest bank signalled the end of what I had previously characterized as stage 1 of the historical financing regime for cotton production in the republic and the beginning of stage 2. With no sovereign agreement to bring in further capital to finance production of 29  Importation of cotton seed was, and still is, highly regulated and restrictive in Tajikistan  33  cotton, the intermediaries needed to arrange for direct agreements with the international purveyors of cotton for working capital. To this end, Kreditinvest took its role as the institution that channeled these funds between the foreign purchaser/financiers and the local intermediaries. The process was made easier, given that the intermediaries were also shareholders of Kreditinvest, through the previous debt-equity swap as well as the private capital injection. An agreement for purchase and sale of cotton fibre between the intermediary and foreign purveyor of cotton was conducted on the basis of prepayment, which was sought to lend onwards to farmers in order to provide working capital for the production of raw cotton. The prepayment was subsequently provided by the intermediaries to Kreditinvest, on a loan basis, and then subsequently loaned back to the intermediary by Kreditinvest, in country. Why? On the surface, the two loans would seem to be a wash? Given the large amount of foreign currency that has been flowing into the country, particularly through informal channels for remittances, the government has always enforced a regulation that foreign currencies may not be used for settlement within country and exchange of foreign currency between personal or commercial domestic accounts is not permitted. While the legislation was aimed at tackling inflation, the law implicitly aided the intermediaries in potential avenues for capital flight. The loan from Kreditinvest to the intermediary, in country, can be denominated, either in the form of foreign currency or Tajik somoni. In the case of the former, the loan may be denominated in dollars or other foreign currency, but the funds are disbursed in Tajik somoni, at the prevailing exchange rate, and if the loan is to be utilized within country. For the intermediary, the advantage of having the loan denominated in local currency increases in an environment where there is continual devaluation. To see this more clearly, consider that average exchange rates for US dollars against Tajik somoni have been 2.2 (2004), 2.55 (2005), 3.08 (2006), 3.44 (2007), 3.46 (2008) and approximately 4.5 during the first quarter of 2010. Supposing that the intermediary, after obtaining prepayment, provided Kreditinvest with a $100,000 loan in 2004. Excluding interest, for the sake of exposition, the loan to the intermediary (denominated in somoni) would have been 220,000 somoni which, if called in 2005 would have been worth $86,280. The intermediary could, therefore, erase the cost of the in country loan and still retain cash within the institution, which could be called upon in a subsequent period. Differential interest charges would, of course, eat away at some of this gain but the example serves the purpose of illustrating how the intermediary could finance the production of raw cotton without using any of his private capital. Were the in country loan to be denominated in dollars, the intermediary would have received 220,000 somoni in 2004 but would have to repay 255,000 somoni in 2005 to extinguish his $100,000 loan from Kreditinvest in country. The system ensured that the intermediary gained in an environment where the local currency was losing value and at the expense of the domestic financial institution. The cost of working capital to the intermediary was, therefore, much lower, relative to the option of raising funds for working capital domestically.  34  So long as the intermediary can honour his agreement to ship cotton fibre to his international purveyor, as per the terms of his agreement, and in an amount valued at greater than the amount of prepayment, the intermediary lives to trade for another season. If the market favours the intermediary and cotton fibre prices provide a profit to the intermediary, he has a choice to retain his foreign currency earnings outside of the republic or to invest it back in Tajikistan. Most traders opted for the former and continue to do so even today, given what are perceived to be financial, investment and other risks within the republic and the wider region in general.30 In addition, given devaluation of local currency against the US dollar, the intermediary continues to have a stake in Kreditinvest, as an institution that facilitates his business in country, and which allows him to take his profit out of the republic, without any hassle or physical handling of currency. This threat of capital flight ostensibly prompted the republican government to institute measures to prevent such a system. In addition, given lower market prices for both cotton and aluminum (from 2006), foreign exchange earnings were dwindling, while the pressure of import expenditures was increasing (c.f. figure 4). A measure aimed at licensing all cotton fibre exports in order to prevent capital flight, as well as enforcement of export taxes on cotton were instituted in 2006 and took the form of a central clearing agency. A similar regime was also placed for the export of aluminum wire and both commodities were required to register contracts at the Tajikistan Universal Goods Exchange (TUGE) or more well known as the Birzha31 . A more formal description of the workings of the Birzha are given in the next chapter, the institutionalization of which I define as the beginning of stage 3 in the cotton financing scheme within the republic.  2.1.3 Stage 3: Preferential financing Stage 3 comprises a new area in financing for cotton. With the cessation of lending by Kreditinvest in 2008, due to insolvency issues as well as through recommendations from an external audit, commercial banks were being requested/persuaded to borrow from the Ministry of Finance and to on lend to the cotton sector in 2008.32 (IMF and World Bank (2009)) claim that the amount called upon for onward lending to the cotton sector was reported to be 140 million Tajik somoni (approximately $31 million), which represented approximately 22% of bank 30  Many of these conjectures were confirmed on the basis of a telephone conversation with a Tajik cotton trader and financier who resides in Moscow but wished to remain anonymous. 31 A conversation with a Tajik cotton trader based in Moscow, and who asked to remain anonymous, stated that the contracts registered at the Birzha are not binding and that the function of the Birzha is limited to the stipulation of “administrative” prices on which the export tax is calculated. The “administrative” price has little correlation with actual price received at the time of export and as discussed in the next chapter. 32 One area uncovered through the audit was that the National Bank of Tajikistan had provided guarantees and pledges to foreign lenders and somoni denominated loans to Kreditinvest. International Monetary Fund (2008) claim that these guarantees and pledges were equivalent to approximately 85% of the national bank’s gross foreign assets at the end of 2007, thereby putting the solvency of the national bank in great jeopardy.  35  capital as of June 2008. With close to 44% of loan repayments for cotton production in default at the end of 2006, onward lending in this amount is clearly an area of heightened risk for the banking system. These three stages are indicative of the prominence that the government has placed upon the financing of cotton production since the end of the civil war. The preoccupation with financing cotton production, at the expense of financing other crops or non agricultural sectors has at least two potentially detrimental effects. First is the importance of cotton as a main export crop and a source of foreign exchange, which leads to political pressure and interference from local governmental organs to enforce certain edicts on the production of cotton. This clearly has an impact upon efficiency within the agricultural sector. Secondly, the pressure to finance cotton, in an environment where default rates are high, induces instability within the banking and financial system, and leads to lower available credit for all sectors, not just for cotton. I will argue that the debts which are currently being laid at the feet of the farmers are not debt, but rather the cost of doing business for the intermediaries (gins). If accepted, there is no debt crisis within the cotton sector of Tajikistan, but rather a need to revisit policies around the financing of cotton production and the functioning of the cotton value chain. I will further argue that within the cotton value chain, all agents, including farmers, are benefiting from the production and marketing of cotton. While the distribution of benefits may not be fair, a lack of opportunities for employment and income generation keep a certain number of farmers in the field and in the pursuit of collective cotton production. An important part of this discussion on credit also relates to available credit for non-cotton producing (private) farmers as well as individuals within the republic. Access to banks continues to be low by international standards and in 2006, there were about 220 bank branches, translating to about one branch for 30,000 people on average (International Monetary Fund (2008)). In view of mountainous terrain and blocked passage for at least 4 months of the year, access to banks continues to be a critical issue within the republic. Given the importance of remittances, access to banks and official money service agencies also remains a critical link in the flow of cash and credit to areas outside of the capital and main urban areas. Where branches are accessible, limited credit services and products prevent banks and credit agencies from turning remittance deposits into stable sources of financing and revolving funds. Where credit products are available, they may not be suitable or practical to meet the needs of small scale farmers. Even in the case of microfinance products, where collateral could be on the basis of a group loan guarantee, a mismatch in loan cycles to production cycles may inhibit the effectiveness of the loan product being advanced. Most notably, however, is the requirement of a 20% required reserve ratio which, when coupled with significant financing to the cotton sector and poor performing loans, limits the amount of available credit within the economy.  36  I will argue that the production of cotton, therefore, provides an avenue for collective farmers to access credit, a portion of which can be diverted out of cotton production and into the production of privately produced goods. This access to credit from cotton production is one explanation for why farmers continue to produce land on collectively farm land, and a first look at land and labour markets is therefore important prior to advancing this argument.  2.2 Land and Labour markets 2.2.1 Land Given the large numbers of rural residents and mass emigration, one reason for the lack of development in a formal labour market likely relates to the process of land reform and decollectivization of the kolkhoze and sovkhoze structures. Despite the promulgation of a number of government decrees and ordinances, the process of decollectivization and privatization of agricultural land has been arduously slow. In most cases, there has been a lack of consensus at the farm level on how to apportion capital assets, agricultural land and debt. Where success has been had in terms of consensus on apportionment, there is often little difference between the former collective system and the newly created organizational structure. Only in a limited number of cases has there been a clear move toward independent family farming on individual parcels of land. For the most part, farmers have continued to band together in arrangements where (it would seem) former collective directors have assumed the position of manager in a variety of legal associations including joint stock companies (open and closed) and dekhan(peasant) farm associations.33 Collaboration with local governments (hukumats) is one key mechanism for ensuring the collateralization of the cotton harvest as well as a steady and stable supply of cotton for the ginnery. While national and local governments continue to make public statements regarding a hands off approach on the planning of cotton production, most observers are in strong agreement over the deliberate manner in which cotton plans are constructed and implemented. Indeed, the strongest indictment are official publications of cotton plans and realization of the cotton plan by oblast (province) and by rayon (district). Table 4 details the plans by regions for the period 1998-2007.34 Actual production of cotton 33  This is based upon the authors observation in the districts around Dushanbe as well as the southern areas of the republic and is a common perception among many industry analysts. The notion that management has been hijacked by former collective directors is, however, disputed in a report by The Centre for International Studies and Cooperation (November 2006) in their analysis on cotton production enterprises in the northern Sughd oblast (province). An open joint stock company can accept new members/investors whereas a closed joint stock company cannot increase its membership base. A dekhan farm association is simply a collective or cooperative of individual dekhan farmers. 34 In addition, the hukumat’s tax base is also comprised significantly of tax revenues obtained from the production of raw cotton. Transparency is not a virtue of this system and thus little information is available on the construction of the tax on cotton production. What is known is that there is a per tonne tax on raw output,  37  Table 4: Land under cotton, planned cotton production and actual yields (1998-2007)  Republic  Sughd  Khatlon  RRS  1998 1999 2000 2001 2002 2003 2004 2005 2006 2007 1998 1999 2000 2001 2002 2003 2004 2005 2006 2007 1998 1999 2000 2001 2002 2003 2004 2005 2006 2007 1998 1999 2000 2001 2002 2003 2004 2005 2006 2007  Land (ha) 247347 247800 238608 257371 269197 284317 293516 288655 262893 254830 74593 72064 69088 75267 80834 85508 88176 86392 80474 75867 151129 154992 150287 160505 165367 173527 178761 177036 161222 161230 21625 20744 19233 21599 22996 25282 26489 25227 21197 17733  Plan (tons) 600000 600000 600000 600000 600000 610000 610000 610000 547000 550000 187200 187200 187200 187200 187200 190000 190000 190000 171300 171300 360000 360000 360000 360000 360000 365000 365000 365000 329000 330500 52800 52800 52800 52800 52800 55000 55000 55000 46700 48200  Yield (plan) (tons/ha) 2.4 2.4 2.5 2.3 2.2 2.1 2.1 2.1 2.1 2.2 2.5 2.6 2.7 2.5 2.3 2.2 2.2 2.2 2.1 2.3 2.4 2.3 2.4 2.2 2.2 2.1 2.0 2.1 2.0 2.0 2.4 2.5 2.7 2.4 2.3 2.2 2.1 2.2 2.2 2.7  Realized output (tons) 383654 313125 335427 452735 515478 537358 556766 447918 437898 419786 128657 115107 120958 145347 154871 148466 159424 155526 131402 124830 217717 171658 177950 255092 305754 334063 334079 240884 258073 259643 37280 26360 36519 52296 54853 54829 63157 51508 48423 35313  Realized yield (tons/ha) 1.6 1.3 1.4 1.8 1.9 1.9 1.9 1.6 1.7 1.6 1.7 1.6 1.8 1.9 1.9 1.7 1.8 1.8 1.6 1.6 1.44 1.11 1.18 1.59 1.85 1.93 1.87 1.36 1.60 1.61 1.72 1.27 1.90 2.42 2.39 2.17 2.38 2.04 2.28 1.99  plan (%) 63.9 52.2 55.9 75.5 85.9 88.1 91.3 73.4 80.1 76.3 68.73 61.49 64.61 77.64 82.73 78.14 83.91 81.86 76.71 72.87 60.48 47.68 49.43 70.86 84.93 91.52 91.53 66.00 78.44 78.56 70.61 49.92 69.16 99.05 103.89 99.69 114.83 93.65 103.69 73.26  Source: State Statistical Committee, Government of Tajikistan (2004), (2007), (2008)  38  has not met planned production for the majority of the period 1998 to 2007 and this can be attributed to at least three reasons. Firstly, the plans may have been overly optimistic, a possibility that gains significant acceptance on the basis of historical evidence of Soviet planning in the agricultural sector wherein taut plans were often met through padding of invoices and receipts as well as other creative schemes. That the plans have been adjusted downward recently may signal acceptance of this tautness. Secondly, the vagaries of weather, pests and natural disasters, are not uncommon in the region and significantly affect crop output. But these are random occurrences. The inability to fill state plans is more regular than random. Thirdly, if the plan was reasonable, then diversion of inputs out of cotton production may help to explain the poor performance of cotton production relative to plan. One component of this thesis is an assessment of the plausibility of the latter. Some argue that poor quality inputs and the late delivery of inputs are responsible for low yields and the inability to fill the plan as well as a part of the debt problem on cotton farms. But, if late delivery of inputs occurs continuously from year to year, this cannot be used as a reasonable explanation for poor performance against plan given that the hukumats would be intimately aware of this limitation and could adjust their plans accordingly. This adjustment, for example, could involve an edict for greater amounts of land to be devoted to cotton production or more vigilant oversight on cotton production in their area of jurisdiction. That there has been a stable requirement to place 70% of land under cotton production may signal that this is the correct ratio under which constituent farmers on cotton producing entities can meet their personal and household demands for food and personal living expenses. This can be done either through production of grain and food crops on the remaining 30% of the land or through the leasing out of this 30% of land to private farmers on sharecropping agreements. All land is owned by the state and the sale of land is prohibited in the republic. Leasing of land is, however, acceptable and there are functioning land lease arrangements within the republic. Most of these agreements are conducted between private farmers, and the new collective, on the 30% of land that is not devoted to cotton production. Those individuals who engage in lease agreements are typically young males, who did not inherit any land in the collective, and who have not emigrated. The agreements are, in large part, share cropping agreements where the collective provides the necessary inputs and land to the private farmer in exchange for a portion of the crop harvest (grain or vegetables). These lease agreements are predominantly conducted on those collectives that have low labour inputs or where the labour is largely female.35 In the case of the latter, these are collectives placed within districts where emigration of males is high. The demographics of labour within collective structures is not documented and there is little collected and remitted by the gin, and a land tax which is significantly lower for land planted in cotton than in other crops. The lower land tax on cotton is undoubtedly an aim at indirectly influencing the production of cotton. 35 The emigration patterns are such that males typically leave during the cotton planting and growing season and return to Tajikistan during the cotton picking season in late fall.  39  information available on the characteristics of villages where emigration is high or low. A number of district officials had stated during field interviews that the government is cautious about a full scale decollectivization process on the basis of an argument that there would a large number of rural residents sitting idle and that this was a recipe for civil unrest. Cotton is still the dominant crop within the republic, is labour intensive in production, and wages paid for employment in the agricultural sector are greatly influenced by the choices made in the production of cotton. This thesis argues that the structure of the labour force, lack of alternative employment options and poor functioning credit markets are important factors for sustaining the current system of cotton production. The only viable alternatives to working within the collective (or farm associations) are either emigration and a search for better paying employment or independent family farming, both of which are interlinked. Smaller households may opt to remain within a collective for reasons provided in the section on relevant literature and include, among other reasons, the inability to farm privately due to household structure and demographics, risk or the desire not to accept any prorated debt that has accumulated on the collective farm. For larger households, the choice to farm privately may result in surplus labour within the household over and above the requirements to till privately held land. Given that one out every seven Tajiks are reported to have emigrated for employment purposes would suggest that emigration may also be a response to small parcels of privately owned land and surplus labour. There is little discussion on the possibility that a more complete decollectivization process and a stronger move toward independent farming may trigger another round of emigration if the newly apportioned parcels of land are below a certain level to accommodate the existing work force. To understand this more clearly, consider that eligibility for a share of the collectives assets and land requires continued membership and work within the collective. Each collective is, therefore, endowed with its own team of workers, who are potential beneficiaries of land and assets. The movement of workers between collectives is thus mitigated as members would risk losing their share of the assets. The efficiencies gained from a functioning labour market (both allocative and technical) are restricted in this case. Given that the production of cotton is labour intensive, a move toward food crop production is likely to release a significant amount of labour off the land. The precise quantum of labour shed will, of course, depend upon the relative marginal productivity between cotton and food crop production, an area that i discuss in the following chapter. The impact upon real wages in a decentralized and independent agricultural production sector will be of particular importance and more so given the current characterization of agricultural wages as “slave” wages.36 36  see for example, Stern (2008).  40  2.2.2 Labour Emigration to former soviet republics and Moscow is extensive and more than 1 million Tajiks are estimated to be living abroad. The National Bank of Tajikistan estimates that remittances account for 12% of the country’s GDP while others have estimated a range between 20% to 50% of GDP.37 The emigrants can be largely characterized as male dominated blue collar workers engaged in the construction industry or heavy labour based industries. Jones, Black and Skeldon (2007) distinguish Tajik labour emigration on the basis of three characteristic waves:  • Wave 1: The exodus of the educated and the urban “intelligentsia” from Dushanbe and other urban areas. The demographics would include the Russian, German, and Jewish diaspora as well as native Tajiks from the health, education sectors and the highly technical and skilled sectors of construction, oil and gas, and specialized industry. • Wave 2: The relatively older (and predominantly male) rural population seeking outside income to sustain households comprised of young children and the elderly. • Wave 3: The current emigration of younger males and females who have recently completed high school or matriculated and seeking entry level positions on the basis of limited or no professional experience. Of the three waves described, the second related to emigration of males has led to a a large number of female headed households within the republic, and in many remote villages, this has led to a number of social problems. Notwithstanding the large number of female headed households in the republic, a striking feature of the population is its youth and more particularly that the median age of the population is 21.6 years. With an almost equal sex ratio, approximately 35% of the population is under the age of 14. In view of dilapidated educational facilities, limited budgets, and a previous soviet legacy of educational training for its population, there is a growing fear of a historical legacy in which a generation of children may be less educated than their parents. Transportation corridors for illicit drugs from Afghanistan to Moscow and certain western European countries continue to exist and pose serious social issues for the development of the republic and its population. The incidence of HIV and AIDS continues to increase and stems from an increasing sex trade as well as the growing narcotic problem. Medical facilities are ill equipped or dilapidated and most trained medical professionals are either engaged in relatively better remunerated occupations or have emigrated to Russia in search of better opportunities. The soviet legacy of seeking treatment as opposed to concentrating upon prevention continues and is exacerbated by low incomes and the inability to afford preventative measures. 37  Kireyev (2006)  41  A formal market for agricultural labour does not function well in Tajikistan. However, the national government does institute a minimum (agricultural) wage rate and piece rate wages for cotton picking are mandated by local governments. Figure 5 depicts real wage rates for the agricultural sector.38 Given the dominance of cotton in the agricultural sector and since independent farmers who have extracted their land from the collective structures are residual claimants to profit as opposed to formal wage earners, it is reasonable to assume that figure 5 depicts real wages for cotton workers in the collective farm system. Figure 5: Real agricultural wages by month, 2000 - 2007 16  14  12  10  2000 2001 2002  8  2003 2004 2005 2006  6  2007  4  2  0 Jan  Feb  Mar  Apr  May  Jun  Jul  Aug  Sep  Oct  Nov  Dec  Source: State Statistical Committee of Tajikistan as found on In viewing figure 5, the distinction between wages offered in the pre-harvest season (January to August) with wages offered in the harvest and post harvest season (September to November) is clearly distinguishable. The former being fixed monthly wages while the latter comprises the fixed monthly wage in addition to the piece rate wages for picking cotton. In making the argument that labour is a critical constraint, one needs to distinguish between labour devoted to land maintenance, planting, seeding and other duties not related to the harvest with labour that is devoted to the harvest itself. Manual cotton picking and the labour exerted during this period is of critical importance in the quality of fibre recovered. In the absence of mechanized picking, the number of days between bloom, harvest and ginning has a definitive impact upon 38 Source: State statistical committee of Tajikistan as obtained on their official website Nominal wages were deflated using monthly CPI figures also obtained on  42  the grade of fibre recovered and price received. More specifically, given decreasing world prices for cotton fibre and poor ginning processes, it is quite possible to have a positive and increasing marginal product for labour but a decreasing value of marginal product for labour.39 From an efficiency perspective, this is confusing, considering that real wages have been increasing.40 However, the fact that cotton production has been incurring “losses” at the farm level, coupled with a lack of labour mobility in the agricultural sector would suggest that this notion is quite plausible and leads directly into the allegations that labour is very much coerced, both directly and indirectly, in the production of cotton. Indirectly through state plans dictating a certain area of land to be placed under cotton and directly through the use of edicts and orders which ostensibly force schoolchildren and state civil servants to pick cotton at harvest. Evans (1970) provides a lucid account of historical examples providing the basis for an argument that the essential element necessary for coerced labour is that other sectors of the economy, those not under coerced labour schemes, do not protest to the coercion of labour in the directed sector. In essence, coercion of labour is associated with a general state of labour shortage and in conjunction with, • A sector that produces a public good or service or one in which society places a very high value. • A “shortage” of labour at a particular or given wage rate which is attributable to the non pecuniary aspects of the employment. • That the “shortage” of labour cannot be adequately tackled or solved by raising wages. Cotton in Tajikistan is certainly one for which high value is placed at least within the national psyche. Updates on the cotton campaign (production versus plan) during the harvest season are regularly aired on the national television station ahead of any other national or international news items of note. The national emblem for Tajikistan is comprised of cotton buds, wheat stalk and the crown of a legendary king. More importantly, however, is the attention that cotton and cotton production gains from international organizations and non governmental organizations engaged in Tajikistan. However, there is clear indication that real wages are increasing in the agricultural sector and this would seem to signal economic growth in the economy. In a normal economy, increased wages would lead agricultural firms to economize on labour and this would be one plausible explanation for the observation of a shortage of labour in the field. In contrast, I will argue 39 Marginal product on a volume or tonnage basis may be increasing but given inefficient ginning, storage and handling processes, the price of the fibre recovered may (and has been) decreasing due to world market conditions as well as quality losses. 40 Wages may be still be low by some prescribed standard but the argument made here is that poor wages or slave type conditions do not necessarily follow from the coercion of labour as argued by a number of authors but most notably by Fogel and Engerman (1995).  43  in a later chapter that increased wages in the cotton sector, which represents a significant portion of the agricultural sector generally, may also reflect a growing divergence between private profitability in grain production and private gain from cotton production on collective operations. More specifically, I will advance an argument that wages for cotton picking and for the sowing and maintenance of cotton fields reflect a “plug” that equalizes the gap between private profitability of grain production and private gain from collective cotton production. If cotton prices are falling, and if grain production is relatively more profitable over time, increased wages for cotton production may reflect a desire for vested interests in the cotton sector to maintain a certain amount of land and labour to be devoted to cotton by equalizing the returns from private grain production and private benefits from cotton production. The latter includes an aspect of the gains from the diversion of inputs out of cotton production and into private grain production. Coercion of labour in the production of cotton may be too strong a statement in light of prevailing evidence.41  2.3 Productivity on existing cotton operations Increasing the output of cotton fibre can be brought about in three ways, (i) increasing the outturn ratio of cotton fibre to raw cotton at the ginnery; (ii) expanding the acreage devoted to cotton production and thereby increasing the total volume of cotton produced; and (iii) increasing the yield of raw cotton per hectare, of which fibre is an essential component. In all three cases, the common thread is one of investment and incentive. A lot has been made of the gins interest in delaying the ginning of cotton in order to extend the duration of the loan and to extract further interest income (see for example The Centre for International Studies and Cooperation (November 2006), World Bank (June 2004)). Industry analysts and various documents purport that the average ginning period in Tajikistan is on the order of 200 days in comparison to normal international standards of 100 to 120 days ( Caccavale (2005), World Bank (June 2004), World Bank (December 3, 2005b), World Bank (January 6, 2005a)).42 The physiological characteristics of cotton are of such nature that quality and yield of cotton fibre are reduced at an increasing rate after 100 days. The notion that gins are deliberately delaying the ginning process would be based upon an implicit assumption that the return on the loan extended (at the margin) is greater than the value of the marginal product of cotton fibre obtained for each day of delay until a critical period, which is on average the 200th 41  The conscription of children into cotton picking campaigns is of course a moral issue, particularly when they are deprived of schooling. This thesis does not dismiss that there are acts and conditions which impinge upon human rights issues. The point here is that there are regular adult farm workers, who choose to remain in the collective, while children are forced to join into the cotton picking campaign, a system that has historical Soviet vestiges. Having said that, it is also true that some of these children likely work alongside their parents in the field, thereby increasing family income, through piece rate wages earned in the picking of cotton. 42 While these references are somewhat older, visits to the field confirmed that there have been little if any investment and upgrading in the gins. The ginning period of 200+ days is still the norm, and in addition to older equipment, erratic electricity supply also hampers productivity and efficiency in the gins.  44  day after harvest. The notion and underlying assumptions would also implicitly suggest that the gins have little interest in improving farm productivity (yields) of the raw cotton output. The first assumption is related very much to the manner in which the future contracts are signed and the role of quality in determining the contractual price and the quantum of prepayment. The second assumption is related to the role and return on investment. A discussion on and an understanding of the constraints to productivity enhancements at both the gin and farm level helps to shed light on the validity of these assumptions. Outturn ratio between raw cotton and fibre Increasing the outturn ratio of fibre to raw cotton relates directly to incentives for the gins to invest in new equipment and technology in the ginning process. Current outturn ratios are on the order of 30% in comparison with ratios of 40% in more efficient cotton producing economies.43 Reasons ascribed for the low outturn ratio include, 1. outdated ginning equipment and poor maintenance, 2. theft by the gin through misreporting, and 3. little incentive for the gin to improve efficiency due to monopsonistic purchase ability While the first two reasons are plausible, the latter is more difficult to argue. The monopsonistic aspect of the gin is based upon the manner in which credit is advanced and the ability to collateralize the cotton harvest at the farm gate. In addition, the current tax code provides hukumats with a large revenue base from taxes on ginned cotton, and much incentive for the hukumats to enforce the direction of cotton from the farm gate to gins within the territory of their jurisdiction. Gins, therefore, have a territorial monopoly in the provision of input credit and in the securing of monopsony power in the purchase of raw cotton. Why would the gin not want to improve efficiency in its operations? Given its monopsonistic power, the gin stands to gain fully from any increase in turnover between raw cotton and cotton fibre. Moreover, if it is true that misreporting is occurring, the real rates of outturn must be higher than reported. The question then becomes one related to the economics of investing in new equipment and not upon poor incentives to invest in order to improve efficiency.44 Acreage expansion Figure 6 depicts a gradual increase in the expansion of land devoted to cotton since 1997 but further expansion, if desired, is limited by credit, labour and irrigation constraints. With current (outdated) seed varieties and little or no investment in physical irrigation infrastructure, as well as credit constraints, the intermediaries would need to spread out the prepaid revenue from 43  Source: Personal correspondence with various industry analysts as well as in World Bank (June 2004) and Bale (2008) 44 Information on firm specific ginning costs and the structure of ginning operations are not available in the public domain and industry representatives are weary of talking about such matters.  45  future sales, as production loans, more thinly. Without injecting any of their own private capital, this would likely reduce the productivity of cotton land, thereby yielding little if any benefit from extensive cultivation of cotton. Rural credit markets are generally thin in Tajikistan and given accumulated debt, the options for credit extension outside of the gin are limited. Quite apart from the impact upon land expansion, a lack of credit and financing has also impacted the ability to rehabilitate the very large and expansive irrigation networks which have, for the most part, been dilapidated. Water losses in irrigation are estimated at 40% and and more than 16% of the irrigated land is reported to be unusable due to salinization and water logging (The Centre for International Studies and Cooperation (November 2006)). Despite these challenges, the amount of land devoted to cotton production (as depicted in figure 6) would indicate an equilibrium quantity despite issues of “debt” in the production of cotton. Figure 6: Hectares of land under cotton and grains  Source: State Statistical Committee, Government of Tajikistan (2005),(2007),(2008) Yield enhancement Increasing yields on existing lands requires investment in new seeds, rehabilitation of dilapidated irrigation facilities, and more productive inputs. A sustained reduction in cotton yields between 1985 and 2006, which has widely been attributed to a lack of investment and poor crop rotation practices45 , is clearly evident and presented in figure 7. An analysis of trends in crop yields for the same period is presented in table 5 from which a number of interesting observations can be gleaned. 45 These include investments in the maintenance of irrigation infrastructure as well as new seed varieties and higher yielding fertilizers  46  Figure 7: Yields per hectare of raw cotton and grains  Source: State Statistical Committee, Government of Tajikistan (2005),(2007),(2008) Table 5: Average annual percentage changes in yields of major crops1 Cotton Grain Potatoes Vegetables 1  1985 to 2006 -3.53% -0.18% -1.65% -1.97%  1985 to 1991 -1.03% -0.79% -2.98% -0.68%  1992 to 1997 -2.54% +2.10% -4.10% -1.62%  1998 to 2006 1.60% 6.28% 7.26% 4.94%  Source: Raw data obtained from  Firstly, the observation that yield declines for the period 1985 to 1991 can be couched in the knowledge that Gorbachev’s reign as general secretary of the communist party between 1985 and 1991 was particularly hard on the agricultural sector and a number of policy changes were either being contemplated or implemented during this period.46 Chief among these policies was the transformation of a large number of kolkhoze and sovkhoze into “lease brigades” which, in essence, were self managed farms. Five points characterized the “Gorbachev” agricultural reforms,47  • Small cooperative groups and families would be allowed to lease farmland, mostly within the supply of and with technical support of parent collective farms, but with some brigades operating individually; 46  The break at 1991 is appropriate given that independence was attained in 1991. 1992 to 1997 reflects the period of the civil war. 47 Adapted from Kramer (1989).  47  • The most inefficient collective farms would be allowed to go bankrupt and then either dismantled and leased or annexed to more successful neighbouring farms; • Military manufacturing facilities would be called upon to provide inputs and equipment aimed at modernizing the food processing sector; • The government agency Gosagroprom, which Gorbachev had created in 1985 in which five ministries and two state committees were amalgamated in order to reduce bureaucracy was to be shrunk and all grass roots decision making was to be shifted to local officials; • A programme for enticing workers back to the countryside which consisted of improvements in rural infrastructure (roads, hospitals, schools ...) as well as a promise of continued city residence permits in the event that workers decide to move back to the city in the future. The overarching obstacle facing Gorbachev’s reforms was that with wages guaranteed, secured housing and some assurance of employment, the incentives offered from piecework payment and uncertain independence in farming operations were not persuasive enough to move the reforms forward. The potential for increased salary was dampened by the excess demand economy in which savings accumulated due to a lack of purchasable durable goods and other consumer items. Independence in farming operations was uncertain and in most cases, a failure, given that the parent collective farms still issued plans and cropping patterns.48 The desired result of increased productivity on farms did not materialize and both Figure 7 and Table 5 support this claim. Of particular interest is that many of these same disincentives continue to exist in present day Tajikistan, particularly those related to state plans and cropping patterns. The second interesting observation from table 5 is that the civil war was particularly hard on the agricultural sector, but that the period after the civil war has resulted in a swing towards positive and increased productivity for all products in table 5. In the case of cotton, declining productivity continued on a yearly basis, but certainly on an order much lower than witnessed during the Soviet period, and even positive gains in the post war period. If one considers the damage wrought during the civil war, social conflict between 1992 and 1997, as well as the stagnation of agricultural productivity during the period of the war, there is much to be said in the strong productivity gains for all mentioned products during the post civil war period. In choosing to highlight productivity trends for potatoes, grain and vegetables in comparison with cotton, the implicit argument is that productivity gains are being made in important food crops and that investments in high yielding inputs as well as improvements in land quality are being made on both large dekhan farms as well as private individual plots. Figure 8 depicts 48 In the case of grain alone, the fact that grain yields were increasing, even during the period of civil war (1992-1997) in post-Soviet Tajikistan, and have continued to increase thereafter, is one indication of what can be argued to be a “failure” of the Gorbachev policies pursued.  48  the relative wheat yields between various forms of agricultural enterprises in Tajikistan. Figure 8: Grain yields on independent farms, collectives and dekhan farm associations  3  25 2.5  TTonnes perr hectare  2  C ll ti Collectives  1.5  Private Farm Associations  1  0.5  0 1999  2000  2001  2002  2003  2004  Source: State Statistical Committee, Government of Tajikistan (2005),(2007),(2008) The “collectives” referred to in figure 8 comprise of all state enterprises that have still not engaged in a decollectivization process and include former sovkhozes, seed farms, or agricultural research stations. “Farm Associations” comprise of those enterprises that have undergone some type of restructuring or division of land but who still function on the same principles or operational standards as the former kolkhoze structures. “Private” refers to farms where individuals have extracted their land from former state structures and/or represent small plot “Presidential lands” or “kitchen gardens” which were provided to citizens for the purpose of household consumption. The most striking observation from Figure 8 is the almost constant differential between yields on private farms and those on collectives and farm associations. In noting that these are grain yield comparisons and that there is no direct interference from external agents in the production of grain, the consistent differential in yields is very likely a result of differential quality seed input and the very real possibility that independent farmers are obtaining “newer” generation seeds on the open market as opposed to collective and farm associations that are relying on inter generational seed stock produced and stored on farm. Why then do the collectives not plant the newer and higher yielding varieties of wheat stock? One answer that seemed to surface from field visits was that the newer and higher yielding varieties are “thirsty” for inputs, 49  particularly fertilizer, and that a lack of credit availability for the purchase of new wheat seed varieties as well as fertilizer for the production of grain mitigate productivity enhancements for grain on the collectives.  2.4 Fertilizer usage and input diversion Figure 9 depicts the application of mineral fertilizers in cotton production and the associated yield of cotton per unit of land. Three items are of immediate interest. Figure 9: Republican fertilizer use and cotton production  2.50  2.00  1.50  Mineral fertilizer use (100kg/hectare) Production of cotton (ton/hectare) 1.00  0 50 0.50  0.00 1998  1999  2000  2001  2002  2003  2004  2005  2006  2007  Source: State Statistical Committee, Government of Tajikistan (2005),(2007),(2008) First and foremost, is the interesting feature of increasing application of fertilizer per unit of land under cotton production. While one would expect variations of application per unit of land due to the vagaries of weather and possible replanting, sustained increases between 2001 and 2005 are somewhat more difficult to understand. What would seem to be clear is that the provision of (in kind) credit is not based upon any technical considerations but rather upon the quantum of credit per unit of land planned for cotton production. This has been confirmed through various field discussions with industry stakeholders. In other words, in a year where fertilizer prices are low, it is quite possible for the gin and the collective farm to negotiate a higher (than technically required) delivery of fertilizer. The theoretical arguments surrounding “loan pushing” are very apt in this case and would certainly not be resisted by the collective farm given that the fertilizer can be liquidated for cash (to be used immediately) or diverted 50  to non-cotton crop production. The second interesting feature from figure 9 is the role that weather and external factors play in the production of cotton as seen for the years 1998 to 2001, in which flooding damaged a relatively mature crop (1998-1999) and drought damaged the initial cotton seed (2000) requiring a second planting in those years. Third, an important observation from figure 9 is the rate of application for fertilizer against recommended norms of application for cotton. Current demonstration plots in Tajikistan are utilizing and recommending an application rate of fertilizer on the order of 200kg/ha for nitrogen, 60kg/ha for phosphorous and 60kg/ha for potassium based fertilizers. The statistics used in constructing figure 9 do not disaggregate on the basis of mineral composition and thus, likely, reflect a composite of all three types of fertilizer. Using as a base reference, the recommended rate of nitrogenous fertilizer application, the statistics would indicate that farmers are under applying fertilizer in the production of cotton. Yet, an observation of farm records, discussions with farm managers and on the basis of various documents49 , farms are obtaining fertilizer on credit (in kind) at rates on the order of 400 or 500 kilogrammes per hectare. In addition to greater procurement of fertilizers relative to application, there is also a widely held notion that fertilizers and inputs are often delivered late and to the detriment of the productivity (The Centre for International Studies and Co-Operation (2006), Caccavale (2005), Pomfret (2007), World Bank (June 2004), World Bank (December 3, 2005b)). Field interviews conducted during the course of this study could not corroborate this notion but neither could they discount the same. However, even if fertilizer and inputs were delivered late and not applied, they would certainly be accepted by the farm given that they have been acquired on in kind credit and given that interest has been accumulating on the delivered items. The relevant question is one of whether the inputs are stored and utilized in the following year or whether they are applied in productive uses other than cotton. That there is a strong sentiment in working documents by the World Bank, the Asian Development Bank and other independent studies about the (continual) problems related to late delivery of inputs would suggest that inputs delivered late were not held for the following year or at least not in their entirety. If storage and utilization in the following year was occurring, then the problem of late delivery of fertilizer would only be a one shot productivity decline in the first year that the inputs are delivered late. In all following years, a stock would likely be held and maintained to initiate the growing season with subsequent late deliveries coming in to augment the growing season. A conjecture that at least some of these late delivered inputs are utilized in non-cotton growing activities is also, therefore, plausible.50 49  see for example page 27 in The Centre for International Studies and Cooperation (November 2006) The late delivery of fertilizer may be ascribed to the fact that there are extensive subsidies on fertilizers provided by the Uzbek government to their cotton farmers during the cotton growing season and that while some of these “cheap” fertilizers are leaked into Tajikistan during the sowing season, a larger supply is leaked 50  51  Figure 10: Republican fertilizer use v. cotton and grain yields 2.10  1.90  1.70  1.50 Average cotton yields (ton/ha) Fertilizer use in cotton (100kg/ha)  1.30  Average wheat yields (ton/ha)  1.10  0.90  0.70  0.50 1998  1999  2000  2001  2002  2003  2004  2005  2006  2007  Source: State Statistical Committee, Government of Tajikistan (2005),(2007),(2008) Figure 10 depicts reported fertilizer applications on cotton fields along with cotton and grain yields. Average grain yields are a sector wide average and include all types of farm entities (farm associations, joint stock companies, collectives and private farmers). Given that there is a consistent divergence between private yields and yields on various other collective type enterprises, as previously shown in figure 8, the use of a sector wide average poses little difficulty in scribing the following arguments: • The variability in yields for both cotton and grains have exhibited a similar pattern for much of the period analyzed. The drop between 1998-1999 can be attributed to the impact of the civil war and resulting damage to infrastructure and market linkages, while in 2004-2005 the drop is attributable to inclement weather. While the impressive gains in productivity between 1999 and 2003 would seem to have tapered out, total volume of domestic production has remained relatively constant since 2002 for both grains and cotton given the increases in land areas for cotton and grain (c.f. figure 6). • Visually lagging grain yields to fertilizer applications (in cotton production) by one year suggests a relatively strong correlation between the (reported) application of fertilizer on cotton fields and grain yields. into the republic after the uzbek cotton seeds are sowed and fertilizers applied. Given similar weather patterns and cropping seasons, the Tajik sowing season is typically delayed relative to Uzbekistan.  52  Estimating a very simple linear production function for cotton and wheat against fertilizer use in cotton provides the following results: Cotton=  1.32  +  0.25 Fertilizer  (t-stat)  (5.69 )  (1.49 )  (p-value)  (0.00 )  (0.17 ) R2 =0.22  Wheat=  0.67  +  0.70 Fertilizer  (t-stat)  (2.24 )  (3.24 )  (p-value)  (0.06 )  (0.01 ) R2 =0.56  While based upon a very small sample and simple model, the results indicate that fertilizer (reported to be) applied in cotton production is a better predictor of wheat production than cotton production. The observation that marginal product of fertilizer in the wheat function is much higher, in relation to cotton, can be explained by arguing that there is a less than optimal level of (total) fertilizer delivered to the farm to produce both cotton and wheat efficiently. When this is the case, fertilizer can still be diverted out of cotton production, but remain in short supply, relative to demand, for the amount of land placed under wheat. In other words, in the absence of diversion, the marginal product of fertilizer in wheat production may likely be higher than the coefficient of 0.70 obtained in the simple regression. The results add credibility to the conjecture that fertilizer and other inputs delivered for the production of cotton are diverted out of cotton and into the production of other crops, particularly wheat. This chapter has outlined the background on various factor markets and a brief introduction to the institutional structure of cotton financing in Tajikistan. The discussion in this chapter and relevant statistics support a conjecture that the current system for financing cotton production provides (subsidized) access to inputs for grain production. In addition, with wages paid for cotton production and harvesting, the system acts as a hook for the gin to ensure that a relatively stable supply of raw cotton is produced by farmers. The story thus far would signal that farmers are engaged in the production of cotton, a loss making crop, largely because of thin credit markets, and governmental regulations that promote cotton production in order to earn much needed foreign exchange. I will argue in the coming chapters that while cotton is a loss making crop, the incentives for producing cotton at the farm gate are such that farmers are indifferent between producing cotton collectively and producing grain on privately held lands. The manner in which credit is extended, both in kind for material input provision as well as in cash for wages, plays a large role in maintaining this indifference.  53  Chapter 3 is devoted to understanding the nature of credit provision for cotton production as well as the nature of production on a “model” farm that produces both cotton and grains. Utilizing farm level data, I elaborate upon the nature of private gains from cotton production, through input diversion, and provide evidence to support a conjecture that (i) individual farmers may not be concerned about “debt” at the collective farm level, and (ii) that private benefits are enough to wipe out any collective “debt” that accrues during a cotton production season. Chapter 4 strengthens this conjecture through an analysis of survey data on 135 farms, with an argument that farmers producing cotton collectively are likely to experience an equal standard of living, relative to farmers producing only grains and other food crops.  54  Chapter 3  Production and profitability The conjecture raised in this thesis is that the gin ties together the services of (non- monopolistic) credit provider and monopsonistic purchaser of cotton, but at the same time, implicitly finances non-cotton crop production through its cotton financing scheme. The gin does not have an incentive to enter into marketing arrangements for non-food crops and nor does it encumber any other product or commodity other than raw cotton at harvest. Food crops, particularly grains, are consumed at the farm level and any surplus is sold into a competitive local market.51 Using the background on various (republican) cotton financing stages, as discussed previously, I will argue in this chapter that through its (cotton) financing scheme to the collectives, the gin ensures (i) it meets its own contractual obligation to deliver a certain amount of cotton to international purveyors from whom it receives advance payment prior to planting, (ii) in ensuring that a minimum level of basic food needs for individuals within the collective is met, through the implicit financing of inputs to produce non food crops, the ability to act as a monopsonist in the purchase of raw cotton is enhanced and (iii) that the gin protects itself against any financial loss through the mechanics of its lending arrangement with the collective. I expand and explain each of these three items through various analyses within this chapter. Chapter 4, which follows, analyzes survey data for 135 farms to assess the plausibility of the conjectures raised and argued within this chapter. I will further argue that while “debt” is being ascribed to cotton producing farmers and farms, a more accurate portrayal would be to classify “debt” as accrued losses of collective type farm enterprises that produce both cotton and food crops. Farm entities that are non-cotton producing do not, for the most part, accrue losses over cropping seasons as will be shown in the next chapter. It is likely the case, therefore, that these are not financial losses in the strictest sense and merely costs that are incurred by the gin in the acquisition of raw cotton from the farm. In other words, “debt” is part of the process rather than an outcome which entails a financial obligation. A fuller description of production processes, market value chains and the nuances that drive incentive structures within this sector are provided in more detail, and assist in contributing to and refining previous and current claims that are sometimes based upon general observation. 51  The gin can, and does purchase grain from collectives, but there is no evidence that monopoly power is exercised in the purchase of grain and, indeed, evidence suggests that grain is purchased at market rates.  55  This chapter is divided into two parts. In the first, a discussion is undertaken on (i) the origins of farm debt, (ii) what role the gin plays within the current environment and (iii) detail on the institutional structure for the production, processing and marketing of cotton within the republic. The second part of this chapter is an analysis and discussion that gets to the question of why the gin resorts to interlinking and what advantage does interlinkage provide? Why does the gin not become a monopoly supplier of credit to all farmers (independent and collective) and a monopsonist purchaser of both cotton and non-cotton production? Why can the gin not benefit from advancing credit for grain as well and also collateralize the grain? Monopoly power can be enhanced on the basis of collateralizing the output (grain or cotton) and the moral hazard problem is essential solved because without any cotton or grain, no credit will be advanced in the following season. Is there something fundamentally different between collateralizing cotton versus grain? This chapter argues that one explanation for why farmers remain within this current system is that the production of cotton provides access to working capital for the production of a variety of agricultural products and commodities. The stability of the system is underpinned by an environment where credit markets are thin and weak, where state mandated wages act as supplementary security in protecting livelihoods, and where incentive structures for all stakeholders are aligned.  3.1 Institutions and cotton “debt” One area that has not been extensively analyzed in the study of the collective farm is the interdependency between collective good production and private good production. More specifically, there are two types of interdependency on a soviet style collective farm, (i) output interdependency and (ii) input interdependency. In the case of the former, production for the state on collective plots of land invariably yielded joint products that were used as inputs for private production of agricultural produce and livestock. Wheat chaff, cotton stalks (also used for heating), and cotton meal (joint product from the crushing of cotton seed for oil) as well as other types of fodder were used for feeding privately owned animals but which accrued from the pursuit of collective production. More importantly, however, and the crux of this thesis is the argument that collective production provided the means for obtaining critical inputs for private production.52 Where credit markets are thin and restrictive in terms of types of collateral acceptable, production of the collective crop facilitates access to inputs needed for private production and thus the observation of input interdependency. Fertilizer is one classic example of a fungible input that could be diverted out of, for instance, (collective) cotton production and into private grain 52 Trevisani (2007) notionally discusses the incidences of input diversion in present day Uzbek collective farming while Wadekin (1973) discusses input diversion during the Soviet period in somewhat more detail.  56  production. Potato seed is another example of a fungible input but is less obvious without an understanding of the production process and an appreciation for directed markets. The classic soviet collective was well know for the production of potatoes destined for distilleries and the production of vodka. The variety of potato chosen had specific genetic characteristics designed for the production of vodka but this did not exclude it from being palatable on the dinner table. Verbal discussions with older soviet bureaucrats suggest that collective plots under potato production were, therefore, often under armed guard in order to mitigate “looting” of the potato crop. Yet, where looting of final product may have been mitigated, there was an under appreciation of the value of potato seed, which is produced as a joint product.53 In an abundant harvest, it would have been relatively easy to underreport the harvest and to siphon off the seed to private lands, while still meeting production targets (for potatoes delivered to the distillery), and enhancing private production of potatoes (for table consumption or vodka) in the following season. The point being made is that a thorough understanding of particular production processes and market value chains, and the nuances that drive incentive structures within a collective system are important in the ability to refine and add to previous theories which are based upon generalized observation.  3.1.1 Origin of “debt” At dissolution of the Union, state financing of the cotton sector ended and both sovkhozes as well as kolkhozes were underfunded. To remedy the situation, the government entered into a sovereign agreement in 1996, to obtain a loan through Credit Suisse-First Boston with the backing of Paul Reinhardt, Swiss purveyors of cotton.54 These funds, initially in the amount of $1 million (in 1996) and later expanded to $60 million, were chanelled through the gin for onward lending to farms. With no formal systems for securing collateral at the time, the gin’s provided credit in kind to sovkhozes and kolkhozes with only an understanding the harvest would be handed over as collateral against the loan. By early 2000, the system began to break down as cotton receipts were insufficient to cover the cost of loan repayment. The sovereign debt has since been extinguished but only due to the purchase of the Credit Suisse-First Boston debt by Paul Reinhardt. This debt still remains outstanding and forms part of the current outstanding debt of approximately $550 million, which is currently booked to the account of collective type farms in Tajikistan. Formal crop lien legislation had not been enacted at the time that the sovereign debt agreement was signed and during the initial distribution of the $60 million and till present, there 53  Potato production yields two important outputs (i) potatoes destined for the table market or for the processed market and (ii) seed potato to be used in the following year production. In the case of the latter, it is the smaller sized potatoes which are reserved and set aside over the winter in order to be utilized as seed in the following spring sowing. 54 Paul Reinhardt is a worldwide cotton merchant trading company with a history dating back to 1788. Based in Switzerland, the company also maintains offices in Tajikistan, Uzbekistan, Turkmenistan, Brazil, China, Egypt, Greece, Latvia, United Arab Emirates and USA.  57  are no legally binding mechanisms for mortgaging land. The current crop lien legislation which was enacted in 2001, gives priority for seizing a pledged asset to the initial creditor. In other words, any creditor who provides finance with the forthcoming harvest as collateral, and who does not obtain full payment against the terms of the loan, is eligible to secure a lien against next years crop ahead of any other creditor who may provide credit to the indebted farm in that following year. One avenue for the emergence of interlinked credit and output markets is therefore provided by the current legislation as competition for credit becomes limited in such an environment and given the requirement for raw cotton to be processed in order for value to be realized. Given that Paul Reinhardt continues to be a prominent purchaser of cotton from Tajikistan, on a prepayment basis prior to harvest, and continues to hold a debt note against the Bank of Tajikistan, signals that there are profits to be made within the industry but raises numerous questions on where these profits accumulate.  3.1.2 Cotton contracts and pricing In this section, I concentrate upon stage two of the cotton financing schemes that were highlighted in the previous chapter. This stage was characterized by a period in which foreign exchange earnings were dwindling, due to relatively lower cotton and aluminum prices, and where the threat of capital flight prompted measures for controlling the flow of cotton fibre out of the republic. One of these measures was the institutionalization of an exchange, the Tajikistan Universal Goods Exchange (TUGE). The Tajikistan Universal Goods Exchange (TUGE) is the authority for approval and registration of all contracts pertaining to the purchase of cotton fiber both domestically within the republic and cotton destined for export. All cotton fibre marketed internally within the republic or destined for export is subject to a minimum floor price, calculated daily for each of the five grades, and set on a ex-works basis by using the Liverpool Cotton Exchange price quotation Index for Central Asia. The minimum price stipulates all costs related to taxes, transportation, commission and insurance. A 2% discount is also provided for the purchaser of cotton (international or domestic) as an incentive for the requirement to prepay at the time of executing the contract. In essence, and as argued shortly, the prepayment is the working capital fund to finance the cotton crop for that season.55 At the time of delivery, typically 6-8 months after the signing of the contract, any variance between the value of cotton shipped and the value prepaid is adjusted and cleared. The TUGE will not register any contract, in which the agreed upon price is below the mandated or administrative price set by the TUGE and, therefore, a minimum price is set daily for the purpose of calculating export taxes and other charges related to the export of cotton fibre. An example of pricing is given in table 6 and pertains to prices set on February 22nd, 2006: 55  On this basis, one could argue that 2% is the cost of capital to the credit lending gin.  58  Table 6: Methodica - minimum FOB price calculation for February 22, 2006  Grade 1 2 3 4 5  CIF (US$/ton) 1398.61 1344.82 1264.13 1210.33 1075.85  Price adjust. +4% 0% -6% -10% -20%  Commission (1.5%) 20.98 20.17 18.96 18.15 16.14  Expenses Insurance Discount (1.0%) (2.0%) 13.99 27.97 13.45 26.89 12.64 25.28 12.10 24.21 10.75 23.52  Freight (fixed) 60.00 60.00 60.00 60.00 60.00  Total expense 122.94 120.51 116.88 114.46 108.41  FOB 1275.67 1224.30 1147.24 1095.87 967.44  Source: Government of Tajikistan (2007) The reference, or base price, is based upon grade 2 cotton fibre, with a premium for grade 1 quality cotton fibre and a markdown for grades 3-5 relative to the grade 2 pricing. The price of $1344.82 for the grade 2 reference price was calculated on the basis of a Liverpool price of 60 cents per pound and the FOB price of ($1244.30) is derived from the deduction of commissions, freight, and insurance charges as detailed in table 6). On the basis of this FOB price calculated above ($1244.30), the ex works price of ($946.95) is derived through the deduction of Tajik state standard certification and domestic charges for internal freight and handling. These charges are detailed in table 7 below as an example of actual prices and cost on February 22nd, 2006.  Table 7: Methodica - minimum ex-works price calculation for February 22, 2006 Expenses FOB price (per ton)-Grade 2 Sales tax Commission to TUGE Issuance of certificate by Chamber of Commerce and Industry State Plant Quarantine Inspection Compliance certificate issued by Tajikstandard Railway tariff within the CIS Wagon charges Station charges, handling and storage Total Expenses Ex Works Price  % 10 0.1 0.1 Fixed Fixed Fixed Fixed Fixed  US$/ton 1244.30 122.43 1.2243 1.2243 0.42 1.0 135.38 2.00 10.00 273.67 946.95  Source: Government of Tajikistan (2007) The institutionalization of the TUGE was a response by the government to remedy two problems. First, was the inability to control the export of a strategic crop within the republic. Cotton and aluminum together account for over 75% of the country’s value of exports and without a central clearing agency, it was felt that there were losses in the collection of customs revenue through smuggling of cotton across borders. To be more specific, in the early stages of  59  the protracted transition56 , most enterprises were still owned by the government and it was felt that a central clearing house would avoid the potential for selling at artificially low prices with the intended benefit of the margin between the stated price and the realized price ending up in the pockets of the directors of state agencies. In the current environment, private agencies engaged in the export of goods (primarily cotton) are ostensibly restricted from undervaluing their export and in an attempt at mitigating the flight of capital out of the republic. Second, and perhaps more importantly, was the lack of financing for the agricultural sector and particularly for the production of cotton. In attempting to collect relevant data from official agencies in Tajikistan, one stumbling block was that the TUGE was unwilling to provide any information on the flow of cotton through its agency. There were also, surprisingly, no official documents or publications detailing the workings of the TUGE and volumes of commodities registered through the TUGE. At the state statistical agency (Goscomstat), it was also apparent that national statistical data on cotton fibre exports were collected directly from the gins and not from the TUGE. Given this observation, and in line with a private telephone conversation with a Tajik cotton trader living in Moscow, there is good reason to believe that the TUGE is not living up to its apparent mandate of regulating the flow of cotton fibre out of the republic. The move by the government to facilitate stage 3 of the financing scheme would support this notion. The gins continue to deal with international purveyors on a contractual basis and, it would seem, that the role of the TUGE now is merely one of collecting state revenue through an export tax that is applied to an administrative price set by the TUGE.57 This administrative price may or may not be indicative of the actual prices received by the gins and logic would suggest that if the gins are bypassing the TUGE, the administrative price is lower than the actual price received by the gins. Given this observation, the question at hand is why the gins only engage in cotton and not in other commodities or agricultural products.  3.2 Financing of wheat production Consider the following hypothetical argument in order to assess potential arguments for why the gin may not wish to finance non-cotton crop production. A gin who extends credit for both cotton and grain production has a production function comprised of income earned from the extension of credit to farmers, revenue earned from the sale of cotton fibre and any margin that 56  Due to the civil war, between 1992 and 1996, efforts aimed at privatization, liberalization and stabilization did not really begin in earnest until the late 1990’s. 57 The private trader contacted stated that the administrative price was typically well below the prevailing world price and allowed a margin to be obtained through a separate contract with the international purveyor. Two contracts therefore prevailed. One for regulatory purposes with the TUGE, from which export taxes and other charges were calculated and a binding contract with the international purveyor which detailed volume and purchase price. In the case of the latter, the margin between the administrative price and the world price was remitted to a foreign bank account and in line with figure 1 previously presented in the introductory section. The TUGE, therefore, does not provide the desired and intended impact of undervaluing cotton exports.  60  it may obtain from the sale of grain.58 Let Pc∗ be the final price of cotton fibre received, Pa be the contracted price for cotton set by TUGE prior to harvest and Pc the price paid for raw cotton to the farm. If Pc∗ is lower than Pa , the contracted price prevails in all circumstances. Pw∗ represents the price of grain received by the gin in the local market, whereas Pw is the price paid to the credit taking farmer. Qc is the amount of cotton supplied to the gin, while Q̄c is the amount of cotton contracted. Qw is the amount of wheat purchased by the gin. Lc and Lw are the quantum of loans extended for the production cotton and grain respectively whereas ic and iw are the respective rates of interest charged. The profit function for the gin is given by,  max  Πg = Qc [Pc∗ − Pc ] + Qw [Pw∗ − Pw ] + ic Lc + iw Lw subject to :  Lc + Lw ≤ Pa Q̄c  (3.1)  The constraint reflects the observation and arguments that contracts for cotton and prepayment against these contracts are the basis for financing the working capital requirements for the production of cotton. In other words, the gin does not invest any of its capital for the purpose of extending loans and disburses only that amount which it receives as a prepayment against the future contract to supply a specified amount of cotton. Rearranging 3.1 and embedding the constraint (Lw = Pa Q̄c − Lc ) into the objective function yields, max  Πg = Qc (Pc∗ − Pc ) + Qw (Pw∗ − Pw ) + Lc (ic − iw ) + iw Pa Q̄c  (3.2)  Maximizing 3.2 with respect to Lc , the value of loans extended for cotton production, yields c a first order condition iw = (Pc∗ − Pc ) ∂Q ∂Lc −  ∂Qw ∗ ∂Lw (Pw  − Pw ) + ic . The relative rate of interest  charged for loans in the production of grain relative to cotton will depend upon the difference in profitability (at the margin) of arbitrage in the purchase and sale of cotton and grain. When credit markets are weak or poorly functioning, logic would suggest that gins, working within demarcated territories, would resort to differential pricing for loans extended towards cotton and non-cotton production. Yet, this is not observed and corroborated on the basis of discussion with farm managers and industry analysts. The gin extends loans for cotton production and with, what would seem, full knowledge that a portion of these loans are siphoned or diverted out of cotton production and into non cotton production. One clear explanation in this regard relates to monitoring and the inability of the gin to observe the full application of inputs into dedicated cotton production. The incentive to divert would increase if the rate of interest for loans extended for the production of wheat is higher than for cotton. What is confusing, however, is why the gin adheres to a practice of providing loans for inputs in kind, 58  For simplicity, profit from the ginning operation is omitted and consistent with observation, loans are assumed to be provided in kind.  61  such as fertilizer, at rates which are much higher than existing norms for application of fertilizer on cotton fields. Evidence on this is provided shortly, but the issue revolves around a question of why the gin does not adhere to scientific or accepted norms for fertilizer application in cotton production and offer distinctive loan packages for cotton and non cotton production. When the gin cannot adequately monitor fertilizer application and if the rate of interest for non cotton production is higher than cotton production, the potential for leakage of inputs out of cotton production is predictable.59 A more convincing explanation, however, for why the gin may be averse to extending loans for non-cotton production within this environment lies in the pricing structure for grains (the largest non-cotton product). Tajikistan in a net importer of grain and has been for most of its history. Since 1996, the World Food Programme, along with a host of other international organizations have been distributing wheat and flour to rural residents in order to improve nutritional standards and to stave off the onset of hunger in remote and isolated mountainous villages. Given a liberal importing regime for wheat and wheat flour, local prices for wheat and flour tend to remain relatively low. Moreover, given the availability of imported wheat throughout the year, as well as the custom of storing wheat and flour within households for security, the tendency for local wheat prices to rise in the non harvest season is limited by the landed import price.60 Where wheat prices are relatively flat or perhaps even falling for parts of the year, there is little to gain by the gin in enforcing the collateralization of wheat or grain as security against loans received. In addition to leakage (of both inputs and outputs), the inability to arbitrage in the local or international markets, for wheat, is also one likely reason for why the gin may not exercise monopsony power in the purchase of wheat. Turning back to the optimal rate of interest for wheat loans, c iw = (Pc∗ − Pc ) ∂Q ∂Lc −  ∂Qw ∗ ∂Lw (Pw  − Pw ) + ic  when the ability to arbitrage in the grain market is limited, (Pw∗ − Pw ) is effectively zero and thus iw will always be greater than ic particularly if the gin exercises monopsony power in the purchase of cotton. Interest rates on cotton and grain will only converge if either (i) the gin does not exercise monopsony power and pays the farm the administrative price for cotton or (ii) when farms are efficient in their application of inputs into cotton production such that  ∂Qc ∂Lc  is zero. The latter is unlikely in a case where credit markets are thin and when diversion of in kind inputs out of cotton production is taking place. 59  This is problematic for the gin in that if it attempts to directly control the market for credit to both cotton and non-cotton crops, it may risk its ability to live up to its contractual obligation of delivering a specific amount of cotton to its contracted purchaser. 60 If wheat production is uneconomic for most of the year, the decision to grow cotton in order to siphon off inputs into grain production is predicated upon the conjecture that wages earned in the harvest of cotton and used to purchase food (or basic needs) above the amount of wheat or food crops grown on collective farms is one incentive that keeps farmers on collective type farming entities. This is in line with the conjecture, that farmers grown cotton at a “loss” collectively, in order to obtain productive inputs for non cotton production. As will be shown shortly, wages earned through cotton production are an integral part of the story of “debt”  62  A second potential reason for why the gin may be averse to lending directly for grain and collateralizing the grain harvest is that unlike cotton, grain can be immediately used by the farm household with minimal processing. The incentive for the farm to hoard or to not truthfully report the full value of the grain harvest increases when food security is of concern and if the price paid by the gin is lower than available in the local domestic market. Raw cotton has the advantage that it is of little use to the farm household who cannot access the services of another gin (territorial monopoly) and given that the processing can not be effectively be conducted on farm (high capital investment). Providing loans solely for the production of cotton and collateralizing the forthcoming cotton harvest provides little risk of loss (with the exception of weather related crop failure), enhances the possibility to gain from arbitrage, provides profit from the ginning services rendered and provides profit from the extension of credit. In strengthening this argument on why the gin may find little advantage in exerting any market power in grain production, the following section analyzes more carefully, the issue of marginal product of cotton production with respect to input provision and within the context of input diversion.  3.3 Model The analysis below is based upon observation within the current cotton sector in post-Soviet Tajikistan. Prevailing data and studies are used to test the conjecture that incentive structures may be aligned for both (i) the credit lending gin who collateralizes and purchases the harvest of cotton and (ii) the collective farm that accepts credit (some in kind), and produces both cotton and grain on a farm in which individual members are shareholders with entitlement to land, but who have banded together in a spirit that is reminiscent of the former Soviet kolkhoze. As in the southern US factorage system, discussed within the literature review, there is a prevalence of debt but continued advancement of loans over successive cropping seasons. In both cases, the inability for a mortgage system and property liens to play a role in lending suggests that debt is relative to cash flow and not to assets. The literature on southern cotton would suggest that planters were very much forced into the production of cotton either through the enhancement of capital assets in the form of slaves (ante bellum) or through the spirit of chance which provided the hope of unshackling themselves from debt to the country store (post bellum). The fundamental difference between the manner in which the ante and post bellum south literature portray the debt problem and the portrayal of debt in the production of cotton in post-Soviet Tajikistan is that in the current situation farmers have a choice in the production of crops and that the gin allows the production of crops for private consumption as well as crops which are collateralized against the credit advanced. Farmers in the post bellum south were  63  not accorded the same privilege of choosing their optimal crop mix or land use plan. Tying together these pieces where the two systems mirror each other and instances where they diverge, one could argue that both cases are reflective of the literature on bankruptcy for profit. More specifically, within the spirit of the analysis given by Akerlof and Romer (1993), the post bellum cotton situation can be described by the notion of going for broke or “gambling on success”(looting) whereas the situation in post-Soviet Tajikistan is very much related to a system where firms (farms) have an incentive to go broke and to profit at the expense of the public purse.61 I will attempt to show that monopsony power is present when the gin allows the collective to produce private crops, with inputs provided on credit, but collateralized on the output of the collective crop. In the case of work on the collective farm, it is the wage benefits which play a large role in the indifference between alternative production possibilities (private or collective). More specifically, I show that profitability of the collective farm is dependent upon wages but that cotton supply is not. The model that is developed in the following sections argues that the gin extracts surplus by paying the collective farm a price for raw cotton that is lower than the prevailing world price, and thereby capturing the mark-up. This results in a lower than efficient quantity of raw cotton delivered by the farm. In order to compensate for this lower than efficient delivery, the gin, in conjunction with the local government, demands a higher than efficient volume of land to be allocated to cotton production. I further show that the interest rate, one of the tools in the arsenal of the gin, which can be used to influence cotton supply, will either be high or close to zero. A high rate of interest raises revenue for the gin but, ceteris paribus, increases the cost of raw cotton that it purchases from the collective farm. The rate of interest, therefore, will depend upon which effect dominates. To the individual collective farmer, the rate of interest and the price of cotton are of little consequence, as the primary motive for contracting with the gin is to obtain access to credit. Credit obtained for cotton production implicitly cross subsidizes the production of non-cotton crops, particularly grains, and wages are the plug with which benefits for individual collective farmers are equalized with benefits that could be gained through production of grain on privately held lands. Why the gin chooses such an inefficient contracting system, as opposed to joint farm-gin profit maximizing rules, is argued to be based upon reasons of political economy. 61  Given continual discussions and public debates over amnesty for cotton farm debt, one could argue that the threat of liquidation and bankruptcy for indebted farms and farmers is not credible.  64  3.3.1 Farm: Part 1 Consider a stylized collective that produces two goods, a cash crop (cotton) and a food crop (grain).62 There are Fi members within the collective, all of whom hold shares for the collective land. None of the members possess enough capital to fund the working capital requirements of the collective and thus require seasonal loans. The territorial gin extends loans in the form of in kind material inputs (fertilizer, seed, fuel and other operating expenses) and in the form of cash for the payment of wages, both maintenance wages and piece rate harvesting wages. The extension of the loan is contingent upon the collateralization of the forthcoming cotton harvest. The collective farm, on the basis of a business plan submitted each season, requests the quantum of inputs required, the value of cash advances required for wages, and details the number of hectares to be placed under cotton. In the case of the latter, it is the local government (hukumat) that dictates the allocation of land to be devoted to cotton whereas the republican government sets the national plan for raw cotton to be produced. All members of the collective work on the farm and receive a monthly maintenance wage of w̄ as well as a piece rate wage at harvest of w̃X, where X is the amount of raw cotton harvested. The total wage bill for the collective is, therefore, w̄Fi + w̃X where Fi are the number of workers within the collective. The production of cotton facilitates access to working capital for the farm and while the request for inputs is made for the purpose of producing cotton, the collective inflates the amount of material inputs required for the production of cotton in order to facilitate the production of grains. An amount of material inputs N is requested for the purpose of producing cotton but only a portion Nx is utilized in cotton production with the remainder, Ny being diverted into the production of grain. When the hukumats announce the allocation of land to be placed under cotton, they do so on the basis of the indicative plan set for tonnage of raw cotton as announced by the republican government and on the basis of prevailing norms for cotton yields per unit of land. Diversion of material inputs out of cotton production and into grain results in actual (cotton) yields which are lower than the existing norms. Table (8), a duplicate copy of Table (4), indicates that the plan has not historically been met and in most cases production of cotton has been significantly under plan. Denote the production functions for cotton (X) and grain (Y ) as,  X = Zx (Nx )α (3.3) Y = Zy (Ny )β 62  Grain is used as a proxy for all food crops grown on the farm which can be distributed on some share basis among collective members or sold in the local market.  65  Table 8: Land under cotton, planned cotton production and actual yields (1998-2007)  Republic  Sughd  Khatlon  RRS  1998 1999 2000 2001 2002 2003 2004 2005 2006 2007 1998 1999 2000 2001 2002 2003 2004 2005 2006 2007 1998 1999 2000 2001 2002 2003 2004 2005 2006 2007 1998 1999 2000 2001 2002 2003 2004 2005 2006 2007  Land (ha) 247347 247800 238608 257371 269197 284317 293516 288655 262893 254830 74593 72064 69088 75267 80834 85508 88176 86392 80474 75867 151129 154992 150287 160505 165367 173527 178761 177036 161222 161230 21625 20744 19233 21599 22996 25282 26489 25227 21197 17733  Plan (tons) 600000 600000 600000 600000 600000 610000 610000 610000 547000 550000 187200 187200 187200 187200 187200 190000 190000 190000 171300 171300 360000 360000 360000 360000 360000 365000 365000 365000 329000 330500 52800 52800 52800 52800 52800 55000 55000 55000 46700 48200  Yield (plan) (tons/ha) 2.4 2.4 2.5 2.3 2.2 2.1 2.1 2.1 2.1 2.2 2.5 2.6 2.7 2.5 2.3 2.2 2.2 2.2 2.1 2.3 2.4 2.3 2.4 2.2 2.2 2.1 2.0 2.1 2.0 2.0 2.4 2.5 2.7 2.4 2.3 2.2 2.1 2.2 2.2 2.7  Realized output (tons) 383654 313125 335427 452735 515478 537358 556766 447918 437898 419786 128657 115107 120958 145347 154871 148466 159424 155526 131402 124830 217717 171658 177950 255092 305754 334063 334079 240884 258073 259643 37280 26360 36519 52296 54853 54829 63157 51508 48423 35313  Realized yield (tons/ha) 1.6 1.3 1.4 1.8 1.9 1.9 1.9 1.6 1.7 1.6 1.7 1.6 1.8 1.9 1.9 1.7 1.8 1.8 1.6 1.6 1.44 1.11 1.18 1.59 1.85 1.93 1.87 1.36 1.60 1.61 1.72 1.27 1.90 2.42 2.39 2.17 2.38 2.04 2.28 1.99  plan (%) 63.9 52.2 55.9 75.5 85.9 88.1 91.3 73.4 80.1 76.3 68.73 61.49 64.61 77.64 82.73 78.14 83.91 81.86 76.71 72.87 60.48 47.68 49.43 70.86 84.93 91.52 91.53 66.00 78.44 78.56 70.61 49.92 69.16 99.05 103.89 99.69 114.83 93.65 103.69 73.26  Source: State Statistical Committee, Government of Tajikistan (2004), (2007), (2008)  66  where, Zx and Zy (Z̄ = Zx + Zy ) are the amount of land devoted to cotton and grain respectively. Nx and Ny are, therefore, inputs per unit of land. Letting l represent the interest rate on the loan advanced, and Pf as the market price for material inputs such as fertilizer63 , total cost for the collective production of cotton and wheat is given by, T C : (1 + l)(Pf (Nx Zx + Ny Zy ) + w̄Fi + w̃X)  (3.4)  The total cost of production for grain is, however, contingent upon the value of raw cotton purchased by the gin, and who sets the price Px . More specifically, the cost of producing grain is given by, (1 + l)(Pf (Nx Zx + Ny Zy ) + w̄Fi + w̃X) − Px X which is the residual between the value of the loan received and the value of raw cotton surrendered to the gin at harvest. The term Px X represents the collateralized value of cotton at harvest, which is surrendered by the farm to the gin. The intuition at the heart of this specification is that the collective, and its members, utilize the production of cotton as a means for obtaining working capital and material inputs in the production of grains. The maintenance wage, w̄ plays a large role here in providing a basis for livelihoods during the growing season. There is no expectation that cotton will yield revenue to the collective over and above the cost of the credit extended by the gin which includes a component related to the provision of material inputs used in the production of grain. On this basis, revenue for the collective is given by the return from grains64 which has a market value of Py per unit. The profit function for the collective is given by, Πf = Py Y − (1 + l)(Pf (Nx Zx + Ny Zy ) + w̄Fi + w̃X) + Px X which is subject to the production function constraints X = Zx (Nx )α and Y = Zy (Ny )β Maximizing with respect to Nx and Ny yields first order conditions, Py  ∂Y = (1 + l)Pf Zy ∂Ny  and  (Px − (1 + l)w̃)  ∂X = (1 + l)Pf Zx ∂Nx  (3.5)  The first is the standard result that the value of the marginal product for grain resulting from an increase in the application of fertilizer to grain production equals the credit price for the 63  1 + l can also be interpreted as the price for the full bundle of inputs into both cotton and grain production. I use the term fertilizer as a proxy for all material inputs into the production process. 64 It is not necessary for the collective to sell the grain in local markets and that distribution among the members for consumption is equally valid as there is economic value to the grain disbursed when valued at market rates.  67  fertilizer. The second condition for efficiency states that when a unit of fertilizer is increased to cotton production, the credit price for fertilizer must equal the net marginal value between the farm gate price received for cotton (Px ) and the extra cost of wages (w̃) incurred in the delivery of the extra amount of cotton harvested and resulting from the extra unit of fertilizer applied. In order to validate the efficiency conditions given in (3.5), estimates for marginal products of fertilizer in cotton and grain production are required. Efforts within this study to collect data that would assist in estimating marginal products for both grain and cotton were not successful, for a number of reasons, and largely because of the nature of the cotton industry in Tajikistan and vested interests. Fortunately, however, copies of actual financial records between the gin and a certain collective were acquired and shed light on a number of issues while concurrently assisting in conceptualizing ranges for marginal products of fertilizer use in cotton and grain production. Table 9 is a translated copy of the original document shown in figure 11 and which relates to a summary of credit for the year 2003 as well as actual disbursements of credit and purchase in 2004. The conjecture that collectives utilize the production of cotton as a means for obtaining working capital for the farm and for production of other crops and commodities is bolstered by the observation that the gin finances and extends capital for a number of operational expenses (primarily taxes and utilities) but also the extension of credit for both maintenance wages and harvest wages. In the case of the latter, harvest wages are specifically piece rate wages extended for cotton picking. The preamble to the original document in Russian states that the extension of credit was based upon an agreement for 750 hectares to be placed under cotton and that the credit (for cotton production) is commensurate upon that agreement between both parties. The document clearly reveals that credit is extended for an amount that covers total farm expenses (including taxes, utilities and water cost) and not simply crop financing related solely to cotton. The total amount of land within the collective was given to be 1218 hectares and while 750 hectares was agreed upon as an allocation for cotton, 807 hectares were actually placed under production. The claims that the hukumats enforce collectives to place 70% of available land into cotton production cannot be refuted in this case, as the 807 hectares represents 66% of available land and very close to 70% of land that is not in fallow. The combined price for cotton purchased from the collective and applied against the loan advanced in 2004 was calculated to be $243.40 per ton. Of note is that in the original document, there is no entry for the price per unit and only mention of the total quantity of cotton supplied and the total price paid in US dollars. Interest charges between February 17th, 2004 and January 1st, 2005 were calculated at 4.6% per annum, on the basis of a principal amount of $448,245.20 and 319 days. The total amount advanced was $528,703.84, $62,185.09 was the profit in 2003 which was being held by  68  Table 9: Dzerzhinsky farm financial record 2004 Unit Amount Price ($ us) Last year’s debt (2002) Crop financing (2003) Interest Cotton Sales (2003) Other sales (2003)  Tons  1652.59  319.06  Tons Litre Tons Litre Tons Tons Tons Tons Litre Litre Litre Tons Somoni  121.966 134041 8.402 2220 0.18 227 152.55 120 220 70 200 102.5 68469.2  391.52 0.36 673.71 0.66 1381.94 163.4 161.43 152.57 25 18.09 16.08 213.66  Somoni Somoni Somoni Somoni Somoni Somoni Somoni Somoni Somoni  29300 401962 225969.54 4401.33 3000 11000 72500 34515 18760  Debt (end 2003) Crop financing (2004): 1 Diesel fuel 2 Diesel fuel 3 Diesel oil 4 Diesel oil 5 Grease and lubricants 6 Ammonium saltpeter 7 Carbamide 8 Superphosphate 9 Pesticide ”Polo” 10 Pesticide ”Tolstar” 11 Pesticide ”Omait” 12 Cotton seed for sowing 13 Spare parts Cash advances and payments: 14 Finance charges Wages Harvest wages Electricity costs Water costs Fuel and lubricants Fertilizer Budget taxes Social Service Taxes 15 Interest (17/02/04 to 01/01/05)  Total advanced (2004) Net balance owing Cotton purchases (2004) Corporate debt  Tons  69  1664.341  Balance ($ us) -31762.79 498986.62 11596.8 478820.63 527281.2 13724.52 541005.72 -62185.09  47751.62 48337.00 5660.51 1474.52 248.75 37091.80 24626.15 18308.40 5500.00 1266.30 3216.00 21900.15 23356.75 238737.95 9815.25 138032.92 74701.86 1467.18 1019.09 3786.57 24775.59 11697.24 6396.61 18273.58 289965.89  243.40  528703.84 466518.75 405093.10 61425.65  Figure 11: Dzerzhinsky credit statement 2004  Page 2:  70  Figure 12: Dzerzhinsky credit statement 2004 (2)  Page 2:  71  the gin as added security against the loan advanced, and $18,273.58 was the interest charge over the period, leading to a net principal amount for interest purposes of $448,245.20.65 While the charge of 4.6% was well below market rates for loans, it was positive. Figure (12), is a copy of an original document updating the financial position of the collective on July 1st, 2005 and indicates that the gin further bought from the collective, wheat and cotton seeds, in the following quantities and amounts:  Item 1. Wheat seed 2. Cotton seed  Unit Tonne Tonne  Quantity 2.50 40.0  Price 197.91 213.58  Amount 494.77 8543.36  Returning to the efficiency conditions detailed in (3.5), and rewriting these conditions yields the following: Zx = Zy    Px − ρw̃ Py    ∂X ∂Nx ∂Y ∂Ny  (3.6)  The right hand side numerator in (3.6) represents the value of the marginal benefit from fertilizer use in cotton production while the denominator represents the value of the marginal product from fertilizer use in grain production. The ratio of land in cotton to grain is taken as the actual amount of land in cotton (807 ha) to the remaining land area (411 ha) which yields 1.96. Taking a (revenue) weighted average price for cotton and cotton seed sold66 , Px is placed at $242.64 and the price for wheat is taken directly from the invoice as 197.91. The piece rate wage is taken at $44 per ton from the invoice and is confirmed on the basis of prevailing rates during the study which were 14.5 dhirams per kilogramme.67 Plugging these values into (3.6) with a calculated interest rate of 4.6% gives 1.96 = M Px /M Py . Using the prevailing figures, the ratio of net benefit to price of grain equals 1, and thus, the ratio of land in cotton to grain equals the ratio of marginal product from fertilizer use in cotton to that of grain. Two immediate implications arise from this exercise. First, is that governmental edicts on minimum land allocations for cotton production will define the input use per acre. When the allocation of land to cotton is increased, input use of fertilizer per unit of land will decrease for cotton and/or fertilizer input per unit of land in grain will increase. The total amount 65  Item number 14 ’Finance charges’ in the amount of $9815.25 has proven to be difficult to explain. It is surmised that the item represents payments to the directors of the collective, but has been booked as charges related to the advancement of credit. I have not entered this charge as an interest cost in arriving at an estimate for the annual interest rate of 4.6%. 66 It is common practice for the gins to retain a certain portion of cotton seed recovered in the ginning process as in kind recovery for ginning costs. These seeds are then provided by the gin as in kind credit to the farms in the following cropping year if the farm is deficient in its stock of seed for the following year. 67 The exchange rate during the period of study was 2.9 Tajik somoni per $US. A dhiram is 1/100 of a Tajik somoni.  72  of fertilizer used (summed over utilization in grain and cotton production) may or may not increase and the direction is ambiguous. Second, is the calculation that the marginal product of fertilizer in cotton production is almost two times the marginal product of fertilizer in grain production. The only study uncovered and available in the public domain which calculates the marginal product for fertilizer in Tajik cotton production is that of Tashrifov (2005) who estimated it at 1.64 for the period 1992 to 2002. There is very good reason to believe that the current figure may not be as high and one significant reason for such a high estimate is that the republic was gripped in a bloody civil war between 1992 and 1996 resulting in significant loss in fertilizer imports and production capacity within the republic. Intuition would also suggest that if the marginal product for fertilizer in cotton production is “high”, the gin would find a means of increasing the allocation of fertilizer to cotton production, as increased fertilizer use per unit of land would result in greater income to the gins from both interest earnings and from the sale of cotton fibre on the world market. Any magnitude over 1.0 would provide an incentive for the gin to push for greater allocation of fertilizer per unit of land in cotton. The claims that the gins exert monopsony power in the purchase of raw cotton, that inputs are provided to the farm at inflated prices, and that there are delays in the provision of material inputs to the farm, would not seem to hold unless the collective is constrained in a manner other than on the basis of minimum land requirements for cotton production. When the marginal products of fertilizer for grain and cotton production are both increasing, it must be the case that fertilizer is in short supply, thereby restricting the efficient use of fertilizer on farm. The World Bank, ADB and others have claimed that farmers are often shorted on their deliveries of critical material inputs during the growing season; however, a deliberate shortage in the supply of material inputs, on the part of the gin, does nothing for the gin in the face of land usage restrictions. Without an upward sloping supply curve, the gin has little to gain from deliberately shorting material inputs in an environment where there are increasing returns in the application of fertilizer. The claims which have been launched against the gins and relating to monopoly power or shortages in the delivery of inputs would seem to require some form of constraint (or appropriate incentive) that would bind the collective to producing a certain desired quantity of cotton that allows the gin to meet its contractual obligations with international purchasers of cotton fibre. My immediate interest is upon an analysis of the underlying decisions and choices related to the amount of fertilizer applied in the production of cotton. More specifically, my interest is in analyzing a situation wherein the collective maximizes its revenue from grain production as well as the private interest of its members, and wherein the production of cotton is a means for obtaining financing to facilitate the production of grain. Lowering the amount of inputs into cotton production reduces the gins profit from loan provision, reduces the margin on processing raw cotton, and reduces the gins revenue from the sale of cotton fibre. In confronting this  73  possibility, the gin could exercise monopsony power in the purchase of raw cotton from the collective, a claim that is frequently lodged against the gins in Tajikistan. Yet, the collective farm members have an outside option, and that is to dissolve their membership in the collective and produce grain on privately held lands. From this perspective, monopsony conditions are only likely to be present when the gin can “neutralize” the outside option for private farming by equalizing private returns with individual returns from collective farming. Begin with the proposition that collective farm members have an incentive to stay within the collective so long as the benefits from private production of grain on privately held lands are equal to the benefits from maintaining a stake in the collective farm. The benefits from remaining on the collective farm are three fold. Firstly, collective farm members are provided with a maintenance wage (w̄) to work on the production of cotton in the preharvest season. Secondly, the members are offered the piece rate wage (w̃) during the cotton harvest season.68 Both of these are of direct benefit to the individual members, whereas the cost of the loan, which funded the availability of these wages, are booked to the account of the collective farm as an entity. Third, the collective produces grain (and other produce or commodities) on a portion of the land which is not devoted to cotton production. In the case of the latter, it is the local governmental organs (hukumats) which mandate how much land is to be placed under cotton production. Moreover, the gins do not stipulate a minimum delivery of raw cotton, at harvest, as part of the loan agreement. Cotton yield per unit of land is therefore of primary importance in the settlement of accounts from the loan advanced by the gin to the collective. Returning back to the first order conditions acquired from the farm’s profit maximization problem,  ∂Y Py ∂N = (1 + l)Pf Zy y  and  ∂X (Px − (1 + l)w̃) ∂N = (1 + l)Pf Zx x  these expressions can be expressed in the familiar p = M C form by dividing through by the marginal products, in order to obtain69 , P Z  Py = (1 + l) ∂Y f/∂Ny y and Px = (1 + l)    Pf Zx ∂X/∂Nx  + w̃    Alternatively, and more formally, by recalling the production function constraints, X = Zx (Nx )α and Y = Zy (Ny )β , the farm cost function can be written as,  C(X, Y ; Zx , Zy , Pf ) = Pf Zx  X Zx  1  68    α  + Pf Zy  Y Zy  1  β  + w̄Fi + w̃X  (3.7)  It is also widely recognized that the cotton stalks at the end of the harvest are also of economic benefit to the collective members and used as fuel in an environment where electricity and the supply of gas are both erratic and of poor quality. 69 I am indebted to Hugh Neary for picking up an error in the oral examination draft and for providing a very helpful set of notes which have been used extensively within this chapter.  74  and on this basis, the farm profit function, as a function of outputs X,Y: ΠF = Py Y + Px X − (1 + l)C(X, Y ; Zx , Zy , Pf ) First order maximization conditions with respect to X and Y are given by,   Pf ∂C(·) Py = (1 + l) = (1 + l) ∂Y β ∂C(·) Px = (1 + l) = (1 + l) ∂X    Pf α  Y Zy  X Zx   1−β β  (3.8)   1−α  !  α  + w̃  (3.9)  and by using the production functions, it can be shown that that the two expressions for marginal cost which have been derived are equal. More specifically,   Y Zy  Pf α    Pf Zy Pf (1 + l) = (1 + l) ∂Y /∂Ny β   1−β β  and  (1 + l)  Pf Zx + w̃ ∂X/∂Nx   = (1 + l)  X Zx   1−α  !  α  + w̃  Equation 3.9 is the supply price for output that must be paid to the collective farm in order to produce cotton (X), given the values of other parameters. Faced with an output-dependent supply price, the gin has the ability to exercise monopsony power over the farm. The relevant derivatives of the supply price function are given by: ∂Px ∂ 2 C(·) = (1 + l) = (1 + l) ∂X ∂X 2  Pf 1 − α α α  X Zx  ∂Px ∂ 2 C(·) = (1 + l) = (1 + l) ∂Zx ∂X∂Zx  Pf 1 − α α α  X Zx  1−2α α  1−2α α  1 Zx  !!  −X Zx2  >0  (3.10)  !! <0  ∂Px ∂C(·) = >0 ∂l ∂X  (3.11)  (3.12)  Equation (3.9) can also be inverted in order to derive a supply curve for the output of raw cotton. This can be written as, s    s  X = X (Px , Pf , Zx , w̃, l) = Zx  α(Px − (1 + l)w̃) (1 + l)Pf    α 1−α  (3.13)  and provides the gin (or local governments) the ability to manipulate supply, through the choice of Px (or X), Pf , Zx , w̃, or l, and in order to transfer surplus from the farm to the gin. While the model may be overdetermined in this sense, the literature on franchising (see for example,  75  Mathewson and Winter (1984), Mathewson and Winter (1985)) would suggest that not all five of these instruments may be needed to transfer the surplus. Given values for these five instruments, at the disposal of the gin (and the local government, in the case of Zx and Zy ), the maximized farm profit function can be written as, πf∗ = πf∗ (Px , Py , Pf , Zx , w̃, l, w̄) One important feature of this function is that profitability of the farm depends upon w̄, a fixed cost, whereas cotton supply does not. The gin has an option, therefore, to manipulate the level of the profit of the farm, and therefore farm income, in order to ensure that each individual farmer’s income is never lower (or higher) than the outside option.70  3.3.2 Gin: Part 1 Assuming that the number of farms is fixed, and utilizing a simplifying assumption that X s (Px , Pf , Zx , l) is the sectoral supply of raw cotton to the gin, rather than individual farm supply, the gin will attempt to maximize,  πg = Pw θX s (·) − Px X s (·) + lC(X s (·), Y s (·); Zx , Zy , Pf ) s.t. Zx + Zy ≤ 1  (3.14)  where Y s = Y s (Py , Pf , Zy , l) is the supply function for grain and a constraint reflecting the scarcity of land Zx + Zy ≤ 1. θ represents the implicit cost of processing raw cotton into its valuable components of cotton fibre (or lint) and other joint products (cotton seed, cottonseed oil, cotton meal).71 Alternatively, the gin’s profit maximization problem can also be formulated by choosing X, l and Zx and maximizing, Πg = Pw θX − Px (·)X + lC(X, Y ; Zx , Zy , Pf )  s.t.  Zx + Zy ≤ 1  (3.15)  where Px (·) is the supply price of raw (farm) cotton output.  3.3.3 Gin-farm joint profit maximization Prior to analyzing actual behaviour of the gin, it may be useful to understand what allocation of resources would be required in order for the system to be efficient. For simplicity, leave aside 70  The manipulation of w̄ can also be used to affect the number of farms in the sector but this goes beyond the scope of this model. I thank Hugh Neary for raising this important point in his helpful set of notes. 71 If the farm delivers X units of raw cotton to the gin and the gin turns out λX units of cotton fibre at a unit cost of γ, the net product is (λ − γ)X = θX) which is sold by the gin at the world price Pw .  76  any benefits that accrue to the local governments, and denote the joint profit function for the gin-farm system as, Πf + Πg = Pw θX + Py Y − C(X, Y ; Zx , Zy , Pf )  (3.16)  With land allocations for cotton and grain fixed (through local government edicts), the first order conditions for joint profit maximization, with respect to cotton (X) and grain (Y) outputs are given by, Pw θ −  ∂C =0 ∂X  and  Py −  ∂C =0 ∂Y  (3.17)  The first order conditions in 3.17 are identical to the farm’s profit maximizing first order conditions given in 3.8 and 3.9, when Px = θPw for cotton, and when the rate of interest on credit advanced l equals 0. But while joint profits would be maximized under such a scenario, all of the profits would accrue to the farm and none to the gin. So what are the options, if efficiency is desired? One clear way is for the gin to extend credit on the condition that payment of the loan include a fixed fee charge (a lump sum payment) and this would allow joint profits to be shared between the farm and gin. However, the evidence presented through Dzherzhinsky farm records would suggest that this option has not been utilized and, indeed, it is rare for such contracts to exist in the real world. But what if land inputs were not fixed and the collective farm had a choice in the allocation of land between cotton and grain. What would be the optimal levels of land input, holding output levels of cotton and grain fixed, and in order to attain maximum joint profit. First order conditions from a maximization of 3.16 with respect to land inputs Zx and Zy , given the land constraint (Zx + Zy = 1) are: −  ∂C −µ=0 ∂Zx  and  −  ∂C −µ=0 ∂Zy  (3.18)  ∂C ∂C For each incremental increase in land allocated to X or Y, the marginal benefit ( ∂Z and ∂Z ) is x y  measured by a reduction in variable costs associated with that increase for a given level of output X and Y. This marginal benefit of cost reduction is equated to the marginal (opportunity) cost of using an extra unit of land, given by the land constraint multiplier µ. In the context being analyzed, the collective farm has no choice in the allocation of land between cotton and grain, as it is dictated by the local government, but in a competitive environment, the farm would choose to allocate land (even rent in or out extra land) according to these cost minimizing rules. To complete the understanding of joint profit maximizing rules, allow the farm to choose its land allocation (Zx and Zy ) while recognizing that output supplies X s and Y s vary with land  77  inputs. Maximizing 3.16 with respect to (Zx and Zy ) gives,     ∂C ∂X s ∂C Pw θ − − +µ =0 ∂X ∂Zx ∂Zx  (3.19)      ∂C ∂Y s ∂C Py − − +µ =0 ∂Y ∂Zy ∂Zy  (3.20)  It has been shown above that each parenthetical expression in 3.19 and 3.20 must equal zero in order for joint profit maximization. For profits to be shared efficiently between the farm and the gin, the gin should pay the farm a price for cotton, Px = θPw , charge an agreed upon lump sum fee for credit offered to the whole farm operation, and then allow the farm to rent land in or out, at price µ, and purchase fertilizer (inputs) at price Pf - both in competitive markets. Why the gin adopts a contracting system that departs from this efficient outcome is ascribed to reasons of political economy.  3.3.4 Gin: Part 2 In section 3.2, a hypothetical model was presented to argue why the gin would not be interested in financing wheat production. Continuing with this line of reasoning, particularly that the gin has little interest in Py and Y , and holding the allocation of land fixed, maximization of gin’s profit (given by function 3.14) with respect to X gives, Pw θ − Px − X  ∂Px ∂C +l =0 ∂X ∂X  and from the farm profit maximizing solution, Px = (1 + l)(∂C/∂X) which upon substitution gives, ∂Πg ∂C ∂Px ∂C ∂C ∂Px := Pw θ − (1 + l) −X +l = Pw θ − −X =0 ∂X ∂X ∂X ∂X ∂X ∂X or Pw θ −  ∂C ∂Px =X >0 ∂X ∂X  x The term X ∂P ∂X reflects monopsony power and thus, the gin chooses a level of output, where the  final price that it receives for processed cotton on the world market, is higher than its marginal cost. Surplus is therefore extracted through an inefficiently low price paid to the farm for raw cotton. The low price paid for raw cotton has an impact upon the level of cotton output produced by the farm, and given this low level of output, the first order condition from maximizing 3.14 with respect to Zx is given by,  78  −X  ∂C ∂Px +l −µ=0 ∂Zx ∂Zx  This can be rewritten as,   ∂2C ∂C ∂C − µ = −(1 + l) −X + − ∂Zx ∂X∂Zx ∂Zx  (3.21)  The term within the brackets [·] evaluates to Pf (X/Zx )1/α (1 − α)2 /α2 , which is positive, and thus the expression given in 3.21 is negative. As shown previously, efficiency requires −(∂C/∂Zx ) − µ = 0. That this expression is negative indicates an excessive amount of land placed into cotton production, relative to what cost minimization requires. While the choice of land to be placed under cotton is dictated by the local government organs, likely with input from the territorial gin, the gin has full flexibility in its choice of interest rate l. The first order condition gives, "  1  1 # ∂Πg ∂Px X α α−1 Y β : −X + C(·) = w̄Fi + Pf Zx + Zy ∂l ∂l Zx α Zy  (3.22)  The sign of the expression given in 3.22 is ambiguous and given conflicting effects, l can either be high or zero. The gin has an obvious incentive in raising l given that its revenue base increases on the total amount of the loan advanced. Yet, at the same time that its revenue increases, its cost of procuring raw cotton from the farm also increases, as the supply price for raw cotton is a function of l. When most land is dedicated to cotton, as is the case in the context being analyzed, the marginal cost effect will dominate the revenue effect, therefore resulting in a negative value for the derivative. This would result in an interest rate, l, that approaches zero.  3.3.5 Putting together the pieces In summary, the gin has been modeled as one where it extracts surplus from the farm by paying a “low” price for raw cotton, and through the use of monopsony power. It should be noted here, that an accurate assessment of monopsony power, through the use of comparison between farm gate prices and the world price for cotton fibre is difficult. One particular difficulty relates to inefficient ginning practices. When turnover ratios between raw cotton and cotton fibre is low, average price per unit of raw cotton paid will be lower, and may lead to an assessment of monopsony power that is stronger than actually applied. Similarly, inefficient transportation and picking practices that lead to post-harvest losses can reduce weight and quality, which lead to an over assessment of the level of monopsony power. More specifically, if monopsony power is present, post-harvest losses and inefficient ginning practices will overstate the level of this power, at least in terms of the price paid to the farm. Data analyzed in the following  79  chapter suggests that monopsony practices cannot be rejected, but a full assessment is difficult to undertake due to a lack of hard evidence. The model developed above has also raised an argument that the level of interest charged against loans advanced for cotton production is based upon the which marginal effect dominates - revenue from increased interest income or increased cost from higher procurement costs. The combination of lower product price paid to the farm and high interest rate charges results in a lower than efficient quantity of raw cotton delivered to the gin which pushes the gin (and local governments) to enforce a greater than efficient quantity of land to be allocated to cotton production. While not explicitly modeled, one implicit reaction to the edicts on land use is for diversion of inputs out of cotton production and into non-cotton production, thereby leading to a further decline in expected output. It was postulated early on that individual farmers within the collective were indifferent between producing cotton at a collective loss and producing grains profitably on private lands. Wages play an important role in this regard, and as the model indicates, the maintenance wage is one lever through which the gin can manipulate the delivery of raw cotton, through appropriate incentives for individual (collective) farmers to produce cotton. In testing these assertions, refer back to table 9 of section 3.3.1. First, in kind delivery of fertilizer was in the amount of 499.55 tonnes (items 6, 7, and 8). Second, the cash payment from fertilizer under category 14, valued at an average price per tonne of $159.13, bought an additional 155.69 tonnes of fertilizer. In total, therefore, 655.24 tonnes of fertilizer were available through crop financing. Table (10) is a summary of data obtained from Dzerzhinsky collective farm and which relates to the crop financing statements in Figures 11 and 12. Table 10 shows that brigades were provided a total of 774.57 tonnes of fertilizer which were applied to cotton and other crops. The difference of 119.33 tonnes of fertilizer represents the quantity of fertilizer stored by the collective in the previous cropping season for utilization in the following season’s sowing of spring wheat. This was confirmed in conversations with the staff of Dzerzhinsky collective farm. The definition of diversion used in this study is precisely within this context of material inputs being charged to the collective’s account of cotton production but utilized for crops that are not collateralized by the credit lending gin. As discussed in the background section, recommended norms for fertilizer usage in Tajikistan are on the order of 200kg/ha for nitrogen, 60kg/ha for phosphorous and 60kg/ha for potassium based fertilizers. The 655.24 tonnes of fertilizer obtained on credit, and on the basis of 811 hectares placed under cotton, is clearly well above the prescribed norms for cotton production. Moreover, national statistics would suggest that cotton entities in the republic were averaging 150kg/ha of mineral fertilizers in cotton production during the year 2004. Double checking  80  Brigade Soliev Najmidinow Salimov Abduramonov Tashripov Zaripov Musorfirov Asoev Boboev Nasirov Sodikov Yorova Murodov Akmedov Saedov  Table 10: Dzerzhinsky brigade statistics # Fertilizer Cotton Hectares Yield Workers (tons) (tonnes) (cotton) tons/ha 16 52.4 85.80 55 1.56 34 40.2 77.04 36 2.14 96 66.2 142.56 66 2.16 61 49.0 95.00 50 1.90 56 54.3 133.28 56 2.38 43 49.4 99.11 53 1.87 75 53.2 114.40 52 2.20 75 57.2 126.72 64 1.98 62 62.5 166.80 60 2.78 32 49.4 129.32 53 2.44 54 50.6 124.20 60 2.07 43 41.5 92.92 46 2.02 47 60.3 122.40 60 2.04 58 36.1 86.80 31 2.80 24 52.2 69.69 69 1.01 776 774.57 1666.04 811 2.09  Fertilizer/ Cotton 0.61 0.52 0.46 0.52 0.41 0.50 0.47 0.45 0.37 0.38 0.41 0.45 0.49 0.42 0.75 0.48  figures in order to ensure that national statistics are reasonably accurate and that there are no errors in unit measures, table 11 shows the average amount of fertilizer utilized per unit of raw cotton produced. In comparing the last column in table 10 with the constructed values (A/B) in table 11, it is evident that not all of the fertilizer delivered to the brigades was utilized in the production of cotton and that a significant portion was diverted into the production of other crops. Returning to the study of Tashrifov (2005) mentioned earlier in this paper, the author estimated the marginal product of fertilizer to be on the order of 1.64 and the output elasticity of fertilizer to be on the order of 0.85. These figures would suggest that the ratio of fertilizer use in cotton production to total cotton production was approximately 0.50. This is similar to the average for Dzerzhinsky collective farm as noted in table 10 and calculated at 0.48. Tashrifov’s estimates are based upon a panel data set of 342 observations among 34 regions over the years 1992 to 2002. While there may be skepticism over the integrity of the data provided by the state statistical agency for Tajikistan, there is little reason to believe that the agency would consistently underreport fertilizer usage in cotton production. Anecdotal evidence and conversations with industry observers would also suggest that the figures reported by the state agency are reasonably accurate. One conclusion that we may make is that Tashrifov’s data was representative of the total amount of fertilizer used on farm, obtained on credit from the gins, and not actual application of fertilizer in cotton production. The marginal product of fertilizer use in cotton production may be significantly higher than the 1.64 estimated by Tashrifov and would certainly fall in line with the argument of diversion made in this paper. In the unconstrained case, 81  Table 11: National cotton and fertilizer 1998 1999 2000 2001 2002 (A) Cotton (000 tons) Total 384 313 336 453 516 Sughd 129 115 121 145 155 Khatlon 218 172 178 255 306 RRS 37 26 37 52 55 (B) Fertilizer (000 tons) Total 34.8 16.3 24 18.6 38.7 Sughd 13.7 7.4 7.8 8.7 13.8 Khatlon 17.2 7.0 10.9 6.9 20.4 RRS 3.9 1.9 5.3 3 4.5 (B)/(A) Total 0.09 0.05 0.07 0.04 0.08 Sughd 0.11 0.06 0.06 0.06 0.09 Khatlon 0.08 0.04 0.06 0.03 0.07 RRS 0.10 0.07 0.15 0.06 0.08  statistics, Tajikistan 2003 2004 2005 2006  2007  537 149 334 55  557 159 334 63  448 156 241 52  438 131 258 48  420 125 260 35  43.2 17.2 21.2 5.7  49.2 19 24.5 5.7  55 16.3 34.5 4.7  40.9 14.6 21.2 5  35.2 11.9 20.6 2.7  0.08 0.12 0.06 0.10  0.09 0.12 0.07 0.09  0.12 0.10 0.14 0.09  0.09 0.11 0.08 0.10  0.08 0.10 0.08 0.08  the assertion made was that when the marginal product of fertilizer in cotton production was greater than 1, the gin would seem to have every incentive to increase the amount of fertilizer use in cotton production given extra revenue from credit advanced and from the extra revenue obtained from ginned cotton. Are there plausible reasons in the constrained case for why the gin would accept to have a relatively “high” marginal product for fertilizer use in cotton production? In the first instance, a look at private member benefits within the collective may be constructive in answering this question. Turning back to the crop financing record for Dzerzhinsky farm in 2004, total wages (harvest and maintenance) were $212, 734.78. With 776 workers in the collective, the average wage earned during the cotton crop year was $274 while the average debt per worker was $71.72 A detailed record of the 2003 crop financing is not available but from the summary in table 9, we can surmise that the situation was more favourable for the farm workers in 2003 as the crop financing was roughly on the same order. Given that the total volume of cotton sold was roughly equal, the harvest wages would have been comparable, as would have the maintenance wages. That the situation would have been more favourable is based upon the assertion that there was no debt per worker and that a worker wishing to leave the collective and take her portion of land with her could have done so without any debt being ascribed or prorated. As profits and debts are rolled over from season to season, the workers were not accorded with the benefit of a dividend or sharing of the profit in the 2003 crop year. From an individual member perspective, therefore, it is the benefits from wages as well as the 72 The average debt per worker is based upon the debt value of $54690.35 obtained after the sale of wheat and cotton seed.  82  side benefits that diversion of material inputs provide which are the incentives for remaining within the collective and the benefits are very much in relation to the alternative occupation of producing grain privately. Table 12 is an accounting estimate of the margin obtained by private grain farmers during the 2004 crop year.73 The price for wheat in the estimate of private margins is that obtained by Dzerzhinsky collective in its sale of wheat to the gin. Under the constrained imposed in the model within this study, the benefit between wages per member ($274) and private grain profits ($351) is the minimum value of the benefits from non-cotton production per member within the collective that provides the incentive to stay within the collective. Any amount less than this value would make private grain production more lucrative and profitable.  Wheat Straw  Table 12: Private grain production margins Yield Unit Price Unit 3 tons/ha 197.91 $ /ton 1.5 tons/ha 33 $ /ton Total income  Cultivation Seeds Fertilizers Pesticide(Tilt) Irrigation Harvest Transportation Labor - permanent Labor - casual Tax (Unified tax system)  Amount 3 0.2 0.3 0.5 4000  Unit times tones/ha tones/ha lit/ha m3  10 15 50 1  km $/pp / mth hours/ha year  Price 14 110 170 12 0.004 20 1 0.3 0.3 50  Unit $/time $ per ton $ per ton $ per time $ per m3 $ per ha $ per km person/ha $ per hour $ per/ha Total expenses Gross margin  USD 593.73 49.5 643.23 USD 42 22 51 12 16 20 10 54 15 50 292 351  Consistent and reliable data on grain production and other non-cotton crop production was difficult to obtain from Dzerzhinsky collective for a number of reasons. Various discussions were held with members of the brigades who were reluctant to divulge the amounts of wheat and vegetable crops that were produced. Farm records on non-cotton crop production at the office of the collective’s accountant were also incomplete and in sharp contrast to the meticulous recording of cotton production records. In the absence of reliable data on non-cotton crops, one way to look at the $77 per member difference between private grain profitability per hectare in grain production and wage benefits is in the following manner. At an average cost of $159.14 per ton of fertilizer, the $77 per member translates to 0.48 tonnes of fertilizer per member.74 73 74  Source: Random field visits and from International Finance Corporation office in Khujand. The calculated figure of $77 per member is very close to the per debt figure of $71 per member. Without  83  0.48 tonnes per member and 776 members in the cooperative is equal to 373 tonnes of fertilizer on a collective basis. Excluding the stock of fertilizer brought in from the previous crop year (2003), 655 tonnes of fertilizer ordered on credit in the 2004 crop year, less 373 tonnes, provides an estimated amount of cotton utilized in cotton production (under the conjecture of the constraint imposed in the model) of 282 tonnes. With 811 hectares placed under cotton, this yields 350 kilogrammes of fertilizer per hectare of cotton. The norms for fertilizer use for cotton production in Tajikistan of 200kg/ha for nitrogen, 60kg/ha for phosphorous and 60kg/ha for potassium based fertilizers would seem to be met under this scenario. Sale of the 373 tonnes in the local market would be enough to meet the constraint but there would be much greater gain per member if they utilized the fertilizer in the production of non-cotton crops. The question of why national statistics report fertilizer usage in cotton production at well below the prescribed norms looms large given the above analysis. Several possible explanations can be advanced in this regard. Firstly, the quality of labour on farm has not been included and quality in the judicious sowing of seed, application of fertilizer and in the manner in which cotton is picked has a great impact upon yield and quality of raw cotton. Delays within any of these stages could result in significant loss of production, thereby reducing the amount of piece rate wages during the harvest season. With a cadre of 776 members, Dzerzhinsky farm may very well be quite efficient in its application of labour at critical times and perhaps more efficient (in labour allocation) than other cotton growing collectives within the republic. Second, the state of equipment at Dzerzhinsky would seem to be in much better shape than other collectives visited and the quality of land preparation plays an equally large role in production and in yields of cotton produced. A gin who agrees to advance funds for fuel, lubricants and spare parts is rational and does so with the expectation that the use of equipment for appropriate land preparation leads to greater benefits at the margin for the gin. Third, the quality of land and the extent of land degradation and soil salinization would significantly impact the efficient amounts of fertilizer that could be utilized in cotton production and the marginal yields that would be obtained therefrom. Lastly, substitution of organic manure could also be driving the results indicated by national statistics. On the basis of survey data, The Centre for International Studies and Co-Operation (2006) estimate that the average rate of organic manure applied per hectare of land for various farm types as 412kg(Collective), 960kg(Family) and 1794kg(Individual).75 Judicious and efficient use of other inputs that partially substitute for fertilizer could be one broad explanation for fertilizer use below applicable norms. Diversion of fertilizer use out of cotton production and into private production of grains and other food crops is another explanation. any strong evidence, we can only claim that this is fortuitous and not a deliberate calculation on the part of the gin. Having said that, a calculated attempt by the gin to equalize these figures would support many of the arguments made herein. 75 Collective farms have not undergone the land privatization process. Family farms are private farms which are managed and operated by a group of families. Individual farms are individual family households.  84  The model presented in this chapter suggests that l would be lower or close to zero when there is a significantly high allocation of land in cotton production. The calculated figure of 4.6% per annum for Dzerzhinsky farm during the 2004 crop financing year was well below prevailing rates of 32% per annum at legally sanctioned lending agencies. Field surveys also revealed that the price of fertilizer and other inputs were not significantly higher than market prices. In fact, in some cases, the price provided to the collectives was below market prices.76 A review of the crop financing summary for 2003 as shown in table (9) reveals that the calculated value for l is on the order of 2% for 2003. A fall in the price of world cotton, which led to a fall in the farm gate price for raw cotton, saw a fall in l from 4.6% to 2%. The fall in the farm gate price between 2003 and 2004 as stated in the official invoices from the gin to Dzerzhinsky collective was 24%. According to data provided by the International Finance Corporation office in Khudjand, the A Index for cotton fibre during the period fell approximately 35% as shown in figure (13). Figure 13: A index for cotton fibre (cents/lb)  That the gin did not put the full burden of the fall in the world price upon the collective 76  An analysis of the markup on inputs provided to the collective by the gin is made extremely difficult given the incidence of smuggled fertilizer and other agricultural inputs from Uzbekistan where fertilizer during the cotton season in Uzbekistan is heavily subsidized.  85  supports the contention that the interest rate l may be one lever that the gin uses to steer its way towards maintaining its markup margin in the face of fluctuating world prices and input diversion by the collective. It does not seem coincidental that the calculated value of 2% for l in 2003 is equal to the discount provided to the purveyor of cotton for prepayment against the invoice. This is also very much in line with the argument made that the prepayment discount can be viewed as the cost of capital for the gin. In the case where the spot price on the day of delivery is lower than the contracted price, the gin loses nothing on the transaction, as all costs related to delivery, taxes, and insurance have been accounted for and paid by the farm. The guaranteed price for the gin is the contracted price and not the administrative price set domestically. What the gin does gain in a case where there is no benefit to arbitrage, is a margin to cover its ginning costs, and any profit that it retains from the provision of this service. One implicit underpinning of the discussion thus far is based upon the notion that there is differential in yields of grain and other crops between private farms and the collective and more specifically that private yields are in excess of collective yields for non-cotton crops. The proposition that individual collective farmers divert inputs out of cotton production and into non-cotton production did not entail an assumption that grain yields would be higher on collective farms. On the contrary, lower grain yields on collective farms were to be made up through access to wage incomes and subsidized inputs into the production of grain. The analysis of Dzerzhinsky farm invoices has shown that this proposition cannot be rejected. Figure (8) previously showed that yields on private enterprises have exceeded yields on collective enterprises since the end of the civil war that ensued shortly after independence. Given positive and strong correlation of movement between yields on various enterprises, the difference can only be ascribed to lower fertilizer levels. Varying genetic vigour of wheat seeds and varieties would not provide yield measures that are relatively consistent in terms of the divergence between yields over the period analyzed. There is no publicly available information detailing historical use of fertilizer consumption in wheat production but the trends in yields between private and collective enterprises does not negate the notion that wages earned in the production of cotton and diversion of inputs into private crop production are two factors which contribute towards the persistence of collective farm structures in post-Soviet Tajikistan. In the next chapter, I test this notion through an analysis of data from 135 farms. On the basis of this analysis, I will argue that collective type farmers, producing mainly cotton, may be no worse off in terms of standards of living than private farmers producing primarily grains and food crops. The diversion of fertilizer and other material inputs leading to subsidized production of non-cotton crops is one important cog in this argument. The payment of wages for the production of cotton is equally important and helps to equalize returns between individual collective farm member benefits and private grain farmer benefits.  86  Chapter 4  Relative comparison across farm types Numerous formal visits to the state statistical agency (Goscomstat) yielded sparse information on cotton and other agricultural statistics, and largely due to a disorganized system of publications management. Heightened sensitivity for any information on the cotton sector also played a role. Much of what was obtained resulted from diligent work of an employee within the agency who scoured unorganized cabinets and rooms for published material. Some of these data were used in the preceding chapter to describe the cotton sector and the general economy in Tajikistan. Some were also used in justifying a conjecture that a portion of fertilizer obtained for the purpose of cultivating cotton may actually be diverted into the production of grain and other crops. In the case of the latter, an analysis of national data on fertilizer use in cotton production suggested that fertilizer earmarked for cotton production may actually be a better predictor of wheat production than of cotton production. This chapter extends the argument regarding fertilizer diversion and a comparative analysis of independent farmers versus collective type farmers is undertaken. The objective in this chapter is to test the conjecture that collective type farmers, producing mainly cotton, may be no worse off, in terms of standards of living, relative to private farmers producing primarily grains and food crops. The diversion of fertilizer and other material inputs is one important cog in this argument, as is the earning of wages in the production and picking of cotton. Data on 135 farms, collected for an FAO study conducted by Caccavale (2005), have been provided for use by the author, and are presented in table 23 at the end of this chapter. Caccavale (2005) utilized principal component analysis (PCA) techniques to look at the effect that various farm sizes have upon productivity in Tajikistan. Caccavale (2005) concluded that for farmers producing wheat and other non-cotton crops, “small” farms between 2-15 hectares had higher margins compared to relatively larger farms and that credit and limited market opportunities were constraining factors.77 For cotton farms, no clear pattern emerged and the results were very much mixed. The author did not use the data for any other analysis or purpose, but the following sections extend Caccavale (2005) study with an aim of trying to pin down some of the conjectures made earlier in this study. While the data does not formally allow the testing of market power on the part of the gin, there are indications which suggest that monopsonistic power exercised by the gin cannot be dismissed. 77  The terminology and definition relating to “small” and “large” are directly taken from Caccavale (2005).  87  Of the data utilized, 79 farms are classified as wheat growing farms (type 1), 14 as cotton growing farms (type 2), and 35 farms are classified as mixed (type 3).78 Data on each farm collected is detailed in table 13. Table 13: Variable names for empirical data set DF ha sh cr type wha wt wy cha ct cy income exp debr debfc  Farm number Total land area in hectares Total number of shareholders Total number of crops under cultivation Farm classification (1=wheat, 2=cotton, 3=mixed) Hectares placed under wheat Tons of wheat harvested Wheat yield (tons/hectare) Hectares placed under cotton Tons of cotton harvested Cotton yield (tons/hectare) Total revenue from farm operations Expenditures Accounts Receivable Debts to the credit lending agency (gin)  Thus far, the discussion has been solely upon the existence of two categories of farms: collective type farms which produce primarily cotton and private farms which produce grain and food products. The sharp distinction has mainly been for expositional purposes and for simplicity. In reality, the distinction made in this section between “cotton”, “wheat” and “mixed” farms is more realistic and represents the opportunities that exist for farmers given limited (or thin) access to credit, governmental edicts, and markets for products both domestically and internationally. “Wheat” farms are characterized by either privately run farms or an association of families who farm multiple plots of land jointly. “Cotton” and “Mixed” farms are characterized by collective type production. In both cases, land may have already been subdivided between farmers (shareholders in the terminology of Caccavale (2005)) but the process of final privatization has not taken place. Table 14 summarizes the average amount of “debt” per shareholder among the three farm categories.79 The data has been generated for only those farms where the number of sharehold78 The FAO data set contained 145 farms but 17 of these samples were deleted due to missing pieces of data. The classification of type 1,2,3 farms are my own, as is the notation for type of farms. The remainder of the notation for variables is relatively consistent with Caccavale (2005) with minor alterations. √ 79 The 95% confidence interval has been calculated conventionally as, SE=(Standard Deviation)/( n) where n is the sample size and 95% CI = Mean +/- (1.96 * SE)  88  ers is greater than zero and where there is a positive level of “debt”.80 Of immediate note is the observation that only one of the 79 farms classified as “wheat” report any accrued “debt” and that even within this one observation, the level of accrued “debt” is relatively minimal.81  Table 14: Debt per shareholder among various farm types Farm type 1: Wheat 2: Cotton 3: Mixed  # 1 6 18  Mean 23.10 1435.11 1477.41  Std. Dev. 1689.26 2169.07  Min. 23.10 165.50 23.42  Max. 23.10 4693.15 9392.61  Median 23.10 814.45 884.06  95% Conf. Interval n/a 83.42 - 2786.80 475.35 - 2479.47  Approximately 50% of the farms in the category of “cotton” and “mixed”, in which land has been divided but not privatized, reported accrued “debt”, with no statistical difference in the mean levels of “debt” per shareholder between the two categories. Similarly, retained earnings between the two farm types are also not significantly different. Taking data on each farm type relating to accrued “debt”, revenue, expenses, and accounts receivable, “retained” earnings per shareholder are shown in table 15.82  Table 15: Retained earnings per shareholder among various farm types  Farm type 1: Wheat 2: Cotton 3: Mixed  # 79 14 30  Mean 265.67 (475.46) (466.80)  Std. Dev. 385.44 1307.07 1469.92  Min. (400.00) (4468.85) (6482.68)  Max. 1746.15 1107.14 2434.52  Median 128.57 (49.30) (239.66)  95% Conf. Interval 180.67 - 350.67 (1160.15) - 209.23 (992.80) - 59.20  While mean levels of retained earnings per shareholder are not significantly different between cotton and mixed farms, median and max levels would suggest that mixed farms offer greater potential for profitability, but with more downside risk. How? Given that incomes are most probably understated for mixed farms, as a portion of the grain is consumed by the households and not sold, the downside risk is likely related to the production of cotton. For cotton producing entities, retained earnings per shareholder are essentially nil. The maximum of 1107 Tajik 80  Where there is no shareholder, this is an indication that it is still a government run institution reminiscent of the form Sovkhoze system. As in previous chapters, I denote the term “debt” within quotation marks in order to highlight the argument that “debt” in the production of cotton, is not consistent with a standard definition for debt. 81 The reported “debt” levels are in Tajik somoni which at the time of data collection would have been around 3 Tajik somoni per $us. It is likely that this is “debt” obtained from traditional money lenders or from microcredit agencies who place ceilings that are typically too low for cotton producing entities. 82 Retained earnings are defined as (revenue - expenses -“debt” + receivables.) Figures within brackets indicate negative values.  89  somoni (approximately $300) and the minimum of -4468.85 are outliers which, when accounted for, bring the mean and median levels closer to zero.83 These figures raise an interesting point and that is that pure cotton producing entities are likely not the holders of what is being termed “debt” in Tajikistan. Indeed, the statement supports the current arrangements for financing cotton production and the conjecture related to siphoning or diversion of inputs out of cotton production on mixed farms. When all financing is provided for cotton, collateral is provided with the forthcoming harvest for cotton, and when only cotton is produced, there is no incentive for the gin to foster the perpetuation of “debt” (intentionally or as a result of exercising market power). It is the mixed farms that are likely the holders of “debt” and the amount of land that is potentially available for cotton production within mixed farms is likely of strategic importance to the gins. This is due to the relatively larger amounts of land under “mixed” farming. The mean number of shareholders within the “cotton” category of farms is 440 while, within “mixed” farms, the mean is calculated at 447. Table 16 summarizes the average landholding per farm member among the three types of farms. Table 16: Land per shareholder among various farm types Farm type 1: Wheat 2: Cotton 3: Mixed  # 79 14 30  Mean 2.29 0.51 1.26  Std. Dev. 3.71 0.14 0.99  Min. 0.07 0.29 0.17  Max. 24.15 0.75 4.41  Median 0.90 0.46 0.97  95% Conf. Interval 1.47 - 3.11 0.44 - 0.58 0.91 - 1.61  Utilizing the mean values of shareholders per farm and the average land holding per shareholder, we can deduce that, on average, mixed farms are more than double the size of “cotton” farms. Land holdings per shareholder between the various farm types are statistically different at mean levels. The data would also suggest that there is a continuum of farm sizes within each category and, therefore, it cannot be argued that land size per member is a primary motive for disbanding the collective farm. Indeed, the maximum of 24.15 hectares per shareholder on the wheat category is likely a government breeding or testing station and the mean for wheat and mixed farms are generally of the same order. If “mixed” farms are of strategic importance to the gins, and if the gins are willing to extend greater credit to “mixed” farms relative to “cotton” farms, by virtue of financing non-cotton production, cotton yields must play an important factor in the decision. 83  The figure of -6482.68 for mixed farms is also an outlier but does not change the general argument that mean and median amounts of retained earnings for mixed farms are still much more negative relative to cotton only farms and do not approach zero.  90  4.1 Cotton versus mixed production farms Average cotton yields for cotton farms and mixed farms are shown in table 17. While not statistically significant (at 95%), the mean yields of cotton on “mixed” farms and “cotton” farms are close enough to suggest that there is little difference between average productivity on these two types of farms. The argument also holds true in the case of average cotton production by shareholder as shown in table 18. Table 17: Average cotton yields by farm type Farm type 2: Cotton 3: Mixed  # 14 35  Mean 2.14 1.89  Std. Dev. 0.46 0.61  Min. 1.39 0.69  Max. 3.00 2.86  Median 2.14 2.00  95% Conf. Interval 1.90 - 2.38 1.69 - 2.09  Table 18: Cotton production per shareholder by farm type Farm type 2: Cotton 3: Mixed  # 14 35  Mean 1.00 1.20  Std. Dev. 0.38 0.70  Min. 0.56 0.01  Max. 1.79 3.12  Median 1.02 1.03  95% Conf. Interval 0.80 - 1.20 0.97 - 1.43  In both cases, the gin would be indifferent to financing the production of cotton for “cotton” only farms and “mixed” farms but the amount of financing extended for “mixed” farms will be greater, given the extra extension of credit needed to finance the production of non-cotton crops. Prices paid for cotton or differential rates charged to farms based upon risk or market power must, therefore, also play an important role, in addition to yields. An analysis of prices paid for raw cotton at the farm gate is, therefore, instructive in assessing whether or not there is evidence to suggest monopsonistic practices on the part of the gin. One drawback of the data is that price information for cotton or grain is not given and only total income and expenses reported. Prices, in this case, need to be imputed manually by dividing total revenue reported by total production of raw cotton harvested. Taking those cotton producing entities in which only cotton is produced (14 observations) yields prices for cotton received by each “cotton” farm as shown in table 19. The figures for price of cotton are constructed from the respective cotton production and income figures. The wide range of prices for cotton can be ascribed to at least two factors. First is the issue of quality for both raw cotton and cotton fibre and the turnover ratio. The constructed prices are indicative of the price for raw cotton sent from the farm gate. Where farms are located at some distance from the gins, poor transportation and handling can affect the quality of raw cotton for ginning. Further delays in processing, poor storage of cotton at the gin, 91  Table 19: Imputed cotton prices DF 25 26 27 52 55 56 58 59 60 61 62 63 64 139  Price($/ton) 413.73 563.60 427.77 786.04 530.70 380.39 179.29 1230.48 657.14 780.80 441.91 816.33 1069.94 900.00  Cotton(tons) 9.5 8.0 36.0 968.40 215.00 311.00 198.00 3150.00 168.00 125.00 210.00 147.00 958.00 12.50  Land under cotton (ha) 4 3 12 578 103 169 83 1722 110 90 85 74 438 5  Total Land (ha) 6.15 5.28 16 578 103 169 83 1722 110 90 85 74 438 5  and inefficient ginning processes can also affect the turnover ratio, which is the percentage of fiber recovered from the raw cotton. Where quality and turnover ratios are poor, prices received on the world market and subsequently revenue will be adjusted downward accordingly. The data does not allow us to investigate the issue of quality and turnover ratio in a detailed manner. However, I am informed that the farms surveyed are within a cluster that is not widely disbursed geographically and that the farms are fairly well compacted within a certain district. It is likely, therefore, that transportation and delays in transport are not important in explaining the variances in prices for raw cotton. It is also likely, but not substantiated, that cotton farms within this compact area are being serviced by one gin, given the manner in which gins are sited, and in view of the discussion provided in the background section. Holding this assumption that only one gin is servicing the 14 farms under observation, turnover ratios of raw cotton to fibre are not going to vary between the farms given that the effects of technology and the efficiency of the ginning plant are going to be applied uniformly to all farms. If these turnover ratios do vary, this would signal an added aspect of market power being exerted and in addition to the observation of differential prices being paid for raw cotton at the farm gate. A conjecture regarding monopsonistic practices in the purchase of cotton cannot be rejected, given evidence of differential prices paid as shown in table 19, with or without the added wrinkle of potential “pencil” grading raw cotton and turnover ratios on an individual farm basis.  4.2 Wheat versus mixed production farms Table 20 summarizes wheat yields between mixed and wheat farms. An analysis of the confidence intervals from table 20 indicates that (within 1 decimal point), there is no significant 92  difference between mean yields for wheat production between wheat farms and mixed production farms.  Table 20: Wheat yields between farm types Farm type 1: Wheat 3: Mixed  # 79 35  Mean 2.13 1.86  Std. Dev. 1.57 0.76  Min. 0.47 0.63  Max. 6.43 3.29  Median 1.35 1.60  95% Conf. Interval 1.95 - 2.30 1.73 - 1.99  Mixed farms, therefore, would seem to be performing at a level where yields for both cotton and wheat are on par with those on cotton only farms and wheat only farms respectively. Looking solely upon yields between farm types is, however, misleading in the sense that it masks the benefits of wheat production to the individual farm member. Table 21 summarizes wheat production per farm member for wheat growing farms and mixed farms. While average yields between wheat and mixed farms are not statistically different, wheat production per member is clearly and significantly higher on wheat farms than on mixed farms, given the sheer impact of increased numbers of shareholders on mixed farms. An average mixed farm, within the sample, can be described as having 447 shareholders with an average amount of 209 hectares of land devoted to wheat (0.51 hectares per member).  Table 21: Wheat production per farm member Farm type 1: Wheat 3: Mixed  # 79 30  Mean 2.99 0.87  Std. Dev. 6.05 1.13  Min. .05 0.07  Max. 45.31 4.17  Median 1.17 0.33  95% Conf. Interval 1.66 - 4.32 0.66 - 1.08  Looking back at table 16, the average amount of land on mixed farms was 1.26 hectares per member, suggesting that over 40% of the land was devoted to non-cotton production, compared to under 30% of the land in cotton only farms devoted to non-cotton production. Wheat farms, by contrast have, on average, a little bit more than 2 hectares per person dedicated to wheat production, almost four times more than the amount of land under wheat per person relative to mixed farms. Where formalized credit for agricultural production is limited or poorly accessible, in conjunction with the gins need to acquire a minimum amount of raw cotton to meet their own contractual obligations with international purveyors of cotton fibre, financing wheat production 93  on its own is likely problematic.84 Even in the case of increased access through microcredit agencies, the ceilings on most microfinance loans far exceed the capital requirements for wheat production on these relatively large tracts of land. Notwithstanding evidence of monopsonistic practices in the purchase of cotton and the tying together of credit and output markets, the gin would seem to provide a valuable service in the provision of financing for non cotton production. Are farmers on “mixed” farms better off by growing cotton? One way to answer this question is to analyze the income earning per shareholder by farm type (wheat, cotton or mixed) with the conjecture that farmers are indifferent between working on large collective farms producing primarily cotton and privately owned land producing wheat and other food crops. Of the 14 farms that have been classified as “cotton”, 11 have devoted all land to cotton production, while 3 have more than 70% of their land in cotton and did not report any production of grains or other products. Those farms classified as “wheat” reported only income from wheat production while “mixed” farms reported income from both cotton and wheat production. Summary statistics on income per shareholder by farm type are provided in table 22.  Table 22: Income per shareholder among various farm types Farm type 1: Wheat 2: Cotton 3: Mixed  # 79 14 30  Mean 642.78 689.47 1081.69  Std. Dev. 726.92 521.72 1025.28  Min. 10.03 181.12 141.66  Max. 4525 1694.06 5866.05  Median 410.53 466.89 886.58  95% Conf. Interval 482.48 - 803.08 416.18 - 962.76 714.80 - 1448.58  Table 22 indicates that there are no significant differences in mean income per shareholder between wheat and cotton farms but that incomes per shareholder on mixed farms, are much higher than other farm categories. However, these dispersions between farm types need to considered carefully. Incomes per shareholder on wheat farms reflect only that portion of revenue earned from the sale of grain (and produce) which is over and above the amount that would have been retained by the household for domestic consumption. In the case of cotton farms, income per shareholder is largely reflective of the sale of cotton with some smaller portion of grain or other produce sold on the market. Some grain or vegetables would also be retained for household consumption within cotton growing entities. Incomes on mixed farms are a composite of the sale of cotton and contribution to income from the sale of residual grain and other food products. On this basis, one is led to ask the question of whether it is just fortuitous that mean incomes between wheat and cotton farms are similar or whether it is by design. 84 Clearly, wheat farmers are obtaining credit from some source and so this statement relates more to understanding why mixed farms continue to grow cotton when cotton is not consistently profitable from year to year.  94  Estimates of wheat consumption per capita in Tajikistan are difficult to pin down, and there are varying figures, but 200 kilogrammes per person per year would seem to be the prevailing norm, based upon discussions with personnel within various international organizations in Tajikistan. Taking 200 as a reasonable estimate, and using a figure of $200 per tonne for wheat85 , the cost of minimal wheat consumption is approximately $40 per member. Piece rate wages, on the basis of an average of 1 tonne per person of cotton produced (c.f. table 18) would have been $44 per person, enough for shareholders on cotton farms to cover the cost of purchasing their minimal wheat consumption. In addition, using the Dzerzhinsky statements, maintenance wages on cotton farms would have been approximately $175 per member. Recalling that mean retained earnings on “wheat” only farms (table 15) were approximately $90 per member, the excess maintenance wages provided security for the purchase of other foodstuff and maintenance items over and above the opportunity cost of income from producing wheat privately. It is likely, therefore, that in terms of standard of living (poor as it may be), there is little difference between an individual member on a “cotton” only collective and an individual wheat farmer. For “mixed” farms, the maintenance wage effectively wipes out the mean retained earnings of -467 Tajik Somoni ($155). In addition, the average wheat production per member of 870kg per member (table 21) provides the basic minimum needs for members in terms of wheat production and the ability to sell surplus production on the market.  85 Data collected for this study and specifically from official documents obtained indicate that the price of wheat during this period was $200 per tonne.  95  Table 23: Sampled dekhan farm statistics  96  DF 1 2 3 4 5 7 8 9 10 11 12 13 14 15 17 18 19 20 21 22 23 24 25 26 27 28  ha 1.1 1.0 1.0 1.0 1.0 1.8 1.0 1.0 1.0 1.0 1.0 1.2 1.0 1.0 1.0 1.0 1.0 1.0 0.9 1.0 1.2 4.0 6.2 5.3 16.0 943.0  sh 4 3 4 5 4 4 3 4 4 4 4 4 3 4 3 4 4 5 3 4 4 15 14 12 35 467  cr 4 2 1 2 1 2 2 2 6 1 5 3 2 1 4 2 3 7 6 7 4 6 3 2 2 3  type 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 3 2 2 2 3  wha 0.7 0.9 1.0 0.9 1.0 1.0 1.0 1.0 0.8 1.0 1.0 0.4 1.0 1.0 0.8 1.0 0.9 0.8 0.7 0.8 1.0 0.8 646.0  wt 3.0 2.8 3.0 3.0 2.5 4.0 4.6 2.8 5.0 3.0 4.2 1.5 3.5 3.0 3.5 3.5 3.8 4.0 4.5 3.8 3.8 2.4 1,948.0  wy 4.3 3.3 3.0 3.3 2.5 4.0 4.6 2.8 6.3 3.0 4.2 3.8 3.5 3.0 4.4 3.5 4.2 5.0 6.4 4.8 3.8 3.0 3.0  cha 2.2 4.0 3.0 12.0 278.0  ct 6.2 9.5 8.0 36.0 429.0  cy 2.8 2.4 2.7 3.0 1.5  income 2,490 1,128 1,200 1,560 900 1,900 1,300 2,120 4,573 1,200 2,100 823 1,750 1,200 4,050 2,900 4,000 4,004 3,459 3,387 3,200 4,400 3,930 4,509 15,400 498,100  exp 520 835 300 240 200 700 949 340 590 200 1,206 662 431 170 513 320 450 785 681 685 465 1,641 2,650 6,310 12,200 646,800  debr debfc continued on next page  97  continued from DF ha 29 1,770.0 30 1,344.0 39 583.0 40 1,122.0 41 2,445.0 42 312.0 43 499.0 44 482.0 45 11.9 46 14.0 47 20.0 48 24.0 49 243.9 50 693.5 51 749.0 52 578.0 53 642.3 55 103.0 56 169.0 58 83.0 59 1,722.0 60 110.0 61 90.0 62 85.0 63 74.0 64 438.0 65 498.0 66 199.0  previous page sh cr type 2,077 3 3 1,978 3 3 252 4 3 1,120 4 3 562 4 3 314 4 1 280 4 3 498 4 3 12 1 1 12 2 3 8 1 1 15 2 3 1,408 3 3 900 3 3 1,295 3 3 1,650 3 2 832 3 3 192 1 2 258 1 2 196 1 2 2,288 1 2 204 1 2 222 1 2 182 1 2 254 1 2 642 1 2 580 2 1 424 2 1  wha 285.0 173.0 357.0 442.0 1,910.0 310.0 428.0 407.0 11.9 5.0 20.0 12.0 230.0 71.0 66.0 130.0 492.0 196.0  wt 712.0 461.0 359.0 276.6 2,266.2 287.4 474.5 555.1 16.4 8.0 26.0 25.2 276.0 106.5 94.0 240.0 442.8 176.4  wy 2.5 2.7 1.0 0.6 1.2 0.9 1.1 1.4 1.4 1.6 1.3 2.1 1.2 1.5 1.4 1.8 0.9 0.9  cha 1,480.0 1,156.0 215.0 670.0 493.0 60.0 70.0 9.0 12.0 8.4 620.0 680.0 578.0 509.0 103.0 169.0 83.0 1,722.0 110.0 90.0 85.0 74.0 438.0 -  ct 2,030.0 1,817.7 256.4 915.6 342.1 68.1 61.3 23.4 30.0 11.7 816.0 816.0 968.4 567.0 215.0 311.0 198.0 3,150.0 168.0 125.0 210.0 147.0 958.0 -  cy 1.4 1.6 1.2 1.4 0.7 1.1 0.9 2.6 2.5 1.4 1.3 1.2 1.7 1.1 2.1 1.8 2.4 1.8 1.5 1.4 2.5 2.0 2.2 -  income 1,878,000 2,817,100 411,100 798,600 1,368,200 319,600 289,600 268,500 2,400 1,700 2,900 6,200 775,000 687,000 982,000 761,200 593,900 114,100 118,300 35,500 3,876,000 110,400 97,600 92,800 120,000 1,025,000 139,400 70,600  exp 2,637,000 3,359,400 411,100 319,600 2,400 1,700 2,900 6,200 775,000 687,000 991,000 633,400 452,400 114,100 118,300 33,500 3,876,000 110,400 97,600 92,800 120,000 1,025,000 139,600 60,700  debr 13,900 1,539,000 41,000 64,000 22,500 28,500 14,000 3,200 311,000 10,500 6,900 2,300 3,400 144,000 2,300 2,000 continued on  debfc 1,384,400 2,321,000 2,984,000 2,897,000 2,045,000 42,700 72,000 207,000 176,900 3,013,000 next page  98  continued from DF ha 67 195.0 68 320.0 69 144.0 70 144.0 71 1,580.0 72 2,834.0 73 1,258.0 74 98.0 75 32.5 76 117.0 77 24.5 78 2.0 79 65.5 80 72.0 83 4.5 84 2.1 87 3.4 88 3.0 89 113.0 90 294.0 91 122.0 92 55.0 93 104.0 94 231.0 95 314.0 96 120.0 97 90.2 98 48.5  previous page sh cr type 378 2 1 240 2 1 163 2 1 155 2 1 979 3 3 1,927 3 3 1,150 3 3 1,420 3 1 10 3 1 35 3 1 9 3 1 6 3 1 8 3 1 61 3 1 12 2 1 8 2 1 6 2 1 8 2 1 49 2 1 49 2 1 22 2 1 55 1 1 78 2 1 80 2 1 13 2 1 42 1 1 120 3 3 97 3 3  wha 191.0 320.0 144.0 144.0 835.0 1,152.0 528.0 86.0 22.0 92.0 20.0 1.0 42.0 49.0 1.0 0.5 0.7 1.0 112.0 291.0 120.0 55.0 103.0 230.0 310.0 120.0 17.0 10.0  wt 171.9 288.0 144.0 153.0 701.0 1,999.0 566.0 90.3 22.0 86.3 25.6 1.2 42.0 34.0 1.0 0.5 0.5 1.0 148.0 380.0 216.0 100.0 200.0 287.5 589.0 201.0 56.0 30.6  wy 0.9 0.9 1.0 1.1 0.8 1.7 1.1 1.1 1.0 0.9 1.3 1.2 1.0 0.7 1.0 1.0 0.7 1.0 1.3 1.3 1.8 1.8 1.9 1.3 1.9 1.7 3.3 3.1  cha 745.0 1,561.0 730.0 71.2 37.3  ct 1,803.0 1,274.0 183.0 102.0  cy 2.4 1.7 2.6 2.7  income 96,000 84,200 50,400 38,200 907,600 2,259,000 744,600 166,300 4,300 46,300 2,900 4,800 6,500 23,600 15,200 8,800 15,400 7,600 14,300 17,300 8,500 6,600 22,200 17,000 45,400 9,400 154,600 86,400  exp 95,800 77,000 49,200 37,300 1,310,000 2,322,000 774,300 170,500 3,300 41,800 1,800 4,500 4,700 20,600 15,200 8,800 15,400 7,600 14,300 17,300 8,500 6,600 22,200 16,700 45,500 9,400 155,500 93,100  debr debfc 1,900 1,600 900 800 260,600 309,700 346,200 274,100 109,600 86,400 41,600 32,800 2,300 800 1,300 5,500 11,600 2,500 15,900 9,400 5,600 29,700 22,800 7,800 1,900 5,200 1,400 14,600 continued on next page  99  continued from DF ha 99 48.0 100 53.0 101 72.0 102 54.5 103 83.0 104 45.6 105 356.0 106 87.0 107 4.0 108 3.3 109 391.0 110 151.0 111 10.0 112 6.5 113 492.0 114 1,029.0 115 9.0 116 704.0 117 1,260.0 118 8.0 119 16.0 120 19.0 121 15.0 122 17.0 123 20.0 124 22.0 125 692.0 126 111.0  previous page sh cr type 3 3 3 3 158 3 3 3 3 3 3 3 3 30 3 1 19 5 1 19 3 1 4 3 1 28 3 1 27 5 1 9 3 1 9 3 1 420 3 1 1,146 3 1 5 3 1 273 3 1 585 3 1 5 2 1 5 4 1 8 4 1 6 4 1 4 4 1 9 4 1 8 4 1 120 4 1 120 4 1  wha 10.0 11.0 14.0 10.5 17.0 10.0 330.0 67.0 1.0 1.0 382.0 100.0 6.0 3.0 203.0 162.0 7.0 310.0 120.0 6.0 11.0 11.0 5.0 12.0 10.0 10.0 527.0 50.0  wt 29.0 29.7 36.0 26.0 48.0 23.0 421.0 64.0 1.0 1.0 492.0 80.0 3.0 1.4 350.0 204.0 5.6 450.0 130.0 6.3 14.8 14.8 9.3 9.6 15.0 15.6 885.4 24.5  wy 2.9 2.7 2.6 2.5 2.8 2.3 1.3 1.0 1.0 1.0 1.3 0.8 0.5 0.5 1.7 1.3 0.8 1.5 1.1 1.1 1.3 1.3 1.9 0.8 1.5 1.6 1.7 0.5  cha 36.0 41.0 56.0 41.0 63.0 33.6 -  ct 63.0 82.0 112.0 90.0 163.0 96.0 -  cy 1.8 2.0 2.0 2.2 2.6 2.9 -  income 67,700 77,000 102,000 95,500 193,400 88,400 16,800 7,800 3,200 18,100 11,400 8,400 3,000 3,800 17,400 11,500 2,300 41,500 50,000 13,500 3,100 3,500 2,800 1,900 3,200 2,500 34,100 29,800  exp 57,300 75,500 106,200 85,500 187,400 68,500 15,900 9,400 2,500 11,700 24,300 10,600 3,800 5,600 23,400 35,700 2,300 68,600 86,000 13,600 2,200 2,600 1,600 1,700 3,000 2,300 43,400 27,400  debr debfc 18,000 17,300 1,400 3,700 1,200 12,900 3,000 30,000 50,000 200 1,400 1,700 1,900 31,000 34,100 29,400 39,500 21,300 16,600 continued on next page  100  continued from DF ha 127 1,233.0 128 162.0 129 364.0 130 749.5 131 109.5 132 111.0 133 56.0 134 103.0 135 67.6 136 35.0 137 6.0 138 6.0 139 5.0 140 5.0 141 5.0 142 422.0 143 191.0 144 100.0 145 554.0  previous page sh cr type 104 4 1 271 4 1 112 4 1 108 4 1 110 4 3 120 3 3 60 4 3 105 4 3 74 3 3 52 4 3 8 1 1 8 5 1 7 1 2 7 1 1 7 1 1 208 3 3 43 3 3 75 3 3 570 3 3  wha 1,030.0 83.0 328.0 658.0 10.0 20.0 10.0 20.0 6.0 3.0 6.0 5.0 5.0 5.0 303.0 110.0 29.0 149.0  wt 2,008.0 48.5 328.0 1,382.0 20.0 29.0 12.0 17.6 7.8 4.5 30.0 30.0 30.0 20.0 457.0 135.0 58.0 283.0  wy 1.9 0.6 1.0 2.1 2.0 1.5 1.2 0.9 1.3 1.5 5.0 6.0 6.0 4.0 1.5 1.2 2.0 1.9  cha 97.0 90.0 44.0 80.0 61.0 31.0 5.0 100.0 80.0 70.0 400.0  ct 224.0 169.0 95.0 194.0 138.4 76.0 12.5 115.0 135.0 158.0 863.3  cy 2.3 1.9 2.2 2.4 2.3 2.5 2.5 1.2 1.7 2.3 2.2  income 81,800 43,200 31,200 45,800 97,200 46,000 53,800 93,400 85,300 28,800 7,200 11,950 11,250 6,000 3,500 359,000 254,000 125,000 371,300  exp 103,000 45,500 30,600 37,500 89,800 33,900 44,300 99,000 47,000 23,000 700 2,824 3,500 600 620 325,500 211,700 83,500 339,900  debr 35,500 4,900 43,300 53,100 49,400 32,400 22,600 28,000 28,400 9,600 42,300 83,700 5,700 27,600  debfc 128,200 78,400 73,100 84,200 71,500 30,000 537,700 406,700 171,300  Chapter 5  Discussion This thesis relied heavily upon three pieces of critical data. The first were national statistics, which assisted in characterizing Tajikistan’s economy, as well as in understanding the agricultural sector and the broader economy, more generally. The second were data comprised of copies of original farm invoices from Dzerzhinsky collective farm on the outskirts of Dushanbe. While they cannot characterize the nature of the range of contractual relationships between gins and farms within the republic, these invoices helped to paint a picture of a process and a system. Through this characterization, a theoretical model was developed and tested by using both national statistics, as well as farm level data obtained from Dzerzhinsky farm. In chapter 4, external survey data obtained from Food and Agriculture Organization of the UN (FAO) was utilized to test the conjecture that, at an individual level, there was little difference in livelihoods between private farms producing grains and collective farms producing both cotton and grains. Based upon the analysis and arguments presented in this thesis, the following may be reasonably argued: 1. Farm “debt” would seem to be a problem of “mixed” farms (growing both cotton and grains) and due largely to a lack of access to credit (thin credit markets) and not a problem of unprofitable “cotton production” per se. Analysis of survey data revealed: • no statistical difference between “debt” per farmer as well as financial position of farmers on different farm types • no statistical difference in cotton yields between different farm types • no statistical difference in wheat yields between different farm types 2. Accrued “debt” for cotton is likely the extra cost of inputs that mixed farms have utilized, in non-cotton production, over and above the amount of inputs that were usefully needed in the production of cotton. This amount is the cost of doing business for the gin and the business is cotton. 3. The exercise of monopoly power by the gin, and specifically monopsonistic power in the purchase of cotton, cannot be rejected.  101  4. The gin provides a service to “mixed” farms in terms of facilitating access to credit at rates which would seem to be lower than prevailing market rates and for both cotton and non cotton production. 5. The gin provides cash as part of its loan package in order to finance cotton picking costs. In a residual wage system, there is little certainty of actual income at harvest and this had traditionally led to shirking on Soviet collective farms (Cameron (1973)). The payment of picking wages at harvest was one manner in which the Soviet sovkhozes dealt with shirking. In present day Tajikistan, the role of wage payments for cotton picking would seem to have a different aim. Specifically, the payment of wages for cotton picking ensures that some (minimal) amount of cotton is always produced and that the cash wages act as a plug to close any income gap between an individual farmer producing grains and other food crops on private land and an individual farmer producing cotton and some grains on collectively held land. The first item has been reasonably dealt with in this thesis, but the issue of interlinkage and how it appears has not been prominent in the discussion and nor has a discussion on who owns the debt. Indeed, the analysis has assumed interlinkage to start with rather than explaining how it comes about. The implicit assumption has been that interlinkage was initially born out an environment where financing for agricultural production (particularly cotton) was curtailed after the dissolution of the Soviet Union and as a result of the civil war that ensued in Tajikistan at dissolution. With cotton as a strategic crop for the government, financing for cotton production was initially obtained by a sovereign debt agreement which later developed into a complex system of future contracts and prepayment. Early payment on future contracts for delivery of cotton fibre has become the basis for working capital in the production of cotton and is channeled through the gins or its agents who also negotiate these contracts and market the fibre internationally. Given a push for decollectivization, which was also occurring during the period, the risk of losing land devoted to cotton was real and appropriate incentives were needed if cotton production was going to continue at some significant scale. Cotton provides much needed government revenue in an economy where exports are limited and is a lucrative business for the gins and other enterprises within the value chain. Where rural finance markets are thin, and where land lien legislation is weak, the ability to legally encumber the forthcoming harvest of cotton, as collateral against production loans, effectively opens the door for interlinkage. But what keeps the system alive? Grain production, and particularly wheat, is of primary importance to rural households where wheat consumption per capita is among the highest in the world. Those farmers who did disband and who chose to produce wheat on private lands are generally small and, in most cases, finance production through personal capital, informal money markets, remittances or through microcredit agencies. Cotton only farms are generally still centrally run and there is indication to believe that these farms are break even ventures. The nature of their profitability does not 102  seem to be tied to market fundamentals, but rather to the gins ability to manipulate prices in order to reach break even levels. The analysis conducted provides evidence that cannot reject the claim that monopsonistic practices are alive but there is not enough evidence to formally conclude that monopsonistic practices are indeed occurring. It was argued early on that the gin has no advantage in formally extending its monopsonistic reach and interlinking of finance and output markets to wheat (or grains). In large part, this was argued on the basis that grain markets are domestic markets and that the opportunity for arbitrage is low. Leakage from farm to market is a very real possibility, thereby undermining the gin’s ability to seize all of the forthcoming harvest as collateral and its ability to prosper from arbitrage in local and international markets for grain. Faced with a need to deliver upon contractual obligations, the gin acquiesces to the incentive of mixed farms for the provision of inputs that can be extended to non-cotton production. It does this by formally extending a larger amount of loans for cotton production, a portion of which is knowingly diverted to non-cotton production, and by securing a lien against the forthcoming harvest of cotton. By interlinking finance with the purchase of raw cotton, which has no value at the farm gate without processing, the gin exercises monopsonistic power in setting prices for cotton at each farm. In effect, the gin can price discriminate between farms, and as long as cotton production itself is a break even venture, the accrued debt is essentially the loan cost of the extra inputs applied to non-cotton production. It is in effect, the cost of doing business and the business is cotton. Who owns the “debt”? Clearly, the international purveyor of cotton does not hold any of the debt given that there has been prepayment and an obligation to deliver. When there is a default on the contractual obligations, it is unclear why this system would persist and thus either prepayment would have been discontinued or international purveyors would not have been as keen on Tajik cotton given the uncertainty on delivery. On the gin’s part, as long as the volume of cotton received yields the contractual obligations, there would be no loss to the gin other than the cost of ginning. But this could be reasonably covered through its margin obtained on the provision of inputs in kind or through the many other services that it provides in processing the joint products from the ginning process (cotton seed for oil and meal). It is only in the case where prices for inputs (fertilizer, pesticides, fuel, etc), which the gin purchases on the open market, rise to a level that makes cotton production unprofitable, does the gin expose itself to a loss. In such a case, the gin can manipulate the quantum of funds that it advances for the payment of maintenance wages, in order to try and restore break even levels, but will likely not resort to charging high prices for the inputs provided as credit in kind. Similarly, and for the same reason, the gin will not lower its price paid for farm gate raw cotton in order to compensate for higher input prices that it faces. Farmer profitability depends upon wage costs, among other variables, but the supply of cotton does  103  not. Therefore, manipulating prices paid for raw cotton and inputs provided in kind will affect the output of cotton produced, and this would affect gin profitability directly. Why the gin does not enter into contractual arrangements that are based on joint profit maximizing rules was argued to be based upon reasons of political economy. One assertion, however, is that one reason for the current arrangements is that the system assists in capital flight. By ensuring that profits from the sale of processed cotton fibre remain outside of Tajikistan, the gins ensure that they are not subjected to profit (and other) tax on their business ventures domestically. The current financial regulations do not allow for the transfer of foreign currency out of Tajikistan without a clearance certificate from the relevant tax authorities confirming that the funds to be transferred have been earned legally and that appropriate taxes have been paid on these earnings. From the gin’s perspective, therefore, joint profit maximizing rules would provide little further advantage over the existing arrangements. Despite the inefficiencies generated in the current system of agrarian relations, there is at least one positive aspect of the current contractual relations. Without the provision of prepayment and contract registry, cotton production would likely be curtailed, and large swathes of land would undoubtedly be parceled out to individual farmers for the production of grains, vegetables and fruits (non cash crops in general). This would not sit well with those who earn a very good living from the cotton value chain and there are many prominent feeders within this chain. On small holdings, where non cash crops are produced, small scale loan financing may be available from a host of intermediaries but the desire and incentive to grow cotton is rooted in the potential for a windfall gain from a bumper crop and high prices. In a typical year with moderate prices and yields, growing both cotton and grain does not have a disadvantage in the current system, given the alternative of either growing cotton alone or growing wheat on small scale land holdings. In addition, earnings from both maintenance and piece rate wages provide the benefits of employment to a large proportion of the rural population and security in obtaining minimum levels of a food and domestic goods. In a republic where there is a constant fear of social unrest, and a desire not to have another civil war, the function of the gin plays an important role in “keeping the peace” within the countryside. “Debt”, therefore, would seem to “fictitious” in some senses and continues unabated in this system because the gin internalizes the extra financing for non-cotton production into its production and marketing costs, while the farms do not see “debt” as inhibiting given that there is no credible threat to reinforce payment nor legal legislation for seizing capital assets. The government’s continual announcement for a debt moratorium or debt cancelation further diminishes the threat that “debt” would pose in any other economy with functioning legal systems for enhancing debt recovery particularly when the government is also actively involved in the sourcing and delivery of credit for cotton production.  104  The question of whether the gin is imposing a social cost upon society is a moot point given that there are limited opportunities for agricultural production other than cotton, wheat and small scale vegetable production. Topography and geographical location are major factors in limiting opportunities for agricultural production. Given these constraints, and from a positive perspective, the gin provides capital for the production of both cotton and grains and allows some cash to be injected into the countryside through the payment of wages and other benefits (transportation, contribution to local government treasury, et cetera). That no formal institutions have entered the arena to finance cotton production is indicative of the risks inherent in cotton production and marketing as well as a non functioning mortgage system and land market. The inability to find alternative markets and alternatives to cotton further inhibits diversity in the countryside. A formal analysis on whether legislation can inhibit the ability of the gin to exert market power has not been undertaken, but it would seem that under the current constraints and economic environment, incentives for all agents in the current system (gin, farmers, government, international purveyors) are aligned. At the time of writing, the world price of cotton fibre has risen significantly from a low of 50 cents per pound in 2009 to approximately $1.75 per pound.86 Time will tell, but it is unlikely that this will have any significant impact upon the nature of relations between the gins and collective farms, and particularly upon the allocation of land placed under cotton. The analysis conducted in this study was based upon a model, whereby the gin extracts surplus by paying the farm a price that is lower than the world price and capturing the mark-up surplus. For the collective farm, it is the maintenance wage, a fixed cost, that is the key to profitability of the farm operation and not the farm gate price for cotton. By manipulating this fixed wage, the gin can ensure that each farmer’s income is never lower (or higher) than the available outside option. A higher farm gate price, therefore, is likely to do little in terms of changes to organizational structure and agrarian relations. The sector is locked into a low level equilibrium and inefficiencies would seem to be perpetuated for historical reasons, vested interests in the production of cotton and for reasons of political economy. One explanation for why farmers produce cotton collectively in post-Soviet Tajikistan, despite the accrual of institutional “debt”, is that it is privately beneficial to do so; and with little difference upon standards of living, relative to farmers who privately produce wheat and other food crops.  86 I thank Richard Pomfret for the helpful prompt in thinking through the potential implications of such a rise in the world price of cotton fibre.  105  Bibliography Aghion, P. and Bolton, P.: 1992, An incomplete contracts approach to financial contracting, Review of Economic Studies 14, 473–494. Akerlof, G. and Romer, P.: 1993, Looting: The economic underworld of bankruptcy for profit, Brookings Papers on Economic Activity (2), 1–73. Alston, L. J.: 1990, Issues in postbellum southern agriculture, in L. Ferleger (ed.), Agriculture and National Development: Views on the Nineteenth Century, Iowa State University Press, Ames, pp. 207–228. Amelina, M.: 2000, Why russian peasants remain in collective farms: A household perspective on agricultural restructuring, Post-Soviet Geography and Economics 41(7), 483–511. Bale, M. D.: 2008, Tajikistan: Cotton sector and proposed policy, Technical Assistance Consultant’s Report Project number 41216, Asian Development Bank. Basu, K.: 1983, The emergence of isolation and interlinkage in rural markets, Oxford Economic Papers 35(35), 262–280. Bliss, C. and Stern, N.: 1982, Palanpur:The Economy of an Indian Village, Clarendon Press, Oxford. Bolton, P. and Rosenthal, H.: 2002, Political intervention in debt contracts, The Journal of Political Economy 110(5), 1103–1134. Bonin, J. P.: 1977, Work incentives and uncertainty on a collective farm, Journal of Comparative Economics 1, 77–97. Bradley, M. E.: 1971, Incentives and labour supply on soviet collective farms, The Canadian Journal of Economics 4(3), 342–352. Braverman, A. and Stiglitz, J. E.: 1982, Sharecropping and the interlinking of agrarian markets, American Economic Review . Burkart, M. and Ellingsen, T.: 2004, In-kind finance: A theory of trade credit, American Economic Review 94(3), 569–590. Caccavale, O. M.: 2005, Land reform in tajikistan: Economic effect on private farms, Technical report, FAO. 106  Cameron, N.: 1973, Incentives and labour supply in co-operative enterprises, The Canadian Journal of Economics 6(1), 16–22. Carter, M., J. M.-S. and Luz, K.: 1993, Can production cooperatives resolve the conundrum of exclusionary growth? an econometric evaluation of land reform cooperatives in honduras and nicaragua, in C. Csaki and Y. Kislev (eds), Agricultural Cooperatives in Transition, Westview Press, Boulder Co. Cheung, S.: 1969, The Theory of Share Tenancy, The University of Chicago, Chicago. Domar, E. D.: 1966, The soviet collective farm as a producer cooperative, The American Economic Review 56(4), 734–757. Eswaran, M. and Kotwal, A.: 1985, A theory of contractual structure in agriculture, American Economic Review 75(3), 352–367. Evans, R.: 1970, Some notes on coerced labor, The Journal of Economic History 30(4), 861–866. Fogel, R. and Engerman, S.: 1995, Time on the Cross: The economics of American negro slavery, WW Norton and Company, New York. Gangopadhyay, S. and Sengupta, K.: 1986, Interlinkages in rural markets, Oxford Economic Papers 38(1), 112–121. Government of Tajikistan: 2007, The cotton sector of tajikistan: New opportunities for the international cotton trade. Hallagan, W.: 1978, Self selection by contractual choice and the theory of sharecropping, Bell Journal of Economics 9, 344–354. Hart, O. and Moore, J.: 1998, Default and renegotiation: A dynamic model of debt, Quarterly Journal of Economics 113(1), 1–41. IMF and World Bank: 2009, Republic of tajikistan: Financial sector assessment, Technical Report IDA/SecM2009-0111, IMF and World Bank. International Monetary Fund: 2008, Republic of tajikistan: Financial system stability assessment, Technical Report IMF Country Report No. 08/371, International Monetary Fund, Washington. Ireland, N. J. and Law, P. J.: 1980, Private plot restrictions in a collective farm model, Canadian Journal of Economics 13(3), 475–485. Jaynes, G. D.: 1982, Economic theory and land tenure, in H. Binswanger and M. Rosenweig (eds), Rural Labour Markets in Asia:Contractual Arrangements, Employment and Wages, Yale University Press, New Haven. 107  Jones, L., Black, R. and Skeldon, R.: 2007, Migration and poverty reduction in tajikistan, Working paper C11, Development Research Centre on Migration, Globalisation and Poverty. Kireyev, A.: 2006, The macroeconomics of remittances: The case of tajikistan, Technical Report WP/06/2, International Monetary Fund. Kramer, M.: 1989, Can gorbachev feed russia, The New York Times, April 9th . Lerman, Z. and Sedik, D.: 2009a, Agricultural development and household incomes in central asia: A survey of tajikistan, 2003-2008, Eurasian Geography and Economics 50(3), 301– 326. Lerman, Z. and Sedik, D.: 2009b, Farm debt in transition countries: Lessons for tajikistan, mimeo . Lerman, Zvi, C. C. and Feder, G.: 2004, Agriculture in Transition: Land Policies and Evolving Farm Structures in Post-Soviet Countries, Lexington Books, Oxford. Machnes, Y. and Schnytzer, A.: 1993, Risk and the collective farm in transition, in C. Csaki and Y. Kislev (eds), Agricultural Cooperatives in Transition, Westview Press, Boulder Co. Mathewson, G. and Winter, R.: 1984, An economic theory of vertical restraints, The Rand Journal of Economics 15(1), 27–38. Mathewson, G. and Winter, R.: 1985, The economics of franchise contracts, Journal of Law and Economics 28(3), 503–526. Mitra, P. K.: 1983, A theory of interlinked rural transactions, Journal of Public Economics 20. Newberry, D. and Stiglitz, J.: 1979, Sharecropping, risk sharing and the importance of imperfect information, in J. Roumasset (ed.), Risk, Uncertainty and Agricultural Development, Agricultural Development Council, New York. Oi, W. and Clayton, E. M.: 1968, A peasant’s view of soviet collective farm, American Economic Review 58(1), 37–59. Pant, C.: 1983, Tenancy and family resources: A model and some empirical analysis, Journal of Development Economics (12), 27–40. Pomfret, R.: 2007, Distortions to agricultural incentives in tajikistan, turkmenistan and uzbekistan, Technical Report Agricultural Distortions Working Paper 05, World Bank. Ransom, R. and Sutch, R.: 1972, Debt peonage in the cotton south after the civil war, Journal of Economic History 32.  108  Ransom, R. and Sutch, R.: 1975, The “lock-in” mechanism and overproduction of cotton in the postbellum south, Agricultural History 24, 405–425. Reid, J.: 1976, Sharecropping and agricultural uncertainty, Economic Development and Cultural Change 24, 549–576. Robinson, J.: 1967, The soviet collective farm as a producer cooperative: Comment, The American Economic Review 57(1), 222–223. Ruben, R. and Lerman, Z.: 2005, Why nicaraguan peasants stay in agricultural production cooperatives, Revista Europa de Estudios Latinoamericanos y del Caribe 78, 31–47. Sen, A. K.: 1966, Labour allocation in a cooperative enterprise, Review of Economic Studies (23), 361–371. State Statistical Committee, Government of Tajikistan: 2004, Òàäæèêèñòàí çà 1998-2003ãò ,  .  State Statistical Committee, Government of Tajikistan: 2005, êè Òàäæèêèñòàí, Ñòàòèñòè÷åñêèé ñáîðíèê,  Ñåëúñêîå Õîçÿéñòâî Ðåñïóáëè-  .  State Statistical Committee, Government of Tajikistan: 2007, Òàäæèêèñòàí çà 2001-2006 ãò,  Õëîïêîâîäñòâî â Ðåñïóáëèêå  Õëîïêîâîäñòâî â Ðåñïóáëèêå  .  State Statistical Committee, Government of Tajikistan: 2008, Food security and poverty in the republic of tajikistan. Stern, D. L.: 2008, In tajikistan, debt-ridden farmers say they are the pawns, The New York Times, October 15th . Tashrifov, M.-Y.: 2005, The effects of market reform on cotton production efficiency. the case of tajikistan., Working Paper 05-8. The Centre for International Studies and Co-Operation: 2006, Comparative assessment of cotton growing dekhan farms models in sughd, tajikistan, Technical Report FARMS Project-Tajikistan, The Centre for International Studies and Co-Operation, Dushanbe. The Centre for International Studies and Cooperation: November 2006, Dekhan farms inside/out: A comparative assessment of cotton growing dekhan farm models in sughd, tajikistan, Technical report, FARMS Project - Tajikistan. Trevisani, T.: 2007, The emerging actor of decollectivization in uzbekistan: Private farming between newly defined political constraints and opportunities, in D. Kandiyoti (ed.), The Cotton Sector in Central Asia: Economic Policy and Development Challenges, School of Oriental and African Studies, University of London.  109  United Nations Economic and Social Commision for Asia and the Pacific: 2003, Investment Climate in Tajikistan, Country report. USAID and World Bank: 2007, A survey of collective dekhan farms, family dekhan farms and workers in agricultural enterprises, Technical report. Vanek, J.: 1970, The General Theory of Labor-Managed Market Economies, Cornell University Press, Ithaca. Wadekin, K.-E.: 1973, The Private Sector in Soviet Agriculture, University of California Press, Berkely, CA. Ward, B.:  1958, The firm in illyria:  Market syndicalism, American Economic Review  48(4), 566–589. World Bank: December 3, 2005b, Tajikistan trade diagnostic study, Technical Report 32603-TJ. World Bank: January 6, 2005a, Republic of tajikistan poverty assessment update, Technical Report 30853-TJ. World Bank: June 2004, Tajikistan: Welfare implications of cotton farmland privatization, Technical report. World Bank: May 2007, Project appraisal document on a proposed credit in the amount of sdr 10 million to the republic of tajikistan for a cotton sector recovery project, Technical Report 39494-TJ. Wright, G.: 1978, The Political Economy of the Cotton South: Households, Markets, and Wealth in the Nineteenth Century, W.W. Norton and Company.  110  


Citation Scheme:


Citations by CSL (citeproc-js)

Usage Statistics



Customize your widget with the following options, then copy and paste the code below into the HTML of your page to embed this item in your website.
                            <div id="ubcOpenCollectionsWidgetDisplay">
                            <script id="ubcOpenCollectionsWidget"
                            async >
IIIF logo Our image viewer uses the IIIF 2.0 standard. To load this item in other compatible viewers, use this url:


Related Items