Open Collections

UBC Theses and Dissertations

UBC Theses Logo

UBC Theses and Dissertations

An investigation of secondhand prices : a case of handymax dry bulk carriers Wang, Gerry Y. G. 1993

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

Item Metadata


831-ubc_1993_spring_wang_gerry.pdf [ 5.21MB ]
JSON: 831-1.0086175.json
JSON-LD: 831-1.0086175-ld.json
RDF/XML (Pretty): 831-1.0086175-rdf.xml
RDF/JSON: 831-1.0086175-rdf.json
Turtle: 831-1.0086175-turtle.txt
N-Triples: 831-1.0086175-rdf-ntriples.txt
Original Record: 831-1.0086175-source.json
Full Text

Full Text

AN INVESTIGATION OF SECONDHAND PRICES: A CASE OF HANDYMAX DRY BULK CARRIERS  BY  GERRY Y.G. WANG MA, Shanghai Maritime University, 1986  A THESIS SUBMITTED IN PARTIAL FULFILMENT OF THE REQUIREMENTS FOR THE DEGREE OF MASTER OF SCIENCE IN BUSINESS ADMINISTRATION  in  THE FACULTY OF COMMERCE AND BUSINESS ADMINISTRATION  We accept this thesis as conforming to the required standard  THE UNIVERSITY OF BRITISH COLUMBIA April 1993 © GERRY Y.G. WANG  In presenting this thesis in partial fulfilment of the requirements for an advanced degree at the University of British Columbia, I agree that the Library shall make it freely available for reference and study. I further agree that permission for extensive copying of this thesis for scholarly purposes may be granted by the head of my department or by his or her representatives. It is understood that copying or publication of this thesis for financial gain shall not be allowed without my written permission.  (Signature)  Department of  G  9114"te,1,0-^  The University of British Columbia Vancouver, Canada  Date  ^  DE-6 (2/88)  R‘c  30,  iN3  4-4161-1^AAA.  A,—  ABSTRACT  The behaviour of the secondhand values of ships in the sale and purchase market is a very interesting issue. Almost all existing research related to this issue is qualitative or theoretical. The conclusions have not been subject to empirical testing. In this thesis, the quantitative relationship between the secondhand values and the factors that affect the values significantly is examined for the particular case of handymax dry bulk carriers. A model is estimated to examine the effects of changes in newbuilding prices, the freight market and other market conditions on the secondhand values. The results obtained are very consistent with the most popular qualitative insertions. Through the selected log linear autoregression model, it is found that the secondhand values of the handymax bulk carriers are primarily related to endogenous market conditions within the handymax sector, such as freight rates, newbuilding prices and the age of handymax bulkers. The thesis is a new approach to the investigation of the ship market and freight market. It is of particular value to the participants in the ship sale and purchase market. The data used for the research are composed of about 3,064 observations of the sale and purchase transaction records for the past 10 years plus corresponding collection of other variables.  ii  TABLE OF CONTENTS PAGE ABSTRACT ^  i  TABLE OF CONTENTS ^  iii  LIST OF TABLES ^ LIST OF FIGURES ^  vi  ACKNOWLEDGEMENT ^  vi i  I.^INTRODUCTION 1. 2. 3. 4. 5. 6. II.  BACKGROUND ^ SCOPE AND DEFINITIONS ^ PURPOSE ^ OUTLINE ^ DATA BASE ^ LIMITATIONS AND FURTHER RESEARCH TOPICS ^  1 2 4 5 6 7  REVIEW OF OTHER STUDIES ON SHIP PRICES AND COMPARISON WITH THIS STUDY 1. 2. 3. 4.  INTRODUCTION ^ CAPITAL THEORY AS PERFORMED BY MICHAEL BEENSTOCKS ^ AN ECONOMETRIC MODEL OF SHIP PRICES PERFORMED BY CHAREMZA AND GRONICKI ^ BRIEF COMPARISON WITH THE APPROACH DEVELOPED IN THIS PAPER ^  9 10 15 17  III. INTRODUCTION OF HANDYMAX DRY BULK SHIPPING MARKET 1. 2. 3. 4. 5. 6.  INTRODUCTION ^ A BRIEF INTRODUCTION OF OVERALL DRY BULK CARRIER FLEET ^ IDENTIFICATION OF HANDYMAX BULK CARRIERS ^ DEMAND FOR HANDYMAX BULK CARRIERS ^ DEMAND OF HANDYMAX BULK CARRIERS ^ HANDYMAX FREIGHT MARKET AND REVIEW 1982-1991 ^  iii  19 19 21 26 30 37  IV.  SHIP SALE AND PURCHASE MARKET 1. 2. 3. 4. 5. 6. 7. 8.  V.  41 41 43 45 51 59 65 69  OBSERVATION OF FACTORS AFFECTING PRICES 1. 2. 3. 4. 5. 6. 7. 8. 9.  VI.  INTRODUCTION ^ INTRODUCTION OF SHIP EXCHANGE MARKET STRUCTURE ^ SALE AND PURCHASE BARGAINING PROCESS ^ THE SALE AND PURCHASE DECISION-MAKING ^ S&P MARKET DEMAND AND SUPPLY AND PRICE DETERMINATION ^ NEWBUILDING ORDERS FOR DRY BULK CARRIERS ^ SECOND-HAND SHIP FINANCE ^ SALE AND PURCHASE MARKET REVIEW 1982-1991 ^  INTRODUCTION ^ NEWBUILDING PRICES ^ VOYAGE CHARTER RATES ^ TIME CHARTER RATES ^ PRICES OF SUBSTITUTES ^ LAID-UP TONNAGE ^ SHIP FINANCE ^ SCRAP PRICES ^ OTHER FACTORS ^  72 72 74 76 77 78 80 82 82  ECONOMETRIC MODEL AND EMPIRICAL ANALYSIS 1. 2. 3. 4. 5. 6. 7.  INTRODUCTION ^ GENERAL LINEAR MULTIPLE REGRESSION MODEL ^ VARIABLES INCLUDED IN ANALYSIS ^ DATA BASE ^ THE HYPOTHESIS TO TEST ^ THE COMPUTING SOFTWARE PACKAGE USED ^ EMPIRICAL RESULTS AND INTERPRETATIONS ^  VII. SUMMARY ^  84 84 91 92 95 96 96 105  APPENDIX I APPENDIX II APPENDIX III REFERENCES  iv  LIST OF TABLE PAGE TABLE 3.1 - HANDYMAX DRY BULK CARRIER FLEET BY TYPE ^ TABLE 3.2 - GROWTH OF THE HANDYMAX BULK FLEET (1970 - 1991) ^ TABLE 3.3 - MAJOR BULK CARGOES CARRIED BY HANDYMAX BULK CARRIERS ^ TABLE 3.4 - TOTAL WORLD SEABORNE TRADE IN THE MAJOR BULK CARGOES ^ TABLE 3.5 - TOTAL WORLD SEABORNE TRADE IN THE MINOR BULK CARGOES ^ TABLE 3.6 - HANDYMAX DRY BULK CARRIER FLEET BY SIZE AT 1 JANUARY 1991 ^ TABLE 3.7 - HANDYMAX DRY BULK CARRIER FLEET BY AGE AT 1 JANUARY 1991 ^ TABLE 4.1 - BULK CARRIER NEWBUILDING ORDERS BY SIZE 1985 - 1990 ^ TABLE 4.2 - NUMBER OF SALES OF DRY BULK CARRIERS 1982 - 1991 ^ TABLE 4.3 - PRICES OF SECONDHAND DRY BULK CARRIERS 1982 - 1990^(USD/DWT) ^ TABLE 6.1 - REGRESSION RESULTS TABLE ^  V  23 27 28 29 31 34 35 60 70 70 97  LIST OF FIGURES PAGE FIGURE FIGURE FIGURE FIGURE  3.1 3.2 3.3 4.1  -  FIGURE 4.2 FIGURE 5.1 FIGURE 5.2 FIGURE 5.3 FIGURE 5.4 -  BULK CARRIER FLEET SIZES ^ BULK CARRIER FLEET COMPOSITION ^ TIME CHARTER RATES 1982 - 1991 ^ HANDYMAX DRY BULK CARRIERS PRICES 1981 - 1991 ^ AGE PROFILE OF SECONDHAND DRY BULK CARRIERS ^ HANDYMAX BULK CARRIER PRICES 1982 - 1991 ^ TIME CHARTER RATES VS. SECONDHAND PRICES 1983 - 1991 ^ COMPARISON OF SECONDHAND BULK CARRIERS 1981 - 1991 ^ COMPARISON OF LAIDUP TONNAGE AND SECONDHAND PRICE ^  vi  20 21 41 55 71 74 78 80 81  ACKNOWLEDGEMENT  The writer wishes to thank Dr. Trevor Heaver for his helpful and encouraging supervision and comments during the preparation of this study.  vi i  1  CHAPTER I  INTRODUCTION  1. BACKGROUND  Lloyd's List dated June 31, 1991 had the following report: "Dry bulk carrier values began to increase as a result of the firm tone on the freight market. At least two handymax ships, one of early 1980s and the other of 1976 built, were being finalized at prices higher than last seen, said market sources."  Several questions could be asked with regards to the above statement. Would increases in the secondhand ship prices generally come along with the "firm tone" on the freight market? Why and to what extent? What about other market conditions, such as the corresponding newbuilding prices of dry bulk carriers?  Examination of the records of sales of dry bulk carriers reveals that the activity in the sale and purchase market changes substantially from time to time and, that the secondhand ship prices have a very high degree of variability. This makes it difficult to predict future trends. Shipowners and other operators always find themselves vulnerable in the shipping business when making decision about when and how much to invest in ships.  Unfortunately, there is only a very limited amount of literature to  2 help the investors understand the relationship between the prices of used ships and market conditions. Most of the literature focuses on qualitative or theoretical arguments without giving any convincing quantitative analysis. Academically, considerable research work needs to be done in this area to provide systematic guidance to the shipping community as a whole.  2. SCOPE AND DEFINITIONS  Conventionally, the overall dry bulk carrier market is divided into four sectors or types: the lakesize (10-30,000 dwt), the handymax (30-50,000 dwt), the panamax (50-80,000 dwt) and the capesize bulk carriers ( over 80,000 dwt ). The division is based on the conventional classification widely used in the shipping literature. Though it seems that there is no obvious economic justification for such convention, there are several arguments which seem related to the division.  The most important feature of the handymax and the lakesize bulk carriers is that they can carry a large variety of commodities such as coal, general cargo, forest products, agricultural products, mineral products and even containers, whereas the panamax and larger bulk carriers are normally designed to specialize in relatively few commodities such as grain, coal and minor ores. The lakesize bulk carriers are normally engaged in intra-region trades and trades to and from the Great Lakes.  3 For certain routes, handymax bulk carriers are more economical than the lakesize and panamax bulk carriers. One example is the North West Pacific and Europe trading route where the handymax bulkers are perhaps most economical, allowing for the shipping distances, port condition, type and volume of cargo. The handymax bulkers along with the lakesize bulkers are now widely used within the Pacific Rim of Asia to ship steam coal and other products, principally because of a lack of suitable reception and transhipment facilities, but also because of the relatively short shipping distances involved.  Geographical restrictions also play an important role in the division of dry bulk carriers. The two typical examples are the St. Lawrence/Great Lakes and Panama Canal transits where sizes of ships are physically limited.  The thesis focuses on the handymax type. However, in order to find out the substitution effects in the ship sale and purchase market between the handymax bulk carriers and the lakesize, the panamax bulk carriers, it is necessary to relate the discussion to the lakesize and panamax types. The capesize bulk carriers are not included in the analysis for the reasons that (a) there are few units, most built specially for fixed trading, (b) the market of the capesize bulk carriers itself is very independent.  4 3.^PURPOSE  The primary purpose of this study is to explore, analyze and attempt to find out the factors that are statistically significantly related to the prices of the secondhand handymax dry bulk carriers. In other words, the thesis is to seek a regression model in which the dependent variable, the secondhand prices, is expressed in a functional form by some significant independent variables.  To achieve this goal, it is necessary to review the basic structure of the handymax dry bulk shipping market as well as the ship sale and purchase market, and then to test empirically hypotheses about the interrelationship between the secondhand prices and the proposed factors. As a result of possible existence of autocorrelation due to time series data, a log linear multiple autoregression model and SAS computer program are applied. The results will be of interest to sellers and buyers of ships, and to ship brokers and financiers.  The regression model applied is based on the assumption that there exist log linear relationships between the dependent variable, the prices of the secondhand dry bulk carriers, and a group of proposed independent variables. The hypothesis testing is to test whether the log linear relationships, positive or negative, are statistically significant. The hypotheses include, with the prices  5 of the secondhand dry bulk carriers, (a) a negative relationship of ship age, (b) positive relationships of secondhand prices of the handymax substitutes: the lakesize and the panamax dry bulk carriers, (c) a positive relationship of newbuilding prices of each type of dry bulk carriers, (d) a positive relationship of freight market of each type of dry bulk carriers, (e) a negative relationship of laid-up tonnage of each type of dry bulk carriers, and (f) a negative relationship of financing cost for acquisitions.  4. OUTLINE  This thesis consists of 7 chapters in total. Chapter I gives the introduction of this thesis. Chapter II takes a look at the existing literature in this area and make comparisons between the findings of others and the results of this thesis. Chapter III describes the handymax dry bulk shipping market and its development over the last 10 years while Chapter IV examines the ship sale and purchase market with emphasis on the handymax dry bulk carriers. Chapter IV discusses the possible factors that might be significantly related to the secondhand values. A log liner multiple autoregression model is developed in Chapter V and the hypotheses listed in the previous section are tested using the model. The model accounts for autocorrelation and multicollinearity due to time series data used in estimating the model. The last chapter, Chapter VII, summarizes the main findings and points out possible further research subjects.  6 5. DATA BASE  The data base is composed primarily of the individual records of the recorded sale and purchase transactions of dry bulk carriers ranging from 10,000 dwt to 80,000 dwt that took place during the last 10 years. The observations of such variables as freight rates, newbuilding prices and so on were made on monthly averages due to the limitations of the existence of individual records. Despite this, a big sample size of 615 observations should ensure very reliable regression results.  From the weekly ship sale and purchase market reports of Fearnley, Oslo and Wardley, Hong Kong, 3064 observations were collected for the transactions of used dry bulk carriers for the past 10 years, out of which 615 were for panamax and 856 and 1,603 for handymax and lakesize, respectively. The format of the observations is shown in Appendix I. In order to have full rank for the independent variable matrix used in the regression analysis, only 615 observations were randomly chosen from the lakesize and handymax sub-samples so as to match the size of the panamax sub-sample.  The data on newbuilding prices, freight rates on monthly basis were collected from Shipping Statistics of Institute of Shipping and Logistics, Bremen; the data on laid-up tonnage were collected from  Shipping Economics and Statistics of Drewry Shipping Consultants.  7 6. LIMITATIONS AND FURTHER RESEARCH TOPICS  Although it is believed that data collected on the sale and purchase transactions for used dry bulk carriers have high coverage and representation of the ship sale and purchase market for the past 10 years, there are certain limitations. For example, the reported prices might not be the actual prices paid which in some cases are treated confidentially, and some transactions might not be reported due to privacy and confidentiality of the sales. Apart from this, part of the data base is in monthly averages rather than individual records which are simply not existing or incompatible with each other due to different trading areas, ship sizes and timing. Therefore, the explanatory power of the regression model to reflect the truth of the ship sale and purchase market is limited.  Furthermore, the multiple autoregression model used is based on the assumption that there exist log linear relationships between the prices of used ships and ship age, laid-up tonnage and other independent variables. This assumption, partially made for convenience and simplicity, is expected to be the form of relationship closest to reality.  Despite all those limitations, the model and the analyses developed in this thesis are able to examine scientifically the prices of secondhand ships. It would be very interesting to apply the same  8 methodology as used in this thesis for the handymax bulk carriers to not only the lakesize or panamax bulk carriers, but also tankers, container vessels and other types of ships. It is expected that the results would be quite similar to those obtained in this thesis.  Further research should also be carried out by applying ARMA model (autoregressive moving average model) and find out if there are any changes to the outcomes obtained in this thesis.  9 CHAPTER II  REVIEW OF OTHER STUDIES ON SHIP PRICES AND COMPARISON WITH THIS STUDY  1.^INTRODUCTION  While it is not the intention here to dwell on the defects of any study in particular as most of them would be acknowledged by the respective authors themselves, it would be desirable to compare their methodologies and outcomes with those obtained here.  A considerable literature focuses on the freight rates in international shipping markets. By contrast, very little attention has been paid to the market for ships and the determination of ship prices. Also, the limited literature mainly deals with the market for ships as a whole, ignoring the different types of ships which have their own unique market conditions. The outcomes of the analyses based on the treatment of homogeneity of ships are inaccurate. Nevertheless, some analytical work provides methodological ways to examine the market for ships.  10  2. CAPITAL THEORY AS PERFORMED BY MICHAEL BEENSTOCKS  In his paper "A Theory of Ship Prices" 1 , Beenstock develops a theoretical model for the determination of prices of new ships (newbuildings) by applying generic capital theory to the ship market. He argues that the classical supply and demand study is inappropriate in the case of ships because a ship is a capital asset of considerable longevity. The fundamental feature of his capital theory analysis is the application of the Rational Expectation Hypothesis (REH) based on Muth 2 to the market for ships. Beenstock argues that the use of the REH is sensible first because "it would be rather uncomfortable to assume that ship owners did not make the best forecasts that they could and second, because there is substantial empirical support for REH in the case of markets for other capital assets."  His work is also based on the assumption that there exists a single, competitive, homogeneous freight market. The demand for shipping services is hypothesized to vary directly with the volume of world seaborne trade and inversely with freight rates, whereas the supply for shipping services is hypothesized to vary directly with laid-up rate, carrying capacity, speed and fleet size. Given the competitive market assumption, the equilibrium condition for 1  2  .^Beenstock, Michael (1985), "A Theory of Ship Prices," Maritime Policy and Management Vol. 12, No. 3, 215-225. ^  Muth, J. (1961), "Rational Expectations and the Theory of Price Movements," Econometrica, Vol. 29, July.  11 the freight market is that in each period the supply of shipping services is brought into balance with demand by freight rate adjustments.  Based on similar assumption as made for the freight market, the supply for ships is hypothesized to vary directly with the price of new ships, price of scrap ships and fleet size of existing ships. On the demand side of the market it is hypothesized that ships owners regard vessels as capital assets which must compete with other capital assets in their investment portfolios. As a result, the demand for ships is hypothesized to vary directly with laid up rate, freight rate and cost, expected secondhand ship price, price of new ships, returns on competing assets and wealth. In a competitive market, the equilibrium condition is that in any time period the stock demand for ships is equal to the available supply currently in existence.  The general equilibrium for the freight market and ship market is therefore jointly determined by the two markets. The stationary state for such general equilibrium is that the fleet size does not alter and ship prices do not vary. A set of arithmetic solutions are solved for the freight rates, the prices of new ships and the fleet size.  Dynamic analysis is also performed by Beenstock to allow for the nature of the time path along which the freight and ship markets  12 travel between stationary states in general equilibrium. The analysis assumes crucially that the expectations of ship owners are rational in the sense of Muth. A new set of arithmetic solutions are found out for the same parameters as under the general and static equilibrium.  A brief numerical illustration is presented by simply artificially assuming values of all the parameters concerned such as laid-up rate and so on. Strictly speaking, this might not be considered as an empirical application of the analyses.  Beenstock has certainly provided a comprehensive theoretical approach to the analysis of ship prices. In the view of this author, Beenstock's study is among the most outstanding theoretical work available up to now on the ship price analysis. However, his analysis by applying the capital finance theory has proved itself far away from adequate, accurate and systematic. Regrettably, there has been no further research to improve the approach since his paper was published in 1985.  There are following comments on Beenstock's study:  First, capital theory might be a very good analytical tool to look into ship prices as ships are capital assets or investments of considerable longevity. However, the use of the Rational Expectation Hypothesis has significant limitations in practice. The  13 REH is based on the premise that the ship market is efficient, that is, it is impossible to make economic profits by trading on available information. 3 Substantial practical evidence suggests that a lot of owners are unable to forecast ship price changes. Hale and Vanags 4 have recently applied the idea of cointegration to analyze the markets for secondhand dry bulk carriers and they raise serious doubts about the validity of efficient markets hypothesis. Marlow 5 suggests that shipping companies tend to follow cyclical investment policies. The first reason is that many owners psychologically equate company growth with fleet expansion. The second lies in the spur to new investment in ships afforded by the tax regime in a particular country such as Norway in late 1980's.  Second, his theory assumes that all ships are homogeneous. Obviously, container ships would behave quite differently from oil tankers and even dry cargo bulk carriers, for instance. Therefore, this assumption is a significant limitation.  Third, the freight market demand and supply might be mis-specified or over-simplified in that there might be some more important 3  4  5  .^Smith C.W. (1990), The Modern Theory of Corporate Finance. McGraw-Hill Publishing Company, New York, 004005.  ^  Hale C. and A. Vanags (1992), "The Market for Second-hand ships: some results on efficiency using cointegration," Maritime Policy and Management, Vol. 19, No. 1, 31-39.  ^  P. R. Marlow (1991) "Shipping and Investment Incentives: A Trilogy, Part 2. Investment Incentives for Shipping," Maritime Policy and Management Volume 18 No. 3, 201-216.  14 variables other than those applied in Mr. Beenstock's model such as world seaborne trade volumes, prices of main cargoes and so on.  Fourth, there is no empirical application of his theory though the author is hopeful that "some basic analytical building blocks have been provided which may be helpful in the context of the empirical analysis of ship prices."  Fifthly, very notably, Beenstock's theory is based on a crucial assumption that new and second-hand ship pries are perfectly correlated, i.e. new and second-hand prices move in unison but second-hand prices will be at a discount reflecting simple depreciation. Chapter III reveals that the secondhand and new prices tended to move in the same direction but their ratios change over time. Chapter IV further proves that their correlation is positive, but not even near perfect. Therefore, Beenstock's assumption is over-simplified, which might lead to misleading and inaccurate outcomes of the model. What is more, by assuming perfect correlation, the complexity with regards to the respective developments of both new and second-hand ship prices is deliberately left untouched.  Last, Beenstock's model focuses only on new ship prices. Neither systematic analysis nor possible application of the model is delivered to the market of second hand ships.  15 3. AN ECONOMETRIC MODEL OF SHIP PRICES PERFORMED BY CHAREMZA AND GRONICKI  Unlike Beenstock's theoretical capital theory approach, Charemza and Gronicki present the first version of the econometric model SHACDEM Ll in "An Econometric Model of World Shipping and Shipbuilding". 6 The model tries to quantitatively analyze supply and demand balances in the world shipping and shipbuilding markets. One submodel focuses on the dry cargo shipping market and dry cargo shipbuilding. The submodel is estimated using yearly index data from the period 1961-1977. The following two equations relevant to the ship prices are selected here.  FDS =^0.339 + 1.053 FDS 4 + 0.823 FDL - 0.0360 PB ^ (2.6) (11.4)^(4.4)  FDL = - 0.168 + 0.480 FDL 4 + 0.0404 PB + 1.534 @Qat'? ^ (2.5) (1.9)^(3.8)  FDS =^an index of voyage freight-rates, averaged over 1965-1966 = 1.0^(Norwegian Shipping News, recalculated).  6  ^  Charemza W. and M. Gronicki (1981), "An Econometric Model of World Shipping and Shipbuilding," Maritime Policy and Management, Vol.8, No. 1, 21-30  16 FDL =^an index of time freight-rates, averaged over 19651966 = 1.0 (Norwegian Shipping News, recalculated). PB  =  an average price of a bulk carrier, 70,000 (in the period 1961-1966 an average price of a 30,000 dwt bulk carrier recalculated to be a proxy of a 70,000 dwt bulk carrier), in millions of current dollars (Alcan Shipping and Fearnley & Egers, recalculated).  @QWT =^change of dry-cargo shipments in 1,000 million tmiles (Fearnley & Egers).  Worth notice here in the model is the relationship between the voyage freight rate index and the ship price versus the relationship between time charter rate index and the ship price. The former is negative whereas the latter is positive. The inconsistency might cast some doubts on the reliability of the model. Though it is not the task of this paper to test this inconsistency, the intuitive understanding from practical point of view is that both relationships should have the some directions of effects on the ship prices.  The model so built is purely an econometric analysis towards solving real complicated problems related to the shipping market. As there are so many variables, one being possibly the endogenous or exogenous variable of the other, statistical selection of the so-called significant variables (for example, t-value is over 2.0)  17 is somewhat blind. A good understanding of what is really happening in practice in order to find out the causal relationships between variables is indispensable in terms of choosing the right variables. Simply searching for econometric relationships amongst all the variables appears very sophisticated but the outcomes can sometimes be very misleading or even ridiculous.  The SHACDEM L1 deals with the overall shipping market without differentiating different aspects such as container shipping, dry bulk shipping, tanker shipping and so on. As each aspect has its unique characteristics which need unique treatments, the homogenous treatment can only give some generic and broad insights about the overall shipping market.  4. BRIEF COMPARISON WITH THE APPROACH DEVELOPED IN THIS PAPER  The approach developed in this paper is fundamentally different from Beenstock's capital theory approach. Bearing all the problems related to Beenstock's model, this paper focuses on finding out those variables in both the freight and ship markets that tend to affect the secondhand ship prices directly and significantly, and then exploring the statistical relationship between those variables and the secondhand prices. Another big difference is that, in order to avoid treating heterogenous ships homogenous as so in Beenstock's study, the whole dry cargo market is divided into 3 categories, lakesize, handymax and panamax each of which consists  18 of almost homogenous ships. Handymax is selected as the data base for this study.  Fundamental difference also exists between the approach of Charemza and Gronicki and that of this paper in that the former is based on classic demand and supply analysis whereas the latter is based on direct selection of statistical variables. Apart from this difference and the same problem as in Beenstock's study with regard to the homogenous treatment of all different ships in the market, other main differences lie in that (i) this study deals with dry bulk carriers whereas Charemza and Gronicki covers the whole ship market; (ii) the statistical outcomes of this study are mostly based on individual observation records whereas Charemza and Gronicki's outcomes all result from a data base consisting of only yearly index, in which case the accuracy of the outcomes is doubtful;' (iii) the selection of relevant variables by Charemza and Gronicki is a pure statistical process other than investigating the true casual relationship between the dependent variable and independent variable from both practical and economic points of view.  7  ^  Proctor I.L. (1968) A Statistical Investigation of the Ocean Charter Market, UBC Commerce MBA graduation thesis, 158-160  19 CHAPTER III  INTRODUCTION OF HANDYMAX DRY BULK SHIPPING MARKET  1.  INTRODUCTION  This chapter describes the general structure of the handymax dry bulk shipping market. The overall world dry bulk shipping fleet is described first before focusing on the handymax dry bulk carriers. To follow the normal tradition of market analysis, the market structure, and then the demand and supply of the handymax dry bulk shipping market are examined in detail. A description of the handymax dry bulk shipping market over the last 10 years concludes the chapter.  2.  A BRIEF INTRODUCTION OF OVERALL DRY BULK CARRIER FLEET  Figure 3.1 shows the size of the world dry bulk carrier fleet over the past 10 years. The fleet size increased by twenty percent between 1982 and 1991. There was, though, a sharp increase of supply into the market during 1985 and 1986 as a result of deliveries of the newbuilding orders placed in 1982 and 1983 when the market was booming. The fleet was kept at the same level, a bit lower than the 1985 and 1986 level, mainly through demolition in 1987 and 1988.  20  Figure 3.1  R  < CARR  S  ES  1982-'1991 Cmi 1 I ion DWI ) -  250  200  150  100  50  0 1982  1983  1984  1985  1986^1987  1988  1999  1990,  1991  YEAR Source. Drewry Shipping Consultants  As shown by Figure 3.2, on June 30, 1991, the world handymax bulk carrier fleet constituted the largest proportion of the overall world dry bulk carrier fleet, which reflects the obvious importance of handymax fleet in the whole fleet. The second largest sector is the panamax bulk carrier fleet which accounts for 24.0% or just below one quarter of the total dry bulk fleet which was 210.3 million dwt on June 30, 1991. The third largest sector in terms of deadweight tons is the lakesize bulk carriers. These three sectors covered in this paper consist of about 75% of the total dry bulk fleet in deadweight tons. The capesize sector accounts for the  Figure 3.2  21  remaining 25%.  3. IDENTIFICATION OF HANDYMAX BULK CARRIERS  3.1 PRINCIPAL DIMENSIONS OF THE HANDYMAX BULK CARRIER FLEET  The ability of handymax bulk carriers to trade widely, calling at many ports with a variety of bulk cargoes, is one of the main  22 attractions of handymax-sized tonnage. The particular designs of their length, beam and draft make this versatility possible.  HANDYMAX LENGTH  1,034 vessels out of the total 1,454 (on January 1, 1991) have overall lengths between 180 and 200 meters meeting regulations of the most important transit ways such as the Panama Canal and St. Lawrence Seaway locks 8 . The handymax class is capable of using all major bulk ports in terms of the draft being the main determinant of the size of ship able to berth and load or unload bulk cargo.  HANDYMAX BEAM  Handymax bulk carriers, with very few exceptions, have a beam measurement of less than 32 meters, allowing them to transit the Panama Canal where the maximum beam is 32.3 meters with length overall and draft being maximum 274.3 and 12.03 meters respectively.  Handymax bulk carriers have a beam measurement that in general lies between 24 and 32 meters. For ships entering the St. Lawrence Seaway, the maximum allowable beam is 23.16 meters. Because of this constraint, the number of handymax vessels that are thus able to trade into the Great Lakes of North America is relatively small. 8^  Source: Drewry Shipping Consultants.  23 HANDYMAX DRAFT  Most handymax bulk carriers have a laden summer draft of between 10.5 and 12 meters. Therefore, generally passing through the Panama Canal does not constitute a problem but it does with the St. Lawrence Seaway where the draft limit is only 7.92 meters. As a matter of fact, handymax bulk carriers which fit the other dimension requirements will invariably undertake the passage of the Seaway locks part-laden, and complete loading to their full draft a process known as "topping-off" - at ports further down the estuary.  3.2^TYPES OF HANDYMAX BULK CARRIERS  TABLE 3.1 HANDYMAX DRY BULK CARRIER FLEET BY TYPE AT 1.1.^1991 NUMBER  '000 DWT  General Purpose Vessels  845  32,824  58.1  Specialised Vessels Of Which Wide Hatch Access Carriers Gantry Craned Carriers Woodchip Carriers Car/Bulk Carriers Cement Carriers Self-unloading Carriers Ore Carriers  609  22,297  40.5  382 81 63 67 7 5 4  13,527 3,232 2,610 2,317 275 199 137  24.5 5.9 4.7 4.2 0.5 0.4 0.2  1,454  55,121  100.0  Total  Source: Drewry Shipping Consultants Ltd. (1991)  DWT (% OF FLEET)  24 In general, there are two types of handymax bulk carriers: general purpose and specialised. Table 3.1 shows the composition of handymax bulk carriers by type on January 1, 1991. General purpose vessels account for 58.1% of the overall handymax fleet. Of the remaining 40% for the specialized type, the wide hatch access carriers  9  account for about 24.5% in terms of dwt.  GENERAL PURPOSE HANDYMAX BULK CARRIERS  The general purpose handymax bulk carriers are suitable for operation in a wide range of cargoes such as agribulks, coal, ores and minerals, fertiliser materials, etc and in some cases for "semi-bulks" such as steel products and packaged timber.  The design of general purpose handymax class will not be such that it limits their trading in specific sectors. However, the ships are not of uniform design. There are a variety of different designs, offering variations in such design characteristics as: draft, length, beam, cubic capacity, number of size of holds, cargo handling gear, etc. Shipyards develop standard bulk carrier designs which they can conveniently build in series at competitive cost while at the same time allowing the owner a degree of flexibility within the specification. Examples of famous existing designs are Tsuneishi Tess-40 design and Oshima OS-40 design.  9  .^Generally log/lumber carriers but are also capable of loading containers.  25 Good general purpose designs should have the following attractions:  - More competitive cost as a newbuilding, - Potentially low operating and maintenance costs and better of availability of spares, - Characteristics widely acceptable such good fuel consumption and cargo handling gear to charterers and major bulk shipping operators, and - Increased probability of a good re-sale value.  SPECIALIZED TYPES OF HANDYMAX BULK CARRIER  Table 1.1 indicates that more than 40% - or 22.3 million dwt - of the handymax bulk carrier fleet is of a specialized design which means that each ship is specially built to take advantage of features for certain trades. However, in reality, many of the vessels have the flexibility to interchange between the specialised trade for which they were designed and other traded commodities as and when market conditions dictate such a strategy.  As it has already been pointed out, over half out of the specialized type are wide hatch access carriers. Such designs always have container-fitted features. The main cargoes of these  26 vessels are sawn lumber and containers, but pulp, newsprint or paper board are often combined with the timber cargo.' °  4. DEMAND FOR HANDYMAX BULK CARRIERS  World dry bulk trade where the demand for handymax bulk carriers is derived is traditionally separated into two basic groups of commodities, "major" and "minor" bulk trades. Major dry bulk trades include iron ore, coal, grain, bauxite/alumina and phosphate rock; whereas minor dry bulk trades include manufactured fertilisers, potash, sulphur, tapioca, forest products, sugar, gypsum, manganese ore, concentrates, salt, cement and petroleum coke.  4.1 THE MAJOR DRY BULK TRADES  The continued popularity of the handymax class of bulk carrier especially the flexible, multi-purpose designs outlined in Section 3 of this Chapter - is a measure of the broad-based, profitable employment available to ships of this size. A characteristic of this sector of the market is the diversity of employment available to owners, with handymax bulk carriers participating to a greater or less extent in virtually all the major or minor bulk trades.  Detailed analysis of each type of specialized handymax bulk carriers is beyond the scope of this paper. For details, one can refer to Drewry's "Handymax Bulk Carrier" (1991).  27 It is evident that in the previous decade there has been a contraction of trading opportunities in some traditional commodity trades and expansion in others. An increased emphasis on the minor bulk trades has been spotted; thus the handymax bulker is seen carrying a far wider range of commodities than larger bulk carriers such as Panamax and Cape size bulk carriers.  What is evident is that the operations and market for handymax bulk carriers is quite different from that of the larger bulk carries of Panamax size and upward, where future demand and productivity are dependent upon developments in the seaborne trade of relatively few commodities. TABLE 3.2 MAJOR BULK CARGOES CARRIED BY HANDYMAX BULK CARRIERS (Million Tonnes) 1980  1987  1988  1989  Iron Ore Coal Grain (1) Bauxite/Alumina Phosrock  30 58 130 30 41  28 68 102 26 34  24 67 109 27 37  25 74 96 28 34  Total  289  258  264  257  (1)^Including wheat, maize, barley, oats,^rye, sorghum and soya beans. Source: Fearnley Ltd, Oslo, Norway  28 TABLE 3.3 TOTAL WORLD SEABORNE TRADE IN THE MAJOR BULK CARGOES (Million Tonnes) 1980  1987  1988  1989  Iron Ore 314 318 344 365 Coal 188 285 309 310 Grain (1) 198 196 217 218 Bauxite/Alumina^48^45^47^49 Phosrock^48^42^44^41 Total  ^  796^886^961^983  (1) Including wheat, maize, barley, oats, rye, sorghum and soya beans. Source: Drewry Shipping Consultants Ltd, London, England. Table 3.2 features the five major bulk trades: iron ore, hard coal, grain, bauxite/alumina and phosphate rock, and shows a proportion of each trade shipped by handymax bulk carriers. It is apparent there has been a stable and small decline in the proportion of major bulks traded in handymax bulk carriers. Table 3.3 indicates that the world sea-borne trades of major bulk commodities have increased over time. Using data in Table 3.2 and Table 3.3 jointly, the overall market share of handymax bulk carriers has declined substantially over time, from 36.3% in 1980 and 29.1% in 1987 to only 26.1% in 1989.  29 4.2 THE MINOR BULK TRADES  Requirements for the major bulks will have an important influence over the future requirement for handymax vessels. There are, however, a wide variety of other cargoes which are shipped by handymax bulk carriers, ranging from minor ores like manganese and chrome to manufactured products such as steel. Trading opportunities for 30-50,000 dwt ships will to an increasing extent depend upon the growth of trade in minor bulk commodities, as, collectively, the trade volumes moving by sea will be substantial. TABLE 3.4 TOTAL WORLD SEABORNE TRADE IN THE MINOR BULK CARGOES (Million Tonnes)  Manufactured Ferts Potash Sulphur Tapioca Forest Products Sugar Gypsum Manganese Ore Concentrates Salt Cement Petroleum Coke  1986  1987  1988  1989  40.5 13.4 10.6 7.6 147.0 24.6 12.1 8.3 8.9 16.5 39.7 15.5  46.2 19.2 11.0 5.7 149.0 27.0 12.5 6.7 9.9 17.0 43.5 14.1  49.4 20.0 13.4 7.5 157.0 27.9 13.5 8.2 9.5 17.6 41.0 15.1  50.0 17.0 10.3 9.1 153.0 28.1 13.5 8.4 9.3 17.7 40.0 15.5  Total Source: Drewry Shipping Consultants Ltd Table 3.4 shows the traded volumes on selected minor bulk routes. The majority of the bulk carriers employed in the transportation of  30 the minor bulks tend to be in the small-to-medium size range as the relatively small proportion of cargo is handled by ships above 50,000 dwt. In 1979, 79% of shipments of minor bulks - or some 480 million tonnes of trade - took place in vessels below 50,000 dwt.  11  5. SUPPLY OF HANDYMAX BULK CARRIERS  5.1 DEVELOPMENT OF THE HANDYMAX BULK CARRIER FLEET  The term 'supply of bulk carriers' refers to the available capacity for carrying cargoes in bulk from one or more ports to one or more ports. The handymax bulk carriers first came out of the world's shipyards in the early 1950s. Since that time, the growth of this sector of the world bulk carrier fleet has been almost continuous. Only in 1987 and 1988, was this trend interrupted, because the high level of scrapping which followed the freight market slump resulted in deletions exceeding new additions to the fleet. The fleet did not exceed the 1986 capacity until 1991.  Drewry Shipping Consultants Ltd. (1991) Handymax Bulk Carriers, Drewry Publishing Ltd, England, 37. Also for the analysis of the characteristics of the individual minor bulk trades, trade patterns and shipping volumes and role of the handymax bulk carrier.  31 TABLE 3.5 GROWTH OF THE HANDYMAX BULK FLEET (1970-91) ( in millions)  Year  Handymax Fleet million dwt  1970 1975 1980 1981 1982 1983 1984 1985 1986 1987 1988 1989 1990 1991  15.4 24.3 35.7 35.4 39.8 42.9 46.6 53.0 55.0 52.6 52.2 53.7 54.8 55.1  Total Bulk Fleet million dwt 63.1 98.5 138.2 140.6 158.8 170.6 179.3 191.6 197.3 190.7 190.7 196.8 203.4 207.0  Handymax percentage 24.4 24.6 25.8 25.2 25.1 25.1 26.0 27.7 27.9 27.6 27.4 27.3 27.0 26.6  Source: Drewry Shipping Consultants Ltd., London  From the Table 3.5, it is apparent that there has been a net increase of nearly 40 million dwt to the fleet of handymax bulk carriers since 1970. A considerable amount of this growth has occurred since 1980. Between 1980 and 1991, the handymax fleet witnessed a net addition of 19.4 million dwt. This was a 54% increase in fleet capacity.  Handymax bulk carriers were able to increase their percentage share of available tonnage in the world bulk fleet through the 1970s into the mid-1980s, as new tonnage was added. The increase in market share was largely a reflection of the progressive increase in shiploads in the bulk trades with handymax types gradually  32 replacing Lakesize (i.e. 10-30,000 dwt) and smaller (under 10,000 dwt) vessels.  However, a number of factors combined to result in a slowdown in the rate of growth of the handymax fleet in the period 1986 to 1988. They include: the increase in the price of newbuildings due in general to the steep depreciation of the dollar versus Japanese Yen; weak freight market conditions and rising scrap values which increased the incentive to dispose of vessels which under different market conditions may have continued trading. During 1987, there was a significant reduction in handymax tonnage - a trend evident in 1988 although at a much reduced level. With only 12 handymax vessels on order, newbuilding activity was extremely low in 1986, and this trend continued in 1987 and 1988  12  . A resurgence in  activity was evident, however, in 1989 when 58 vessels were contracted. This particular year proved to be a good one for owners and operators as both freight revenues and vessel values maintained high levels. The optimism of shipowners associated with this revealed a resurface of contracting activity, that led to prices of newbuildings reaching substantially higher levels - a trend which continued into 1990.  During 1989 the dry bulk freight strengthened, and with revenues and fleet utilisation high, this led to a certain amount of  Drewry Shipping Consultants (1991) Handymax Bulk Carriers, Drewry Publishing Ltd, England, 10.  33 optimism about the future and the profit potential of newbuilding. However, by early 1990 the effects of slowdown in the world economy had started to manifest itself, resulting in a downturn in the dry bulk market.  5.2 THE MARKET ROLE OF HANDYMAX FLEET  To a great extent, the evolution of the handymax bulk carrier is the result of its changing market role - a product of an extremely diverse range of commodity trades in which the handymax class bulker is known to participate. It is difficult or simply impossible to obtain exact statistics with regard to the involvement of handymax bulk carriers in terms of trading opportunities as substitution can very easily be made between Lakesize and Panamax or other classes.  A critical influence on the requirements for various vessel size classes in the dry bulk trades has been - and will continue to be the overall increase in the size of cargo consignments throughout most of the seaborne commodity trades. It is this trend that without doubt has shaped demand for handymax fleet in the major and minor bulk trades.  As pointed out in section 4 of this chapter, the relative importance of handymax fleet in the ocean transport of the major bulk commodities has been declining. In terms of the movement of  34 iron ore, handymax involvement declined substantially in the early 1970s leaving ships of this size with a minor role. The residual level of employment open to the handymax carriers has recently shown evidence of further erosion, as their substitutes Panamax bulk carriers are now being utilized for more iron ore shipments.  5.3 HANDYMAX BULK CARRIER FLEET JANUARY 1, 1991  The world handymax bulk carrier fleet numbered 1,454 units with a total carrying capacity of more than 55 million dwt, as at 1 January, 1991. The handymax sector contains 31% of the total number of vessels in the world bulk carrier fleet above 10,000 dwt, accounting for about 27% of the total dry bulk fleet tonnage. TABLE 3.6 HANDYMAX DRY BULK CARRIER FLEET BY SIZE AT 1 JANUARY 1991 (tonnage in million dwt) SIZE^('000 dwt)  NUMBER  PERCENT  TONNAGE  30-35 35-40 40-45 45-50  450 558 341 105  30.9 38.4 23.5 7.2  14.7 21.0 14.5 4.9  26.7 38.1 26.3 8.9  Total  1,454  100.0  55.1  100.0  PERCENT  Source: Drewry Shipping Consultants Ltd. Table 3.6 provides a breakdown of the handymax bulk carrier fleet by size. The fleet is concentrated in the three sectors below 45,000 dwt; these contain 91% of the tonnage and 93% of the number.  ^  35 The 35-40,000 dwt size range has the largest concentration of the vessels, comprising 450 or 30% of the vessels in service at the beginning of 1991 and 38% of the total tonnage within the handymax range. The 30-35,000 dwt size band has the next largest concentration of ships, making up nearly 31% of the handymax fleet numbers and 28% of the total tonnage. The 40-45,000 dwt size category ranks next in order of importance, with 341 ships amounting to 14.5 million dwt. As Table 3.6 indicates, the smallest sector is the 45-50,000 dwt range which, in terms of the total handymax fleet, only accounts for 7% of the vessels and 9% of the total dwt in this sector. TABLE 3.7 HANDYMAX DRY BULK CARRIER FLEET BY AGE AT 1 JANUARY 1991 (tonnage in '000 dwt) AGE SIZE^0-4^5-9^10-14^15-19^20+^TOTAL ('000 DWT) NO. DWT NO. DWT NO. DWT NO. DWT NO. DWT NO. DWT 30-35^13^429 131 35-40^38 1461 261 40-45^46 1948 186 45-50^13^604^48  4282^87 2856 158 5174^61 1677 450 14720 9836 137 5168 ^86 3214^36 1365 558 21034 7884^48 2059^28 1165^33 1397 341 14463 2243^12^554^15^706^17^797 105 4904  Total^110 4452 626 24247 284 13637 287 10259 ^147 5236 1454 55121 PERCENT^7.5 8.1 43.1 44.4 19.5 24.7 19.7 18.6 10.1 9.5 100^100 Source: Derived from Drewry Shipping Consultants Ltd.  The data in Table 3.7 provide a general outline of the age profile of the handymax bulk carrier fleet. The most important feature of the above distribution is that ships under 10 years of age account for about 50% of the total ships in terms of both tonnage and number of ships. This indicates the handymax fleet is relatively  36 modern and young. The largest section of the fleet is formed by vessels between 5 and 9 years of age, ships of this vintage representing 43% of the total number of ships in the handymax fleet. Only 10% of the ships are more than 20 years of age. Older ships form a much higher percentage of the 10-30,000 dwt sector than of other sectors of the handymax bulk carrier fleet.  It should be emphasised that only 7.5% of the total fleet is less than 4 years of age, signifying a marked slowing of newbuilding activity in the past five years. However as has been noticed, the newbuilding of handymax bulk carriers since the beginning of 1991 have been stimulated substantially. Further discussion of this is in Chapter III.  A substantial proportion of the vessels in the handymax bulk carrier fleet, no less than 36%, is between 35-45,000 dwt and is less than ten years of age. There is evidence to suggest that there has been a movement away from standard designs of bulk carrier at the lower end of the handymax range - that is, 30-35,000 dwt - and towards 40-45,000 dwt vessels which maximise dwt capacity at their design draft. The fleet age analysis reveals that over two-thirds 68% - of the total number of vessels in the 40-45,000 dwt band are under ten years of age and of the vessels less than four years of age, 42% are in this dwt range. Therefore, many of the most modern bulk carriers in the fleet fall into in the 40-45,000 dwt size category. Two reasons might be suggested for the above change. The  37 first one might be related to the increasing demand for 40-45,000 dwt tonnage arising from cargo volume structure which makes this size more economical than 30-35,000 dwt range. The second reason could lie in the modern ship design change in the direction of achieving maximum carrying capacity at certain draft. The most popular designs are TESS-40 series of Tsuneishi Shipyard and OS-40 series of Oshima Shipyard in Japan which are both about 43,000 dwt bulk carriers standard designs.  6. HANDYMAX FREIGHT MARKET AND REVIEW 1981 1991 -  Dry bulk shipping is an industry that has a market which functions under conditions that are not dissimilar to the theoretical model of perfect competition. ° As a usual competitive market, the dry cargo freight market consists of many participants on the demand and supply sides who interact to determine the level of freight rates.  Some argue that the demand for bulk cargo tonnage is relatively insensitive to the freight level, indicating that the elasticity of demand to the freight level must be very small. In other words, the demand curve will be almost a vertical line. 14 On the other hand, others argue that the demand for bulk cargo tonnage is very 13  Metaxas B.N (1971) The Economics of Tramp Shipping, The Athlone Press of the University of London, 19-20  14  Shimojo T. (1979) Economic Analysis of Shipping Freights, Hayashi Obundo Printing Co., Ltd., Kobe, Japan, 77-78.  38 sensitive to the freight level especially when the proportion of the freight cost is high in the market price of the final product of the cargo and there is no alternative means of transportation. Then, the demand curve would be almost horizonta1.  15  The common  consensus is somewhere between these two extremes, depending on the factors that derive the demand for bulk shipping. Nevertheless, the writer feels that there exists a distinction between the short-run demand curve and long-run demand curve. In short run, the demand curve tends to be steeper as the importers and exporters might be unable to find ways to avoid high rates whereas the owners might have to live with the depressed rates. However, in long-run, the situation would be different. Shippers and owners have time and means to adjust the structure of their trade, their manufacturing and so on, and as such the long-run demand curve tends to be flat relatively.  In the short-run, the supply of dry bulk tonnage is elastic to the freight level when the freight level is low or moderate, and inelastic when the freight level is high. The owners can adjust themselves economically to the market freight level through layingup their tonnage or reactivating their laid-up tonnage-up. In the long-run, the owners can increase or decrease their tonnage by means of scrappage and new orders. A typical short-run supply curve stays flat for low freight levels and then gradually becomes steep  15 ^ .  Metaxas B.N. (1971) The Economics of Tramp Shipping, The Athlone Press of the University of London, 44-45  39 after the market freight rises above the laid-up level. The supply can not be greater than the total existing tonnage available as owners are unable to increase their capacity in the short-run through new orders which usually need about 2 to 3 years of leadtime. In the long-run, the owners have time to adjust their capacity in accordance with the freight level. The higher the freight rate expectations, the higher the supply. Therefore, the long-run supply curve should be of increasing form.  In general, the short-run demand and supply for handymax dry bulk tonnage determine the short-run freight including voyage rates, trip time charter rates and even rates for period charter with relatively considerable length. Similarly, the charter rates for long-period charters are set by long-run demand and supply. Though it is not the subject of this thesis, it would be interesting to find out relationship between long-run and short-run conditions.  Figure 3.3 depicts the development of representative one-year time charter rates of the three major sizes of bulk carriers for the past 10 years. For the beginning years of 1980s, the time charter rates were fluctuating moderately. However, the market began falling at the end of 1984. This was mainly due to a sharp increase of the bulk carrier supply  16  and low demand for bulk carriers. The  slump of 1985 and 1986, forced many shipowners into bankruptcy because of high costs and low revenues (charter rates). However, 16  •  ^  See section 2 of Chapter 2.  40 things began picking up in 1987 when the demand for bulk carriers was encouraged by the increases of both major and minor dry bulk cargoes and the decrease of the world bulk carrier fleet through scrappage during 1985 and 1986. The market reached its top in 1989 when most owners made huge operating profits. During 1990 and 1991, the time charter market was at high level, which was a signal to a lot of ship investors that there would be no force (from either demand or supply or combination) to drive the market down. A high volume of transactions of ship sale and purchase took place during this two years despite a certain degree of disturbance created by the short Gulf war. Figure 3.3  TIM  ^  ART ER RATES 1Y82-199' BASIS 1-YEAR TIME CHAPTER  11  10  9  8  7  6  5  4  3  2 1982  ^  1983  ^  1984  ^  1985  ^  1986^1987 YEAR  -  10 - 30, 00 0 DWT BULK CARRIERS^  -  50- B0,000 DWT BULK CARRI ERS  ^  1988  ^  1989  ^  1990  30 - 50,000 DWT BULK CARRIERS  ^  1991  41 CHAPTER IV  SHIP SALE AND PURCHASE MARKET  1.  INTRODUCTION  Chapter II reviews the handymax dry bulk shipping market. This chapter focuses on the ship sale and purchase market, its mechanism and decision-making process. It also provides a review on the sale and purchase market of the past 10 years.  2.  INTRODUCTION OF SHIP EXCHANGE MARKET STRUCTURE  There are three main players in this market: buyer, seller and broker. Normally a buyer and a seller do not work directly for a deal. It is the broker who works as a match-maker between the seller and the buyer and puts the detail of the deal together.  When an owner has a ship for sale, he or she will notify his or her broker or brokers to "put the ship in the market for sale". The broker(s) will circulate the main specifications of the ship together with a suggested price and brief description of the current employment to various brokers all over the world. Places where the ship is inspectable are also provided. The ship is now "in the market".  42 The broker(s) report responses from the buyer(s) to the owner. If a buyer shows interests in this ship, he or she will first ask via his broker for permission for inspection of the ship at a convenient port of call along her itinerary. The inspector is normally the buyer's own technician or a professional ship surveyor. If the ship is found in satisfactory conditions or the right type for the buyer, the buyer might present an offer which might be outright or with subject clauses.  When an owner goes to the market to buy a ship, he or she will ask his or her broker or brokers to "put the purchase enquiry in the market". Quite the same as the above, the broker(s) will circulate the main requirements of the ship to purchase to various brokers all over the world. The buyer is said to be "in the market" for a certain type of ship(s). The broker(s) provide various candidates to the buyer for discussion. Once suitable ships are chosen, the next thing to do is to ask the sellers to grant permission for inspection. The inspection results would determine whether or not the buyer would proceed with an offer.  In most cases, the owner would appoint an exclusive broker who is to market and sell the ship. However, sometimes, the owner might ask several brokers to market and sell the ship. The brokers can be either local or anywhere in the world, depending on the relationship between the owner and the broker(s). The same broker appointment system applies to buying a ship.  43 The biggest world ship sale and purchase centre is still London which has traditionally been the world's most important sale and purchase centre for over 300 years. Oslo, New York, Hong Kong and Tokyo are becoming more and more important.  There are numerous ship sale and purchase brokers all over the world. Normally if a deal goes through, each broker earns a commission equal to 1.0% (one percent) of the ship purchase price. If a ship is sold for USD10.0 million, the broker earns USD100,000.00 which is quite substantial as almost no capital investment is needed for shipbrokering.  3. SHIP SALE AND PURCHASE BARGAINING PROCESS  Like a small fair, buyers and sellers bargain with each other on prices as well as on other terms and conditions. The difference is they normally do not hold talks face-to-face but through their brokers by fax, telex and telephone.  The negotiations normally take the form of offer and counter offer. The initial offer usually consists of price, deposit, place and time for delivery, drydocking, method of payment, ship classification and subjects. Those main terms and conditions are the basis for the sale contract, namely Memorandum of Agreement, in brief, MOA. The most popular sale contracts are Norwegian Saleform 1987 MOA and Nippon Saleform 1977 MOA.  44  Negotiations are mainly centred around those main terms and conditions. The number of counter offers depends on the specifics of each case. When price and main terms and conditions are accepted by both parties, the ship is said to be committed, in which case the brokers will notify the market. MOA will be drafted by the Sellers' broker accordingly for further discussion and confirmation.  When MOA is agreed and signed, the next thing to do is to make a deposit which is normally 10% of the total purchase price. The deposit is paid to a joint account at a bank nominated by the seller under the names of both seller and buyer. The deposit will be held there until the delivery of the ship at which time the deposit is released to the seller as part of the sale price. The interests earned are normally to be for the buyer.  90% of the sale price is paid upon exchange of the bill of sale and protocol of delivery signed both seller's and buyer's authorized representatives at the site of delivery.  A sale is completed when all the money is paid and the ship is delivered.  45 4. THE SALE AND PURCHASE DECISION-MAKING  4.1 THE BUYERS  In the secondhand market it is possible to identify four typical types of buyers. The chart below provides a summary of the principal buyers.  GOVERNMENT BUYERS  PROPRIETARY FLEET OPERATORS  RELUCTANT BUYERS  PROSPECTIVE BUYERS  SPECULATIVE TRADERS  1. Speculative traders looking at asset appreciation. Such buyers are normally uninterested in complexities of trading and are likely to place the vessel under the control of a management company. An investor may even allow the ship to operate at a loss if of the opinion that any trading loss will be outweighed by the profit on the capital investment. Alternatively, the investor may place the ship in lay-up (i.e. effectively reducing expenditure solely to the needs of  46 insurance, care, crew costs and maintenance) with the intention of selling when the freight market improves.  2.  Proprietary fleet operators would see ship acquisition as more desirable than chartering since transportation forms a part of their integrated production and distribution operation and ownership allows them to maintain direct control over their trading costs.  3.  Government buyers buy the ships to meet needs of bilateral or multi-lateral trade agreements, conserve freight exchange through use of the own ships and/or generate foreign exchange from chartering out state vessels.' Good example is the China Shipping Company Ltd. (COSCO) whose fleet is constantly expanded through acquisitions of either newbuilding ships or used ships.  4.^Reluctant buyers - normally mortgagees or foreclosing banks dealing with distress sales. They are forced to "buy" ships when the owners go bankrupt.  Of the above, the first three types of buyers have alternatives to secondhand ship acquisition. Speculative buyers have the option to place a newbuilding. Proprietary fleet operators and government 17 . ^  Ademuni-Odeke (1988) Shipping in International Trade Relations, Avebury Gower Publishing Company Limited, England, 234-235  47 buyers have other options such as chartering necessary tonnage. The prospective buyer, therefore, in deciding whether to enter the secondhand market, would consider the merits and disadvantages of secondhand ship acquisition.  Advantages of Secondhand Acquisition  The buyer will obtain a vessel at a fixed price which may be considerably lower per dwt than an equivalent new vessel, provided the secondhand ship is not too young.  On completion of the purchase deal the vessel is basically available for service commencement. Even if the buyer has to undertake alterations to enhance the trading prospects of the vessel, service entry will only be delayed for a relatively short period. Given a quick delivery date, the vessel can soon be introduced to meet market demand. This is significant in cash flow terms as the vessel starts to earn money sooner than a new vessel which may have a lead time of a few years.  Furthermore, the amount of capital required is lower than for a corresponding newbuilding order and the prospect of immediate trading eases the burden of loan repayment". For financiers, there  B.  However, because of the ability to put the vessel into service almost immediately, secondhand vessels may sell for higher prices than that of newbuilding when the freight market is booming.  48 appears to be greater security in the secondhand market than in the newbuilding, particularly as they are dealing with the immediate term rather than possibilities a few years ahead.  Finally, secondhand tonnage is beneficial if starting a new service, as it enables the operator to test the market in low capital risk than with a newbuilding. In the event of the service proving successful, new tonnage can be introduced.  Disadvantages of Secondhand Acquisition  Financiers may impose stringent conditions prior to providing the requisite finance. These could include limitation on other investments, disposals from the existing fleet, all earnings to be paid into an account at the lending bank, maintenance of corporate liquidity, a preference for borrowers with an existing diversified fleet.  Secondhand tonnage tends to have higher operating costs than new vessels. These costs include crew wages, supplies, insurance and administration expenditures, and periodic expenses like drydocking and survey costs. As a vessel ages, its operating life shortens and maintenance/survey/operating costs will be higher. For example, ship insurance premiums will increase with age, particularly for vessels over fifteen years old.  49 A buyer may have difficulty finding a vessel with the requisite specifications - the vessel may have slow speed, limited cargo capacity, poor cargo transshipment facilities, etc. In such circumstances the buyer would undertake modifications; a new selfunloading system, for example, may need to be installed in a bulk carrier to let her self-unload the cargo very efficiently. A typical example would be that a couple of Canadian shipping companies (Canadian Steamship Line and Upper Lakes Shipping) have been seeking to purchase less-than-5-years-old panamax bulk carriers for conversion. However due to the requirement for depth being minimum 19 meters for the installation of self-unloading system, they can hardly find a suitable candidate. When they finally decided to turn to newbuildings, they found the newbuilding price is "considerably higher than anticipated"  19  .  While a secondhand ship may be cheaper to purchase than a new ship, financial aid for a new ship such as building or shipyard subsidies or grants will probably be generous and readily available. In addition, in many countries the tax relief benefits on secondhand tonnage tend to be less favourable than new tonnage.  20  19. In February, 1992, Upper Lakes Shipping of Canada wrote to China State Shipbuilding Corporation in Beijing that the price offered for a panamax self-unloading dry bulk carrier was considerably higher than anticipated. As a result, they decided not to place the newbuilding order. m.^Branch A.E. (1988) Economics of Shipping Practice and Management, Chapman and Hall, England, 59-62  50 4.2 THE SELLERS  So far the discussion has focused on the demand side of the market without considering factors influencing the selling decision and hence the market supply of tonnage.  1.  To recoup a profit over book value or initial cost. Owners for some reason may wish to make their financial statements look better. One of the common ways is to sell ships which have market values higher than the book value or initial cost. The profit generated from selling ships is sometimes very vital for owners who especially have not done well in operating the ships. If the owners are public companies, such excessive capital gain might help to strengthen or sustain their stock prices.  2.  To realise current benefits if the market trend points to the available sale price being in excess of the benefits of perceived future trading income 21 . One important factor here is that the ship maintenance costs and expenses tend to increase as the ship is getting older. For certain owners who have specific replacement policies such as to keep their fleets at favourable age structures and sizes as part of their long term strategic planning, they may think a particular ship n.^Also see Murray Frank and lain Cockburn Market Conditions and Retirement of Physical Capital: Evidence From Oil Tankers November 24, 1991, Seminar, Commerce of UBC.  51 is just old enough to sell, allowing for market sale price, of course. However, as the author has observed, the actual decision-making process is very complicated, involving substantial cost-benefit analysis, since capital associated with a ship is usually in millions of US dollars.  3.  The shipowner may be facing a distress or enforced sale. The latter can result from arrest for non-payment of bunkers, dues, etc. or through foreclosures by prime mortgagees. Such cases are observed when shipping is in recession during which freights revenues can not cover costs resulting in financial problems. Ships sold for such reason are usually sold at discounted prices via either public bidding or private tender.  4.  A vessel type may become redundant within an owner's fleet because of a shift in company policy - say, to cease operating in a certain trade for reasons including lack of cargo, inflated operating costs and so on. Alternatively, a shipowner may be better served by disposing of an existing vessel and replacing it with a new vessel which incorporates modern design, equipment, fixtures and fittings.  S. MARKET DEMAND AND SUPPLY AND PRICE DETERMINATION  The fundamental determinants of price are the supply and demand position in a vessel's trading market sector. It is more than  52 likely, therefore, that secondhand values respond to and mirror the pattern exhibited by freight rates which are mainly determined by the combined forces of demand and supply.  22  The vessel itself, however, is also a major determinant of price. In general, secondhand prices decrease with age. The older the vessel, the greater the wear and tear, prospect of mechanical breakdowns, fuel consumption, and the shorter the remaining work life and payback and profit-generating period. An exception could arise where age relates to survey status, the value of a vessel having just passed its survey (particularly, a major special survey) may be greater than one where a survey is due.  Few ships have identical features; some differences can have a marked effect on their effectiveness in particular trades. Prices reflect those differences. For instance, buyers seeking a handymax bulk carrier may pay a premium for a vessel which is fitted to handle containers.  Also notably, vessels suitable to transit canals and seaways may command higher values than slightly larger units. The latter's additional capacity often fails to compensate for added voyage expenses associated with alternative routes.  n.^Drewry Shipping Consultants (1990) Bulk Fleet Growth and Tonnage Supply in the 1990s, Drewry Publishing Ltd, P61.  53 Some shipbuilders have a reputation for "problem ships" - this may be related to building standards and use of poor quality materials, especially in some socialist countries. On the other hand, vessels identifiable as good design and built by first-class yards may command a premium above normal market price. The difference is always made in the sale and purchase market between ships built in Socialist countries and those built in western yards including Japan, Korea and Taiwan.  The number, record and reputation of a vessel's previous owners have an effect on price. Previous employment affects levels of wear and tear, whether repair and maintenance work has been kept to a minimum, and whether the vessel is a continuously surveyed unit. Sometimes, the vessel's price is discounted if a vessel's main engine has breakdown or other damage records which can easily be discovered from class records inspection at the classification society which the ship belongs to.  Finally, other factors include: reactivation costs if the vessel is laid-up, whether the vessel is being sold with charter commitments; if the transaction is of a judicial nature, for example, the vessel bought free of all mortgagees and liens.  54  Figure 4.1 HANDYMAX DRY BULK CARRIERS PRICES 1981-1991 NEWBUILDING VS. SECONDHAND 1100 1000 900 800 700 600 500 400 300 200 100 0 1982  ^  1983  ^  1984  ^  1985  ^  1986^1987  ^  1986  ^  1989  ^  1990  ^  1991  YEAR _w_Handymax Newbuilding Price _._Handymax Secondhand Price Note 5 years old for secondhand ships,  5.1 UPPER LIMIT - NEWBUILDING PRICE PLUS IMMEDIACY PREMIUM  The upper boundary on the secondhand value of a vessel is the newbuilding price for that type of ship adjusted by a premium that might be paid for the immediacy of delivery offered by a trading vessel of very young age when there is a strong market. If owners believe the gap between new and secondhand prices is too narrow, they will choose newbuilding as a better option, taking into account the necessary leadtime for newbuildings.  55 Figure 4.1 provides an illustration of newbuilding and secondhand price (5 years old) trends for representative sizes of bulk carriers for a period from 1982 to 1991. It is obvious that both secondhand and newbuilding prices were declining steadily from 1982 to 1986. It was in 1986 when the ship market reached the trough and started to climb up. In 1982, a representative 5-year-old handymax bulker could be bought at about USD225 per dwt, whereas the corresponding newbuilding price was about USD900 per dwt. In 1986, the prices dropped to only USD105 per dwt and USD450 per dwt or by about 53.3% and 50.0% respectively. Since 1986, the market had been rising quite substantially. The market rising trend slowed down in 1989. In 1989, the newbuilding price was restored to almost the same level as that in 1982 whereas the secondhand price was restored to 1982 level about 2 years earlier. In 1990, both newbuilding and secondhand prices reached record high since 1980. 1991 followed the trend and marked a high start for the 1990s.  The ratio of secondhand to newbuilding prices in 1982 was about 25% as compared to only about 20% in 1986. During the period between 1982 and 1986, the newbuilding prices dropped relatively more than that of secondhand prices. Though both secondhand and newbuilding prices began climbing after 1986, secondhand values were rising faster than that of newbuilding prices until 1989. The ratio of secondhand to newbuilding prices reached about 45% in 1989, the highest level in 1980s. In terms of absolute values, there was a difference of around $9 million between the newbuilding and  56 secondhand prices of the representative 5-year-old Panamax bulk carrier in the first quarter of 1989. The secondhand values maintained the 1989 level throughout 1990 and 1991. The ratio actually narrowed to around 40% in 1991 as a result of the newbuilding price increase from about USD900 per dwt in 1989 to about USD1,050 per dwt in 1991.  The newbuilding prices fluctuated more than secondhand values in absolute values and in the long term in 1980s. However, it should be emphasized that the secondhand values are more volatile than those of the newbuildings in short term in terms of period of, say, a week or a month. Despite the obvious disparity in the development of the secondhand and newbuilding prices, the pair were roughly moving in the same directions.  5.2 LOWER LIMIT  -  SCRAP VALUE LESS POSITIONING COSTS  The price floor for secondhand ships is easily identifiable as the demolition market price of an undamaged ship with trading certificates adjusted by the positioning costs that are associated with the locations of the ship and the demolition site.  5.3 FLUCTUATIONS IN THE PRICES  Prices fixed in the course of interactions between supply and demand in the sale and purchase market have an unstable character.  57 They often fluctuate. Fluctuations of prices may take a form of sporadic changes or seasonal fluctuations.  SPORADIC CHANGES  Sporadic changes, also called irregular fluctuations, are short changes of ship prices. Short-run changes in supply and demand for ships are the reason for these fluctuations which rarely last longer than a period of a few months. The reasons for sporadic changes of supply and demand for ships may be of a political nature (war psychosis, political crisis and unrest, etc.), or of a speculative nature (gaining of extraordinary capital profits through buying or selling). Generally, these brief fluctuations of prices have a limited influence on the shipowners investment policy since the period of these changes is too short, and it would not be wise for the owners to introduce far-reaching changes in their tonnage and their operations.  Those shipowners who have diversified fleets and who pursue a speculative operating policy can be favoured or hurt by sporadic fluctuations more than the carriers with long-term shipping policy, depending whether they can accurately forecast the future market development.  Though sporadic changes of prices have an impact on the general situation of the sale and purchase market the forces behind the  58 fluctuations are not, however, determinants of long term market trends.  SEASONAL FLUCTUATIONS  These are fluctuations which are repeated at certain periods of the year. The causes of such fluctuations lie both in the climatic conditions (e.g. freezing of certain ports, harvest season for certain commodities, increased demand for energy in winter etc.) and in the nature of commercial transactions (e.g. completion of transactions before the end of the financial year for PR Chinese buyers). Compared to sporadic changes, seasonal fluctuations may have a more stable character. Seasonal fluctuations may also exert influence upon the shipowners' investment policy. Particular shipowners try to order ships whose parameters enable them to carry alternative cargoes and find employment outside seasonal shipments. Multipurpose ships, having a high degree of versatility with respect to the nature of the cargo have their merit in this respect. Certainly, all negative consequences of seasonal fluctuations cannot be avoided. They favour shipowners with an flexible operating policy and penalise those who run highly specialised ships or who are unable to adapt themselves quickly to a changing market situation.  59 6. THE NEWBUILDING MARKET AND VESSEL DELIVERIES  TABLE 4.1 BULK CARRIER NEWBUILDING ORDERS BY SIZE 1985-1990 Size Range 85^86^87^88^89^90* ( M dwt)^No. dwt No. dwt No. dwt No. dwt No. dwt No. dwt 10-30 30-50 50-80  69 58 42  1.6 2.2 2.7  41 12 6  0.9 0.5 0.4  13 18 21  0.2 0.7 1.3  15 12 45  0.4 0.5 2.9  56 58 32  1.1 2.4 2.0  6 29 6  0.1 1.1 0.5  Total  169  6.8  59  1.8  52  2.4  72  3.8 146  5.5  41  1.7  * First six months only. Source: Drewry Shipping Consultants Ltd, London. So far, the discussion has focused on the secondhand sale and purchase market. The ship newbuilding market is inevitably correlated with the secondhand market. Table 4.1 indicates there were many more newbuilding orders placed during the year of 1985 than 1986 when there were just above one quarter the quantity of orders. However, new orders gradually rose throughout 1987 to 1989. The statistics for the first 6 months of 1990 show that the newbuilding orders began sliding. This general tendency applies to each type of dry bulk carriers.  When prospective shipowners wish to buy tonnage, they weigh up carefully the advantages and disadvantages of purchasing a newbuilding versus a secondhand ship.  60  6.1 ADVANTAGES OF NEWBUILDING  First of all, the benefit of purchasing a newbuilding is that it offers a longer trading life - often around twenty years.  A  further advantage of buying a newbuilding is that the shipowner  can specify the design of the ship, depending on the particular needs of the route it is intended to trade on. It must, however, be remembered that this can prove a disadvantage when it comes to the resale of the ship. The decision as to size and propelling machinery will be determined by the factors involved in the particular trade such as the nature of the cargo to be moved, the cost and availability of fuel and the minimum carrying capacity required. Thus, in the various bulk trades, the nature of the cargo tends to favour large vessels which can take advantage of the economies of scale.  When the future trading pattern of the vessel is less certain, the shipowner may tend to opt for a standard design ship and introduce modifications suitable for its purpose. The concept of the standard design ship is that the hull form, principal dimensions and basic characteristics for a certain type of ship are pre-designed for quick delivery from production line techniques. This will have cost advantages at the design and building stage and may, through market familiarity, assist in obtaining favourable charters. When a vessel constructed is intended for long-term charter to industrial users,  61 then the ship design is likely to be determined by the specifications of the charterers.  Also the financing deals offered on newbuilding contracts are often very attractive, demanding less equity from owners while giving them longer loan repayment schedules and even perhaps, in some cases, a two year moratorium on the loan. Some shipbuilding nations offer very attractive government subsidies for the purpose of supporting domestic shipbuilding industries such as OECD.  What is more, on the financing side, some shipyards may receive government aid for political reasons. An example of this is the operating aid criteria within the European Economic Community Sixth Shipbuilding Directive. This type of aid permits member states to provide direct subsidies to yards to make EEC newbuilding prices competitive with the Far East. The directive sets a "common maximum ceiling" expressed as a percentage of the contract price before aid. For 1990, this ceiling has been set at 20% with a lower ceiling of 14% for vessels costing less than 6 million ECUs (in November 1990, 1.385 ECU = one pound). Special development aid is permitted for orders form the third world nations in the form of more favourable credit terms. One good example now is the disputable credits for China Ocean Shipping Company (COSCO) to order 4 units of 3,800 teu full container vessels at German yards. (Because of the dispute within Germany, COSCO have just secured  62 similar orders with Hitachi in Japan under the quite similar credit terms offered by the Japanese Government.)  This proves to be a great advantage in the decision to purchase newbuildings in that government aid can reduce shipbuilding contract prices to artificially low levels.  6.2 DISADVANTAGES OF NEWBUILDING  The arguments against the contracting of new vessels centre around the total cost which, in all but exceptional circumstances, will be higher than that of secondhand tonnage.  A further related problem with the purchase of newbuildings is the length of "lead time"; this is the time taken between the contract being placed and the vessel being delivered. Delivery times vary depending on the type and size of vessel being built, and the volume of work in the yard. When long lead times are involved there is always the risk that market conditions might change between the date of ordering the date of delivery, and that these changes can be to the disadvantage of the shipowner.  Furthermore, as the new construction payments are always paid over next 2 to 4 years in terms of foreign currency, say, Japanese Yen via bank or other financial loans, there is a risk associated with the foreign exchange rates. Such foreign exchange risk can be very  63 substantial if the exchange rates go against a certain shipowners. However the risks can be reduced to the minimum acceptable level by using various hedging techniques such as swapping or future contracts. It would be very dangerous to have open positions for the newbuilding payment instalments.  6.3 BUYERS OF NEWBUILDINGS  Due to the high initial capital outlay for a newbuilding, the number of companies which can afford to purchase new tonnage is limited. The range of bulk fleet newbuilding ownership has, therefore, become more and more concentrated over the years. For the moment the principal interest in new orders is derived from a limited number of purchasing groups.  1.  Large multinational corporations such as the oil majors, who can afford newbuildings and will use the ships regularly for their own purposes for the majority of the ships' life. This is generally in addition to taking tonnage on time charter if demand arises.  2.  Investors looking purely at asset value appreciation are interested in newbuildings as commodities and often have little interest in the trading option. These investors tend to order newbuildings when they anticipate the freight rates are  64 going to rise. These newbuildings are likely to go on to be operated by management companies based in Hong Kong or Cyprus.  3.  There are some ship owners who order newbuildings in order to charter out long-term (between 5 - 10 years) as soon as the ship is built. Normally in those cases, the charter contracts are already secured before orders for newbuildings are placed with the shipyards.  4.  There are some ship owners who aim at expansion of their fleet and operation scale. Those owners together with other buyers may be primarily inspired by the fiscal regime they face. For example, in some countries, the shipowners can enjoy very good capital cost allowance (CCA) when the tax matters are involved.  5. There is also a form of state aid which leads to the government itself investing in newbuilding tonnage. This injects money into domestic shipyards and covers captive trading needs. This also has the advantage of assisting with foreign exchange. An example of this is orders placed by COSCO of China for various types of ships from the Chinese shipyards.  65 7. SECOND-HAND SHIP FINANCE  Shipping companies may have retained income in terms of cash available for their own investment in ships. Normally shipowners are unable to finance new tonnage whether secondhand or newbuilding entirely from their owners resources owing primarily to the low capital return on existing tonnage and the high rate of inflation in shipbuilding costs. Many owners supply up to one-third of the capital for a new project from their own funds and raise the residue from external supplementary sources.  The external investment funds for ship finance come from the various sources which can be categorized into three main types: government grants and subsidies, financial markets and specialist financing institutions.  7.1 GOVERNMENT GRANTS AND SUBSIDIES  Governments or government agencies in some countries grant domestic loans for the construction of new ships in domestic yards. They also provide grants in the form of direct subsidies or incentives for the modernization and/or expansion of national fleet and domestic shipyards. The degree and form of such grants or subsidies vary from one country to the another.  66 7.2 FINANCIAL MARKETS FINANCING  The most direct method is to arrange a private placement with a suitable lender, but this is expensive as the lender needs to evaluate the "credit risk" of the borrower. A much simpler way is to raise financing in the three basic financial markets.  Money Markets  Money markets trade in short term debt (less than a year) issued by companies, governments, etc and known collectively as 'money market instruments' - in effect these are IOUs.  Capital Markets  Capital markets trade in longer term debt finance instruments known as bonds or debentures. Companies who need long term finance issue bonds which repay a specified sum of money on maturity, say in ten years. Interest rates reflect the credit rating of the issuer.  Equity Markets  Equity markets trade in shares. Companies can raise equity capital by public offers of shares.  67 These markets act as a sophisticated 'risk filter' and are organized in much the same way as the voyage charter market, consisting of networks of banks, brokers and dealers linked by telex, telephones and computers and deal in the various bonds, shares, etc. They are highly regulated and to raise capital a company must achieve recognised standards of credit-worthiness and offer a competitive rate of return.  By capital market standards the finance required by shipping is not particularly large. The world securities market (debts plus equities) was worth about USD6.2 trillion in the 1990s, of which half consists of shares issued on the New York Stock Exchange, so the much quoted USD200 billion ship investment in the 1980s accounted for less than 3%. 23  7.3 SPECIALIST FINANCIAL INSTITUTIONS  The shipowners have traditionally preferred to raise finance from the range of specialist financial institutions. These institutions use their own funds or their credit rating to borrow in the capital markets and lend on at a profit. This is a little more expensive because of a middleman's margin is involved, but it offers the flexibility and confidentiality that many shipowners value.  23.^  Martin Stopford (1990) Analysis -Ship Finance, Seatrade Business Review May/June 1990, Seatrade Publishing, 2329.  68 Amongst those institutions, the commercial banks provide loans for up to 10 years. Big loans are usually "syndicated" among several banks.  The mortgage banks are also very important in the ship finance market. They provide loans backed by the ship mortgage. By doing so, the risks of the those banks can be reduced.  There are also merchant banks who arrange loans, public offers of equity, private placements and bond issues. Those institutions are mostly used by public shipping companies.  The finance houses or trading houses play more and more important roles, especially in Japan. For example, in Tokyo, big trading houses such as Kanematsu, Nichimen, Summimoto and Nissho Iwai all have a big marine investment department responsible for their marine investment. They provide loans from their own funds under management especially when the financing market is very tight.  There are also leasing companies who buy the ships first and then lease to the shipowners to operate. The owners in return pay to lessors in semi-annual instalments with a down-payment of about 20% of the ship value. The lease term is usually between 5 to 10 years for bulk carriers.  69 8. SALE AND PURCHASE MARKET REVIEW 1981-1991  TABLE 4.2 NUMBER OF SALES OF DRY BULK CARRIERS 1982 - 1991 82^83^84^85^86^87^88^89^90^91 TOTAL Lakesize^130 155 139 140 156 176 200 211 133 163 1,603 Handymax^48^59^66^55^97^82 129 126^86 108^856 Panamax^33^36^63^55^73^83^89^74^51^75^615 Total^211 250 268 250 326 341 418 411 270 346 3,064 Source: Calculated from the sales records 1982-1991. Table 4.2 indicates that there was a steady increase of the number of dry bulk carriers sold and bought until 1989. This trend was also valid for all the three types of bulk carriers. The primary reason the number of sales went down in 1990 was the Iraqi invasion of Kuwait. As a result of this unexpected event, the world sale and purchase market virtually stopped as almost all sellers and buyers were uncertain about the effects of this invasion. They were taking "wait-and-see" attitudes. However the sales picked up in 1991 after the end of the Gulf war and the sale and purchase market was back to normal. TABLE 4.3 PRICES OF SECONDHAND DRY BULK CARRIERS 1982 - 1991 (USD/DWT) 82  83  84  85  86  87  88  89  90  91 AVERAGE  Lakesize Handymax Panamax  314 227 159  341 206 139  213 206 208  176 155 101  111 105 104  174 184 137  357 328 233  471 428  526 423  467 441  AVERAGE  233  229  209  144  107  165  306  407  415  407  322  297  Note: The above figures are prices of 5-year-old ships. Source: Calculated from the sales records 1982-1991.  312  315 270 201  70 It is quite obvious from Table 4.3 that the sale prices in 1985 and 1986 experienced the lowest level during the past 10 years. The prices started at a high level at the beginning of 1980's and then arrived at the valley due to over supply and shrinking demand in the freight market in 1985 and 1986, then picked up in 1987 when  Figure 4.2 AGE PROFILE OF SECONDHAND DRY RULE CARRIERS '1982-1991 15  14  13  12  11  10  9  8  7  6 1982  1983  1984  1995  1986^1987^1988  1999  1990  1991  YEAR •  Lakesize^111 Ha ndyma;,  •  Panama,'^• Overall  NOTE Calculated from the sales records  the recession was over. The sale and purchase market in terms of both prices and number of sales reached the peak in 1989 and the first half of 1990 (before the Gulf war). The market maintained at this firm level throughout 1991.  71 Figure 4.2 shows the age profile statistics of the secondhand dry bulk carriers sold during the last 10 years. On average, the age falls in the margins of 9 and 14 years old. One of the interesting things from the age profile is, after 1984, the lakesize bulkers sold were older than other types of bulkers.  In conclusion, the world sale and purchase market experienced ups and downs in the last decade. Both prices and number of sales fluctuated quite substantially.  72 CHAPTER V  OBSERVATION OF FACTORS AFFECTING SECOND-HAND PRICES  1.  INTRODUCTION  In the previous chapters, the handymax dry bulk shipping market and the corresponding sale and purchase market are examined. This chapter is mainly concerned with the factors that might be significantly related to the secondhand values of a handymax bulk carrier. The qualitative analysis of those factors forms the base for the inclusion of variables that are used in the regression model discussed in the next chapter.  2.  NEWBUILDING PRICES  Owners who originally consider newbuildings may turn to look at the second-hand ships if the newbuilding prices are comparatively high. On the other hand, those who seek to buy second-hand ships may consider newbuildings if the second-hand values are firm. Despite the fact that there is one to two year leadtime for new construction, the secondhand ship sale and purchase market and the newbuilding market generally move in the same directions.  It is not difficult to observe that the relationship between the newbuilding prices and the corresponding second-hand ship prices is  ^  73  a positive one. When the second-hand prices are high, the newbuilding prices are high, and vice versa. During 1985 and 1986, the world shipbuilding industry was experiencing a very low demand for new constructions. At the same time, the prices of secondhand tonnage were very low, too. For late 1980s, the prices of newbuildings and secondhand tonnage both kept going up. During the beginning years of 1980s, while the newbuilding prices  Figure 5.1  HANDYMAX BULK CARRIER PRICES 1982 - 1991 NEWBU I LD I NG VS. SECONDHAND 1100  1000  900  800  700  500  500  400  300  200  100  1983  ^  1984  ^  1985  ^  1986^1987  ^  1988  ^  1989  ^  1990  ^  1991  TEAR Ha ndymax newbu ildi ng^—0— Handymax Secondhand  decreased substantially, the secondhand prices were also in downward trend. The price ratio between a 5-year-old secondhand ship and a corresponding newbuilding ship was 25.5% in 1982 and  74 22.0% in 1986. The ratio rose to 37.5% in 1987, 44.4% in 1989 and 42.8% in 1991. The continuing increase of the price ratios after 1986 is attributed to the fact that the percentage increase in prices of secondhand tonnage was more than that of newbuilding ships over the period of 1987 and 1991. The prices of newbuilding ships went up from USD480/dwt in 1986 to USD1,050/dwt in 1991, or increased by 118%, as compared with USD120/dwt in 1986 to USD450/dwt in 1991, an increase of 275% for the secondhand tonnage. The reason may be that the buyers were willing to pay a high premium for the immediacy of the secondhand tonnage when there was high demand for such tonnage.  Statistically, there is a strong correlation between the secondhand and newbuilding prices with a correlation coefficient being about 89%. The high correlation indicates the close relationship between the two prices. The positive relationship is also tested when multiple regression is developed in next chapter. Despite other variables such as freight rates, it is expected that the relationship remains statistically significant.  3. VOYAGE CHARTER RATES  Voyage charter rates expressed in terms of dollars per freight ton of cargo, reflect the short-term demand and supply of tonnage. They also relate to the relative voyage positions of tonnage such as where and when ships are open. The voyage charter market is very  75  sensitive to the changes of demand and supply. When demand is stimulated by an increase of freight cargoes, the voyage charter rates go up, and vice versa.  Voyage charterers and the owners of tramp tonnage are monitoring the market constantly and thus have very good feel about how and where the market develops. They are just like stock brokers in the stock market.  Normally, the voyage charter rates are higher than the period charter rates, which reflects some degrees of uncertainty associated with the voyage charter. One obvious uncertainty is that owners may not be able to find another voyage charter at the place and time of redelivery of the tramp charter. The voyage charter rates have partly incorporated the uncertainty that the ships may have to be positioned for next employment in ballast.  The voyage charter rates are mainly determined by the short-run demand and supply of tonnage. Ship investment is very capital intensive and a ship is normally worth millions of US dollars. The investment return period is normally 8 years or longer. A rational investor would not be stimulated to buy a ship by the short-run freight market. He or she would look at future profit prospects of the ship in the investment appraisal. Therefore, the voyage charter rates should not have a significant bearing on the ship prices. What is much more important to a buyer in deciding how much to pay  76 is the period charter market which incorporates expectations about the future freight market development. Expectations affect buyer' willingness to pay and thus become embodied in the market prices. The secondhand ship prices should to some extent reflect such expectations.  4. TIME CHARTER RATES  Time charter rates expressed in terms of dollars per day per ship, reflect the long-run or future demand and supply of tonnage. Time charter rates reflect both charterers and owners expectations about the future. The time charter period can range from 3 months to 3-4 years or even longer depending on each contract. Since secondhand ships are purchased for future trading, their prices should reflect the charterers' expectations about the future freight market. This hypothesis is tested empirically in the next chapter by multiple regression analysis.  Figure 5.2 shows the changes in time charter rats and secondhand prices for the past 10 years. When times were bad during 1985 and 1986, both time charter rates and secondhand prices were low. On the contrary, during 1989 and 1991, both were up.  Statistically, the time charter rate indices and the secondhand prices for handymax bulk carrier over the period of 1983 and 1991 are significantly correlated. The correlation coefficient is about  77  Figure 5.2  TIME CHARTER RATES VS. SECONDHAND PRICES 1983-1991 HANDYMAX DRY BULK CARRIERS 500  400  300  200  100  0 1983  1984  1985  _w_Charter Pate Index  1986  1987  1988  1989  1990  1991  ,w_Handymax Secondhand Price  Ca) Time charter rate index 1980 = 100; Cb) Secondhand prices $/dwt  88%, which indicates that there exists a very strong statistical relationship between these two variables.  5.^PRICES OF SUBSTITUTES  Generally when prices of handymax bulk carriers go up, it is expected that the prices of panamax bulk carriers and lakesize bulk carriers to rise. This is mainly due to the substitution effect amongst the three types of bulk carriers: handymax, lakesize and panamax. When handymax prices go up, buyers initially interested in  78 purchasing handymax bulk carriers may turn to look at lakesize or panamax. Similarly, when buyers find lakesize or panamax bulk carriers more expensive, they might switch to handymax. Therefore, there may exist the possibility of substitution among the three types of bulk carriers, which makes the prices of each type consistent with another.  However, since ships are normally designed for certain trades, it might not be economical to deploy panamax size vessels on certain routes where handymax vessels have been trading efficiently. Physical constraints with regards to loading and discharging speed and drafts would make swaps between different sizes of ships much more difficult or even impossible. It is practically very hard to examine the effect of substitution on the prices of each type of ships.  Figure 5.3 plots prices for the three types of dry bulk carriers over the period of 1982 and 1991. It shows that the prices moved in almost the same path. This is further enhanced by the very high correlation between the prices of the three types.  6. LAID UP TONNAGE -  Though a ship is technically is capable of trading, the owners might choose to lay up the ship if the charter revenue can not  79  Figure 5.3 COMPARISION OF SECONDHAND BULK CARRIERS 1991-1991 5-YEAR OLD  Ha ndymax ^_o_ Lakes i ze^Pa namax  cover the running costs 24 . When the ship is laid up, the expenses associated would include costs for a minimum size of crew required on board, very minimal port laid-up charges and other management expenses.  It seems from Figure 5.4 that when laid-up tonnage is high, the secondhand value is low, and vice versa. This relationship sounds  N.  The simplest decision criterion for laying up a ship is when the loss of trading (the difference between trading operating costs and charter revenue) is greater than the laid-up expenses, the owners would be better off to lay up the ship. The laid-up period would mainly depend on the spot market and the owners' expectations about the future freight market development and operating costs.  80  Figure 5.4 COMPARISION OF LAIDUP TONNAGE AND SECONDHAND PRICE 1982-1991 80  70  60  50  40  30  20  10  0 1982  ^  1983  ^  1984  ^  1985  ^  1986^1987  ^  1988  ^  1989  ^  1990  ^  1991  YEAR ^Handymax Secondhand Price (6/100 dwt)^ Lakes ice Laidup ( . 000 dwt) ^Handymax  Laidup  ( '000 dwt)  Panamax  Laidup  (  .000  dwt)  Ca). 5-year-old Handymax Prices (6/100 dwt for the sake of comparision)  logical in that the time when laid-up tonnage is high is the time when freight rate does not justify the operation and/or the secondhand value or scrap value does not justify a sale.  The significance of laid-up activity in terms of explanatory power for the secondhand value is tested in the regression analysis in the following chapter.  81 7. SHIP FINANCE  As discussed in Chapter III, ship finance can take various forms such as bank loan, finance lease, bond issue, etc. Interest rates and length of loan are the two most important features of all those financing sources. Given the length of loan, the interest rates in a bank loan agreement determine how much the instalment payments are; the finance lease interest rates determine the amount of each lease payment; the interest rates in a bond issue determine the face value of the bond. Therefore, the interest rate levels can, at least to some extent, be used an indicator for the cost of ship finance which is an important factor in the ship sale and purchase decision making process. For the same size and type of ship, the newbuilding ship requires substantially more funds than corresponding secondhand ships. When the interest rates are low, the newbuilding may be preferred to the secondhand, and vice versa. Therefore, interest rates do affect the capital strategies, and, thus, the value of the secondhand ships.  However, interest rates differ from country to country. For the sake of convenience, LIBOR (London Interbank Offer Rates) can be used as the representative interest rate. This is practically valid as nowadays most financing institutions offer ship financing on floating interest rate basis to divert some risks. For example, a deal could be LIBOR plus 1.25%. The LIBOR is floating on either weekly or monthly or semi-annually depending on the contract. The  82 1.25% is the premium based on the credits of the borrower. The higher the credits are, the lower this premium is. A very solid reputable borrower might get as low as 0.75%.  The LIBOR interest rate is based on the interest rates of various developed countries which reflect the states of the economies of those countries. The economic situation in each country combinedly affects the shipping market in which the ship values are determined.  Whether or not the interest rates have a significant influence on the secondhand value is to be empirically tested in the next chapter.  8. SCRAP PRICES  As explained in section 2.4 of Chapter III, scrap prices set a floor for the secondhand value. However, secondhand prices of old tonnage, specifically over 20 years old 25 , would be affected by the changes of scrap prices. Since there are very few observations in the data base on'ships over 20 years old, it is assumed here that scrap prices as a variable in the regression model is not significant.  25.^This is based on the observation that the dry bulk carriers sold for scrap during March, 1992 were all built before 1970, ie, over 22 years old.  83  9. OTHER FACTORS  In addition to the factors previously discussed, there are some other factors. For example, taxation incentives are so important that this might be the main criterion in the sale and purchase decision making process. This happened in Norway in 1989 when the Norwegian tax system could be taken advantage of by writing off earnings through purchasing ships.  Technically, all those factors that are not explicitly allowed for in the regression analysis are represented by dummy variables. For the sake of simplicity, only annual dummy variables are used though quarterly or monthly or even weekly dummies 26 could be used.  m. As pointed out by Dr. Murray Frank of UBC, it seems to offer relatively little in terms of interpretation by using monthly dummies, for instance. Statistically it does not add much explanatory power in the regression. See lain Cockburn and Murray Frank, Market Conditions and Retirement of Physical Capital: Evidence From Oil Tankers, working paper in Faculty of Commerce, seminar dated November 24, 1991.  84 CHAPTER VI  ECONOMETRIC MODEL AND EMPIRICAL ANALYSIS  1.  INTRODUCTION  The handymax freight market and the sale and purchase market have been described in previous chapters. Factors that would possibly affect the secondhand values have also been discussed. In this chapter, a comprehensive multiple linear autoregression model is formulated to help to look into how the factors discussed in Chapter IV are related to the secondhand prices. Hypotheses based on the discussion of Chapter IV are tested empirically. Data used to estimate the model consist of records of individual sale and purchase transactions and other variables, which would make the model more accurate and meaningful.  2.  GENERAL LOG LINEAR MULTIPLE AUTOREGRESSION MODEL  Linear multiple regression analysis is one the most widely used of all statistical tools. The log linear model specifies the logarithm of dependent variable such as the secondhand prices for handymax bulkers as a linear function of the logarithms of potential determinants, such as interest rates and time charter rates. There  85 are several advantages of using long linear mode1.  27  First, the  coefficients of selected independent variables are the elasticities. Second, the log-linear function is capable of modelling nonlinear effects. Third, the log-linear model is much simpler than other statistical models such as translog model and logit model. The main drawback of log linear model is that the elasticities are constant across all data points of dependent variable (the secondhand prices for handymax bulkers in this thesis).  The reason for using multiple autoregression model is to adjust for the existence of autocorrelation: that is the error terms of the regression model are correlated over time. As the values of variables in this thesis are time series, autocorrelation might exist. When autocorrelation is present, the ordinary least square (OLS) parameter estimates are not efficient and the standard error estimates are biased. The autoregression model is designed to make adjustments to produce better estimates in these cases. Parameter estimates so produced are usually similar to OLS estimates, but the standard errors can be very different, affecting hypothesis tests.  27  ^  Oum T.H. (1989) Alternative Demand Models and Their Elasticity Estimates, Journal of Transportation Economics and Policy, May 1989, 165-166  86 If there are p independent variables  X2,  • ••  xp and n  observations i = 1,^n, log linear multiple autoregression model is  lny i = Jo () +^+ b 2 *lnx 2 +,^, + bp *lnx p + v i where v i = e i - a 1 vi _ 1 - a 2 vi _ 2 - . • • 1 = the order of the above autoregressive process. In matrix terms, we can write linear multiple autoregression model as follows:  Y = XB + V  where^V is a vector of autoregressive error terms containing true random error terms Efej, Y is a vector of responses (logarithms of dependent variables), B is a vector of parameters or coefficients, whose true values are never known but estimated. X is a matrix of known independent constants in logarithms from observations, which are hypothesized to be relevant explanatory influences of variations in the values of  E is a vector of independent normal random variables with expectations E (e) =0 and variance-covariance matrix ....  87 The presence of this term in empirical economic relations is justified on the following grounds:  1. The matrix X can rarely include all possible factors which influence Y. Consequently, matrix E contains the influences which are omitted either because:  a) their individual importance in explaining the behaviour of Y is very small compared with that of the variables included in X; or  b) it may be either impractical or impossible to quantify these omitted variables.  28  2. The sample observations of both X and Y are likely subject to measurement errors. These errors are covered by E.  3. The form of the model may contain errors. Furthermore, errors are likely to occur due to the n. As discussed in Section 4 of Chapter III, the quality as well as sophistication of a ship certainly have bearing on her price when traded. Those qualitative differences, though marginal in most cases, are usually not measurable. For example, the premium paid upon the feature of the container-fitted for a handymax dry bulk carrier could hardly be identified quantitatively. The error terms together with the dummy variables in the regression model should cover those uncountable measures.  88 effects of the erratic element which is inherent in human behaviour.  Those matrix variables are defined as follows:  lny,^1 lnx l , 1nx 12 ... lnx 1 , 1, lny 2^1 1nx 21 1nx 22 ... lnx 2 , p  Y =^X =^.^.^...^.  lnyn^1 lnx ni lnx a ... lnxiu  b o^e l  B =  b l ^e 2 ^ E =^.  b y^en  V=  vo vi  vn  (a) If autocorrelation DOES NOT exist, the random vector Y has expectations  E(Y) = XB  89 and the variance-covariance matrix of Y is: (72(y) = a21  where I is a unit matrix with the same dimensions with Y.  Let us denote the vector of the estimated regression coefficients 0 0 , 0 1 , ..., Op as B * :  B*  = • •  Op  The least squares normal equations for the general linear regression model are  X'XB * = X'Y  and the least squares estimators are  B * = (X'X) 4 X'Y  which have all the normal properties such as unbiasedness, minimum variance unbiasedness, consistency and sufficiency.  90 (b) If autocorrelation DOES exist, the parameter estimates can be produced by Yule-Walker equations which make adjustments for the autoregressive progress 29 .  Durbin-Watson statistic is used to test whether autocorrelation (strictly first-order autocorrelation when 1 = 1) is significant or not.  n  D -  Et=2 ce, — e Et=2  i4  )2  e i2  The decision rule for testing is as follows:  If D > d u , autocorrelation not significant If D < d u autocorrelation significant If d l  < D < d u , the test inconclusive  where d u d u are lower and upper bounds in the Durbin-Watson Table.  Apart from the autocorrelation problems, multicollinearity might exist if the independent variables are highly correlated amongst themselves. The traditional OLS parameter estimates may not be  n.^See SAS/ETS User's Guide Version 5 Edition, P189.  91 efficient, consistent and unbiased. However, this is no longer a serious problem as most computer software such as SAS PROC REG automatically detects the problem and make suitable adjustments. 3°  3. VARIABLES INCLUDED IN ANALYSIS  As outlined in the previous chapter, independent variables are those which may influence the prices of the secondhand handymax bulk carriers significantly. Whether or not they are statistically significant is to be tested empirically.  y = PH - Secondhand prices for Handymax bulkers; x l = NBP - Newbuilding prices for all bulkers; 31 x 11 = NBPH - Newbuilding prices for Handymax bulkers; x 12 = NBPL - Newbuilding prices for Lakesize bulkers; x 13 = NBPP - Newbuilding prices for Panamax bulkers; x 2 = TC - Time charter rates for all bulkers;  x21 = TCH - Time charter rates for Handymax bulkers; x22 = TCL - Time charter rates for Lakesize bulkers; x23 = TCP - Time charter rates for Panamax bulkers; X3  = p^- Secondhand prices for all bulkers;  xm = PL - Secondhand prices for Lakesize bulkers;  30  ^  See Freund R. and R. C. Littell (1981) SAS For Linear Models, Sas Institute Inc., North Carolina, 42-43.  m.^Including Lakesize, Handymax and Panamax types, ie, 10,000 - 79,999 dwt dry bulk carriers.  92 x 32 = PP - Secondhand prices for Panamax bulkers; x4 = LU - Laid-up tonnage for all bulkers; x41 = LUH - Laid-up tonnage for Handymax bulkers; x42 = LUL - Laid-up tonnage for Lakesize bulkers; x43 = LUP - Laid-up tonnage for Panamax bulkers; x5 = AGE - Ship ages for all bulkers; x 51 = AGEH - Ship ages for Handymax bulkers; xn = AGEL - Ship ages for Lakesize bulkers; x 53 = AGEP - Ship ages for Panamax bulkers; x 6 = LIBOR - 6-month London Interbank Offer Rates.  4.^DATA BASE  Data cover the 10 year period from 1982 to 1991. -  4.1 NEWBUILDING PRICES  Prices in terms of US dollars per dwt were collected on quarterly basis from Shipping Statistics and Economics published monthly by Drewry Shipping Consultants for all handymax, lakesize and panamax types. It is virtually impossible to obtain records on individual orders because most of them are on the private and confidential basis. The prices published by Drewry are the reported indicative contract prices of the main Japanese and/or Korean shipyards where almost 3 quarters of overall world bulk carriers are constructed. There are total of 40 observations for each type of vessel.  93 4.2 TIME CHARTER RATES  Monthly average time charter rates were collected from both Shipping Statistics and Economics of Drewry and Shipping Statistics published monthly by Institute of Shipping Economics in Bremen. There are mainly two reasons for using monthly averages. First, individual time charter rates associated with different zoning, charter period lengths, places of deliveries and redeliveries may not be able to represent the overall freight market. Second, since those charter rates differ very much, it is not possible to obtain repeat observations on time charter rates for fixed zoning, period and delivery and redelivery.  The time charter rates are expressed in terms dollars per dwt per month for one year period.  4.3 SECONDHAND PRICES  Records covering almost all transactions during the past 10 years were collected on an individual basis. The format is name, yearbuilt, dead weight tons, prices. As considerable analysis is centred on the secondhand prices, individual data are essential. In total, 3064 observations (size of the sample) were collected with 1603 for lakesize, and 856 for handymax and 615 for panamax bulk carriers. The big size of the sample gives rise to high analytical  94 accuracy. The sources are mainly Lambert London, Fearnley Oslo, Peirot New York and Wardley Hong Kong.  4.4 LAID UP TONNAGE -  Data on laid-up tonnage in terms of dead weight tons were collected on a monthly basis for each type of bulk carriers. The sources of the data are Drewry's Shipping Statistics and Economics and Shipping Statistics of Institute of Shipping Economics in Bremen. There are totally 120 observations to be used to fit the regression model.  4.5 INTEREST RATES  The London Interbank Offer Rates (LIBOR) for 6-month period loan ask rates were collected to represent the worldwide ship financing costs. Weekly records were collected and monthly averages were calculated. The source is Financial Statistical Trends published quarterly by OECD Paris office.  4.6 TIME DUMMY VARIABLES  In order to allow for unmeasured variations in market conditions for each year, dummy variables are used. The dummies may incorporate political instability in various parts of the world, major policy changes relevant to shipping and so on, which are not  95  observable. Dummy variables are the way to account for those unobservable factors. The use of time dummies may help to get higher goodness of fit of the model. The dummy variables themselves in this analysis do not have any specific meanings per se.  5. THE HYPOTHESES TO TEST  The hypothesis testing in this thesis is to check statistically whether there is a log linear relationship between the secondhand prices for handymax bulkers (Y) and each variable (x) listed under Section 3 of this chapter. One example is to test the existence of log linear relationship between secondhand prices and newbuilding prices for handymax bulkers. Mathematically, the hypotheses (in the form of null) to be tested are summarized as follows:  b1 = 0  j = 1, 2, 3  b2i = 0  j = 1, 2, 3  b 3j = 0  j = 1, 2  b4i = 0  j = 1, 2, 3  or in short, ^bhj^0  where h = 1, ..., 6 and j = 1, 2, 3.  The coefficients obtained are actually the elasticities of each independent variable on the dependent variable since the variable values are converted into natural log values in the actual regression.  96 6.  THE COMPUTING SOFTWARE PACKAGE USED  The Statistical Analysis System (SAS) at UBC Mainframe is used. SAS is an extensive computer software system designed for data management, statistical analysis, and report writing.  7.  EMPIRICAL RESULTS AND EXPLANATIONS  TABLE 6.1 REGRESSION RESULTS TABLE EQUATION 1 PARA T Intercept LAGEH LNBPL LNBPH LNBPP LPL LPP LTCL LTCH LTCP LLIBOR LLUL LLUH LLUP DUMMIES  -0.027 0.009 -0.714 -34.424 0.285 0.414 0.570 0.784 0.006 0.010 0.015 0.588 0.057 2.205 -0.054 -0.115 0.447 0.535 0.459 1.338 -0.138 -0.618 0.085 0.788 0.026 0.313 -0.007 -0.142 YES  EQUATION 2 PARA T 1.215 0.695 -0.717 -35.024 0.761 2.765  0.888  5.523  EQUATION 3 PARA^T 1.189 2.565 -0.714 -35.088 0.558 6.317  1.207  YES  NO  Y-W R2  0.8352  0.8324  0.8259  D-W  1.9421  1.9028  1.8218  16.655  While the detailed regression results are given in Appendix II, Table^6.1 lists the main regression results of three different regression equations. For regression equation 1 where all the variables are included, the t-values for most variables are less  97 than 2, and, thus statistically insignificant. Based on equation 1, insignificant variables are dropped one by one from the regression so that the effect of each variable dropped from the model on other existing variables can be observed. The purpose of doing so is to identify variables that have t-values, at least, close to 2 for each variable. After this process, equation 2 is derived. Equation 2 indicates that the ship age, the handymax newbuilding prices and the handymax time charter rates (representing freight market) together with most dummy variables are statistically significant with goodness of fit at 0.8324. Further analysis by dropping all the dummy variables results in equation 3, which has a goodness of fit at 0.8259. The difference in terms of goodness of fit between the two regression equations is minimal (0.8324-0.8259 = 0.0065), which implies that the dummy variables add very little explanatory power to the regression model. The high values of t statistics are consistent with the high goodness of fit of the model.  The coefficients listed in Table 6.1 are Yule-Walker parameter estimates (after adjusting autocorrelation). Ordinary least square parameter estimates are also calculated and presented in Appendix II.  Out of the three regression equations in Table 6.1, equation 3 is considered the best when R-squared and t-values and simplicity are taken into account. This selected regression equation can be expressed as follows:  98 LPH =  1.18956150 - 0.71432176*LAGEH (t=-35.088) + 0.55825281*LNBPH (t= 6.317) + 1.20719161*LTCH (t= 16.655)  where LPH = log of handymax secondhand price; LAGEH = log of handymax ship age; LNBPH = log of handymax newbuilding price; LTCH = log of handymax time charter rates; t^= t statistic.  There are following comments on the selected model:  1.  The t-values of all the independent variables are far greater than 2.0, indicating high level of confidence that they are statistically different from zero. Adjusted R 2 of the above regression equation is 0.8259, consistent with the high tvalues. Therefore the model fits the data very well. The Rsquared of 0.8259 is considered very high for a micro level data set covering 10-year period.  2.  All the variables in the above model are handymax variables, which indicates that the price of a Handymax bulk carrier is statistically unlikely to be affected by market conditions in the other sectors such as lakesize or panamax. This result suggests a high degree of independence or endogenuity exist for the handymax sector of the freight and ship markets.  99  Furthermore, the endogenuity with regards to the formation of the secondhand values of handymax bulkers suggests that the effect of substitution between the three types of bulker carriers, lakesize, handysize and panamax, is statistically insignificant in the ship sale and purchase market.  The significance of this result is that the buyers and sellers of handymax ships should pay major attention to the development of the handymax market when they sell or buy a handymax dry bulk carrier.  3.  The result that the time charter rates of handymax bulkers is a significant variable indicates that the ship sale and purchase market and the corresponding freight market are interacted. In other words, the ship market is significantly affected by the freight market, which is quite consistent with practical perception about the relationship of these two markets. 32  4.  The newbuilding costs of handymax bulk carriers are also tested to be an important determinant in the model. The positive coefficient indicates that the higher the newbuilding n.  Strictly speaking, the result can not conclude the other way around, that is the ship sale and purchase market significantly affects the freight market. Further but similar study with time charter rates as dependent variable and ship prices as one of the independent variables is needed.  100 price is, the higher the secondhand values is, other things being equal. This is because it is in many cases easy to make a substitution between secondhand and newbuilding tonnage.  5.  The model shows that the age of ships is a significant variable in determining ship prices. This indicates that no matter how market conditions change, the general principle is valid in that the older the ship is, the lower the secondhand value, other things being equal.  6.  Noted in the model are the signs and values of three parameters. They are expected and very consistent with the discussion in Chapter IV. The model obtained is useful in telling the relative contribution of changes of each independent variable to the dependent variable. This makes the overall effect generated from the changes of a combination of market conditions distinguishable. For example, 1% change in all the independent variables may lead to 1.051% change in the secondhand values whereby 0.714% (negative), 0.558% and 1.207% are due to changes in age, newbuilding cost and the time charter rates respectively.  7. The regression result suggests that other variables such as laid-up tonnage are statistically not significant in terms of affecting the secondhand prices. This, however, does not mean that their effects are zero.  101 8.  The inclusion of 10 dummy variables, each for each year, does not seem to add any more explanatory power to the model in terms of either R 2 or t-values for parameters. In other words, the unmeasured variations are not significant in terms of the effects on to the values of the secondhand dry bulk carriers.  9.  Durbin-Watson statistic is 1.8218 for the chosen regression model. As this is greater than the upper bound (d u P.1 1.778), the autocorrelation is not significant. Though autocorrelation is not substantial in this case, SAS PROC AUTOREG has made certain adjustments in producing Yule-Walker parameter estimates.  The testing results of the hypotheses set up in Section 4.5 of this chapter are given here. For the sake of reference, the null hypotheses (H o ) are written as follows:  Ho: b hp = 0  where h=1, 2, 3, 4, 5 j=1, 2, 3  The following null hypotheses have been accepted as their corresponding variables are ruled out from the resulting regression model:  102 ^0  b12  b 13 = 0 b22 = 0 b 23 = 0 b31 = 0 b32 = 0 b,t2 = 0 b43 = 0  and following null hypotheses have been rejected based on the tvalues of each parameter being greater than 2.0:  bil  =  0  b21  =  0  b51  =  0  which illustrates that the secondhand prices for handymax bulk carriers, the dependent variable, are significantly explained by only newbuilding prices for handymax bulkers, time charter rates for handymax bulkers and ship ages for handymax bulkers amongst those independent variables selected in Section 3 of this chapter.  The regression results can also be looked into by the elasticity analysis. To take price and age as an example, the elasticity of age to price is measured by the percentage changes of price as a  103 result of 1% change of age. Mathematically, the elasticity is given as follows:  a (LPH) a(PH) a(AGEH) = / a(LAGEH) PH AGEH =  AGEH a(PH) * PH a(AGEH)  which is the coefficient of age in the log linear regression equation. Therefore, the elasticity of ship age over secondhand price is - 0.71432176.  The fact that the absolute value of this elasticity is less than one may mean that the secondhand price is inelastic to the age. In other words, one percent (1%) age increase would bring down only about 0.7% in secondhand price, given the assumption that there is a log linear relationship between ship price and age.  Apparently, the time charter rates (representing the freight market) have the largest bearing on the secondhand prices. Mathematically, 1% change in freight market would to equivalent to about 2% change in the newbuilding price in terms of the magnitudes of effects on the ship values. In other words, changes in handymax freight market tend to have more effects on the secondhand values than the newbuilding prices.  104 All the results obtained so far are based on the critical assumption that there is a log linear relationship between the dependent variable, ie, the prices of secondhand dry bulk carriers and the independent variables such as time charter rates, ages, newbuilding prices, laid-up tonnage and so on. This assumption is just a simplification of the real relationship between those variables which might take the form of exponential or other nonlinear forms."  D.  Frank M. and Cockburn I. (1991) Market Conditions and Retirement of Physical Capital: Evidence From Oil Tankers, Seminar November 24, 1991, Faculty of Commerce, UBC.  105  CHAPTER VII  SUMMARY  The attempt to investigate the determination of secondhand ship prices is not an original undertaking. This writer does feel, however, the analysis of ship prices in the existing literature is very inadequate. This thesis uses an empirical approach to find important market variables that are related to the secondhand ship prices. The approach is fundamentally different from the traditional price analysis whereby the demand and supply are examined and estimated in order to work out price. The approach is also different from the capital theory approach developed by Beenstock. The ways and results of the investigation of this paper would hopefully be of benefit not only to the academic study in this field, but also to the ship sale and purchase market participants insofar as they may be able to better understand forces that affect secondhand ship prices and possibly assess the future behaviour of the ship sale and purchase market. On the other hand, it is an interesting empirical question in itself to see if the behaviour of the ship market can be explained and economic hypotheses about the ship market can be verified or rejected.  There are several main conclusions that seem strongly supported by the data. The most important one is that the values of the secondhand dry bulk carriers are significantly influenced by their  106 freight market (time charter) which reflects more or less the expectations of the future prosperity of those ships. This concurs with the popular insertion in the daily shipping practice such as the quotation from Lloyd's List at the beginning of this thesis.  A significant positive relationship is found existing between the newbuilding costs and the secondhand prices of handymax bulk carriers. In other words, the secondhand prices tend to be higher when the newbuilding become more expensive.  The secondhand values are also significantly positively related to the age of ships. This means that, no matter how the market conditions change, the general principle is valid in that the older the ship is, the lower the secondhand value, other things being equal.  It is worth pointing out here that this thesis treats different types of ships differently whereas other research work treats all ships homogenous as pointed in Chapter 2. The thesis selects handymax bulkers as the data base for the econometric model.  The main problem of this study the writer feels about the model selected is that the model does not allow for some qualitative and physical variations of ships themselves which apparently affect the ship price. For example, high quality and good design ships with cranes tend to have higher values. A separate study may be worth  107 undertaking to what extent the prices are affected by those endogenous qualitative variations.  In this thesis, only linear regression model is selected. As the data is time series, it would be interesting to apply ARMA model (autoregressive moving average model) and see whether the outcomes would differ.  Further research should also be directed to the other sectors of the dry bulk shipping market, tanker shipping market and container shipping market by using the same methodology used here in this paper. Similar results as obtained for the handymax bulk carriers might be found out for lakesize and panamax dry bulk carriers.  108  APPENDIX I RAW DATA FORMAT AN ABSTRACT OF SALES RECORDS (1982) FOR DRY BULK CARRIERS  109 YEAR MONTH^ 1982 1982 1982 1982 1982 1982 1982 1982 1982 1982 1982 1982 1982 1982 1982 1982 1982 1982 1982 1982 1982 1982 1982 1982 1982 1982 1982 1982 1982 1982 1982 1982 1982 1982 1982 1982 1982 1982 1982 1982 1982 1982 1982 1982 1982 1982 1982 1982 1982 1982 1982 1982 1982 1982 1982 1982 1982 1982 1982 1982  NAME YEAR BLT  1^AMSCELVEEN 1^ARKADIA 1^ ILENA 1^NAI ASSIA 1^NAI CAROLINA IXIA 1^ 1^IRON KESTREL 1^IRON KERRY 1^BENGAL CAREER 1^RIO NUNEZ 1^STONEPOOL 2^AGHIA MARINA 2^GREEN BLESSING 2^ SPRUCE 2^SILVER LIME 2^ GUNGNI 2^OLYMPIA 2^BRIGIT MAERSK 2^BELLA MAERSK 2^KING LEONIDAS 2^GOLDEN ANNE 2^BANNERLOCK 2^BASIL III 2^JALSNTS 2^TAKUYO MARU 2^SOUTH BEAUTY 3^MIDAS APOLLO 3^INDAH JUMBO 3^SAN GEROLAMO 3^MONTIRON 3^MARE TRANQUILLO 3^TORM KRISTINA 3^MARE PLACIDO 3^MOSLAKE 3^MOSRIVER 3^INVERSHIN 3^SHOZEN MARU 3^SHINSHO MARU 3^SHOCHO MARU 3^AEGEAN SKY 3^EAGLE GLORY 3^DUNSTER GRANGE 3^GAUCHO MOREIRA 3^VIRGINIA 3^SANSAN VENTURE 3^KING GEORGE 4^ GAYONG 4^EASTERN LINK 4^KOYO MARU 4^BELL ROSTITS 4^KIKI YEMELOS 4^GEORGIA RAINBOW 4 PENNYSLVANIS RAINBOW 4^ISLAND ARCHON 4^MARGARITE 4^REGAL SABRE 4^NAI CAROLINA 4^FAUSTINA 4^ TRITON 4^MANUELA PRIMA  DWT  1963 1959 1976 1965 1963 1964 1974 1974 1965 1968 1966 1963 1978 1977 1970 1963 1979 1969 1969 1963 1973 1977 1965 1973 1966 1976 1964 1972 1960 1968 1967 1972 1966 1976 1976 1972 1966 1966 1967 1966 1969 1967 1966 1968 1981 1975 1969 1980 1968 1969 1967 1977 1978 1971 1966 1961 1963 1968  19843 21075 25917 26305 26468 26559 26843 27999 35751 43540 44316 16028 16743 18292 18340 19957 22170 24240 24280 25663 25935 27313 31200 51072 53852 65112 15248 16647 17480 19228 25484 25635 25964 28707 29689 29689 36671 37623 39013 40140 41745 42144 46810 57582 58412 78238 16243 16272 16538 20123 20221 24197 24882 25309 25655 26218 26468 27365  1977 1965  27482 28051  PRICE ($MILL) 2.00 0.70 8.60 2.74 2.75 2.70 5.40 5.63 2.60 3.50 3.00 1.35 9.85 7.75 3.80 1.65 8.75 5.11 5.12 2.40 5.90 12.50 2.90 7.00 4.40 9.25 0.75 3.45 0.75 3.30 3.60 5.75 3.50 9.20 9.20 6.20 1.95 1.80 1.50 2.75 3.45 2.60 2.40 4.50 20.70 8.20 3.10 7.00 1.25 3.25 3.45 11.00 11.00  5.50 3.25 1.40 1.65 4.85 8.50 1.85  PRICE ($/DWT) 100.79 33.21 331.83 104.00 104.00 101.66 201.00 201.00 72.73 80.39 67.70 84.23 588.31 423.68 207.20 82.68 394.68 211.00 211.00 93.52 227.49 457.66 92.95 137.06 81.71 142.06 49.19 207.24 42.91 171.62 141.27 224.30 134.80 320.48 309.88 208.83 53.18 47.84 38.45 68.51 82.64 61.69 51.27 78.15 354.38 104.81 190.85 430.19 75.58 161.51 170.61 454.60 442.09 217.31 126.68 53.40 62.34 177.23 309.29  AGE  65.95  17  19 23 6 17 19 18 8 8 17 14 16 19 4 5 12 19 3 13 13 19 9 5 17 9 16 6 18 10 22 14 15 10 16 6 6 10 16 16 15 16 13 15 16 14 1 7 13 2 14 13 15 5 4 11 16 21 19 14 5  110 1982 1982 1982 1982 1982 1982 1982 1982 1982 1982 1982 1982 1982 1982 1982 1982 1982 1982 1982 1982 1982 1982 1982 1982 1982 1982 1982 1982 1982 1982 1982 1982 1982 1982 1982 1982 1982 1982 1982 1982 1982 1982 1982 1982 1982 1982 1982 1982 1982 1982 1982 1982 1982 1982 1982 1982 1982 1982 1982 1982 1982  1982  4 4 4 4 4 4 4 4 4 5 5 5 5 5 5 5 5 5 5 5 5 5 6 6 6 6 6 6 6 6 6 6 6 6 6 6 6 6 6 6 6 7 7 7 7 7 7 7 7 7 7 7 7 7 7 7 7  7  7  7 7  7  DEFIANT PACIFIC PEACE ZINA UNIVERSYET TORUNSKI KISHU MARU MANIFEST LIPCOWY KAREN T ARTHUR STOVE 8 HUNTER IRENES ECSTASY TENSHA MARU NO.5 KARAISKAKI BASIL III EASTERN HORNET CALDERETA DEFIANT MIEKAWA MARU JASAKA THORSDRAKE 1981 ARTHUR STOVE KING WILLIAM PRIMA KING CARCHESTER STROFADES CLYMENIA GOLDEN PIONEER YOZAN MARU OCEAN TAURUS WEIPA MARU GEM RONSONA REGAL SABRE PACIFIC LEADER SPAN TERZA EL GENERAL TRITON CASPIANA ASIAN ROSE ELAT PACIFIC MASTER ASEAN TRADER PRIMA KING SAEMAEUM UJUNG RAJA ROZEL BAY ARKADIA PALAPUR BRILLIANT STAR MARE TRANQUILLO MARE FELICE PRINCE RUPERT CITY RADIANT STAR LYNN RACHEL UPWEY GRANGE SPAN TERRZA IRISH OAK MARINA GRANDE NEW CADMUS GOLDEN ALLIANCE IRON HUNTER  PACIFIC MASTER  1960 1973 1966 1972 1966 1970 1981 1973 1967 1970 1969 1980 1965 1965 1973 1976 1960 1967 1981 1981 1973 1974 1955 1967 1962 1968 1973 1969 1963 1969 1977 1967 1961 1975 1977 1980 1977 1959 1971 1960 1976 1974 1955 1969 1960 1962 1959 1981 1968 1967 1966 1970 1977 1973 1976 1977 1973 1975 1976 1963 1968  1976  36882 38927 40140 50489 53341 54178 64000 66950 69883 15713 19274 26798 30545 31200 33663 34443 36882 43104 60600 60842 66950 78052 12500 14625 15492 16393 19086 19732 19789 23973 26032 26191 26218 26496 26662 27311 27482 28360 30160 31135 56518 10049 12500 13834 15300 16250 20742 23923 24534 25484 25682 25875 25887 26190 26200 26242 26510 27006 39917 44670 55466  56516  1.35 8.00 3.50 6.40 2.50 5.70 17.00 8.00 4.30 2.50 3.40 10.50 3.00 2.70 7.50 8.50 1.40 1.00 17.00 17.10 7.80 7.30 0.45 0.75 1.20 1.80 4.25 3.25 1.00 1.50 7.65 1.75 1.45 7.50 6.35 9.75 7.75 1.15 4.25 1.25 9.10 2.75 0.40 2.10 0.48 0.75 0.50 14.75 1.00 1.70 1.70 4.10 7.50 4.25 6.50 6.32 5.06 6.30 9.00 2.50  2.00 9.10  36.60 205.51 87.19 126.76 46.87 105.21 265.63 119.49 61.53 159.10 176.40 391.82 98.22 86.54 222.80 246.78 37.96 23.20 280.53 281.06 116.50 93.53 36.00 51.28 77.46 109.80 222.68 164.71 50.53 62.57 293.87 66.82 55.31 283.06 238.17 357.00 282.00 40.55 140.92 40.15 161.01 273.66 32.00 151.80 31.05 46.15 24.11 616.56 40.76 66.71 66.19 158.45 289.72 162.28 248.09 240.64 190.87 233.28 225.47 55.97  36.06 161.02  22 9 16 10 16 12 1 9 15 12 13 2 17 17 9 6 22 15 1 1 9 8 27 15 20 14 9 13 19 13 5 15 21 7 5 2 5 23 11 22 6 8 27 13 22 20 23 1 14 15 16 12 5 9 6 5 9 7 6 19  14 6  111  1982 1982 1982 1982 1982 1982 1982 1982 1982 1982 1982 1982 1982 1982 1982 1982 1982 1982 1982 1982 1982 1982 1982 1982 1982 1982 1982 1982 1982 1982 1982 1982 1982 1982 1982 1982 1982 1982 1982 1982 1982 1982 1982 1982 1982 1982 1982 1982 1982 1982 1982 1982 1982 1982 1982 1982 1982 1982 1982 1982 1982 1982  7 7 7 7 7 8 8 8 8 8 8 8 8 8 8 8 8 9 9 9 9 9 9 9 9 9 9 9 9 9 9 9 9 9 9 9 9 9 10 10 10 10 10 10 10 10 10 10 10 10 10 10 10 10 10 10 10 10 10 11 11 11  UNIONA MEILLY ROBERTS BANK KING RICHARD KAKOGAWA MARU STAR MARINE ROSEDALE ALBAFORTH YOAN MARU EASTERN FORTUNE SUGELA LEONIDAS Z CAMBANIS MORI MARU OCEAN BRAVE ALTANO PATRICIA MILROSS CARYATIS ICAROS MONTIGNY WINSUM PHAEDRA MARE TRANQUILLO MARE FELICE NEWHAVEN TOYOTA MARU YUGOH MARU GOHYO (CHIP) TRIABUNNA CHIKUZEN MARU PACIFICO ATLANTICO STONEPOOL QUEENA EDERA ODIN NIIHAMA MARU GOLDEN CAMEO ESCOBS KIAN AN KEN GROVE MONTROSE BELL ROSITA SUGELA MERCY NORSE CARRIER CALLIROY EURO PRINCESS IRISH LARCH NEWHAVEN PHILIPPINE ROSAL KARAISAKI WEDELL CAREER WORLD TRADER SALVIA SHOZU MARU CHIEKAWA MARU SORRENTO DORSETSHIRE FAIR LIZA PETRAIA DOCEPRAIA  1971 1969 1972 1967 1970 1967 1963 1969 1969 1968 1965 1965 1972 1975 1975 1977 1974 1970 1970 1969 1970 1966 1967 1966 1966 1971 1966 1973 1971 1967 1966 1965 1966 1976 1962 1965 1971 1974 1969 1962 1974 1968 1969 1965 1972 1974 1963 1972 1973 1966 1967 1965 1965 1967 1967 1969 1968 1967 1968 1975 1967 1968  59316 59819 67417 77800 80323 15580 16080 19710 20049 23646 25185 27469 31143 34506 34987 34997 41602 17350 17350 19112 21241 25330 25484 25682 26570 30227 32411 36603 41411 41643 44029 44206 44316 45056 45700 60490 76323 78501 11912 16009 17032 18810 20446 24365 25030 25790 26220 26299 26510 26996 28145 30545 35535 36930 40942 58472 58914 67901 80801 11564 16615 19578  6.50 3.95 6.25 4.95 2.50 0.80 0.75 4.20 2.60 2.30 0.85 2.10 1.30 7.90 8.50 8.50 5.40 2.76 2.55 3.00 2.30 2.00 1.50 1.50 1.15 2.70 1.38 2.10 2.70 1.60 1.75 1.10 2.38 9.00 0.65 3.20 2.25 6.25 1.30 0.80 3.00 2.65 3.00 0.85 4.93 3.90 1.00 0.85 4.45 1.15 2.00 1.15 0.90 1.80 1.45 4.00 3.53 1.30 2.50 2.20 1.20 0.35  109.58 66.03 92.71 63.62 31.12 51.35 46.64 213.09 129.68 97.27 33.75 76.45 41.74 228.95 242.95 242.88 129.80 158.90 146.97 156.97 108.28 78.96 58.86 58.41 43.28 89.32 42.42 57.37 65.20 38.42 39.75 24.88 53.71 199.75 14.22 52.90 29.48 79.62 109.13 49.97 176.14 140.88 146.73 34.89 196.96 151.22 38.14 32.32 167.86 42.60 71.06 37.65 25.33 48.74 35.42 68.41 59.92 19.15 30.94 190.25 72.22 17.88  11 13 10 15 12 15 19 13 13 14 17 17 10 7 7 5 8 12 12 13 12 16 15 16 16 11 16 9 11 15 16 17 16 6 20 17 11 8 13 20 8 14 13 17 10 8 19 10 9 16 15 17 17 15 15 13 14 15 14 7 15 14  112 1982 1982 1982 1982 1982 1982 1982 1982 1982 1982 1982 1982 1982 1982 1982 1982 1982 1982 1982 1982 1982 1982 1982 1982 1982 1982 1982  11 11 11 11 11 11 11 11 11 11 12 12 12 12 12 12 12 12 12 12 12 12 12 12 12 12 12  NORSE CAPTAIN NORSE RIVER OCEAN PRINCE PHILIPPINE ROSAL OLAMAR REYNOLDS MARCOPLATA ROYAL FORNAY BORNHEIM KRISTINA VIRGO ORCHID VENTURE SUMMER DREAM ROYAN TUAMOTU OCEAN SOVEREIGN HAVFRU CARA VEER VARUNA NORSE CAPTAIN PAMINA UNIQUE FORTUNE GOLDEN MIRANDA TRONGATE ASIA LARK GENTLE RIVER NORMAN VENTURE  1970 1970 1954 1967 1972 1973 1979 1976 1966 1977 1974 1975 1961 1968 1971 1966 1967 1965 1968 1970 1980 1972 1969 1977 1972 1977 1971  21692 22353 26813 28145 29095 29812 34556 39884 42805 56498 10339 11000 15353 16447 16884 19829 20010 21164 21208 22040 22042 22268 24115 29119 30159 33135 61460  3.50 3.80 0.48 1.30 2.50 5.00 9.40 7.50 1.90 9.00 1.80 2.30 0.50 0.95 1.70 1.20 1.00 0.85 1.75 2.50 10.00 3.60 2.00 4.50 2.70 7.00 5.00  161.35 170.00 17.90 46.19 85.93 167.72 272.02 188.05 44.39 159.30 174.10 209.09 32.57 57.76 100.69 60.52 49.98 40.16 82.52 113.43 453.68 161.67 82.94 154.54 89.53 211.26 81.35  12 12 28 15 10 9 3 6 16 5 8 7 21 14 11 16 15 17 14 12 2 10 13 5 10 5 11  113  APPENDIX II SAS REGRESSION PRINT-OUTS  2^SAS(R) LOG^OS SAS 5.08^VS2/SVS JOB 8H46CINS STEP VSSO2.0 ^ 24^MODEL LPH = LAGEH LNBPH LTCH / NLAG=12; 25^OUTPUT OUT=OUT1 P=LPHIPRED R=LPH1RES; 28^*PROC PRINT DATA=OUT1; NOTE: THE DATA SET WORK.OUT1 HAS 815 OBSERVATIONS AND 40 VARIABLES. 40 OBS/TRK. NOTE: THE PROCEDURE AUTOREG USED 1.74 SECONDS AND 472K AND PRINTED PAGES 4 TO 5. 27^PROC PLOT DATA=OUT1; 28^PLOT LPH1RES'LPH1PRED= . e; NOTE: THE PROCEDURE PLOT USED 0.79 SECONDS AND 402K AND PRINTED PAGE 6. 29^PROC AUTOREG; 30^MODEL LPH = LAGEH LNBPH LTCH 31^ D1 D2 D3 04 05 06 07 08 D9 010 / NLAG=12; NOTE: THE PROCEDURE AUTOREG USED 1.32 SECONDS AND 472K AND PRINTED PAGES 7 TO 8. 32^PROC AUTOREG; 33^MODEL LPH = LAGEH LNBPH LNBPL LNBPP LPL LPP LTCL LTCH LTCP 34^ LLIBOR LLUL LLUH LLUP 01 D2 D3 D4 05 D6 D7 08 D9 010 / NLAG=12; 35^ NOTE: THE PROCEDURE AUTOREG USED 1.83 SECONDS AND 472K AND PRINTED PAGES 9 TO 11. NOTE: SAS USED 472K MEMORY. NOTE: SAS INSTITUTE INC. SAS CIRCLE PO BOX 8000 CARY, N.C.^27511.8000  14:08 MONDAY, MARCH 8, 1993  1^SAS(R) LOG^OS SAS 5.08^VS2/SVS JOB 6H46CINS STEP VSSM2.0^  14:08 MONDAY, MARCH 8. 1993  NOTE: COPYRIGHT (C) 1984 SAS INSTITUTE INC., CARY, N.C. ^27511, U.S.A. NOTE: THE JOB 6H46CINS HAS BEEN RUN UNDER RELEASE 5.08 OF SAS AT UNIVERSITY OF BRITISH COLUMBIA (02837001). NOTE: CPUID^VERSION^63 SERIAL^020804 MODEL^3081 NOTE: SAS OPTIONS SPECIFIED ARE: MACRO 1^TITLE 'SHIPS SAS PROGRAM'; 2^DATA TEMP; 3^infile filet; 4^input PH AGEH NBPL NBPH NBPP PL PP TCL TCH TCP: 5^infile filet; 6^input LIBOR LUL LUH LUP D1 D2 D3 D4 05 D6 D7 D8 09 010; 7^LPH^= LOG (PH); 8^LAGEH = LOG (AGEH); 9^LPL^. LOG (PL); 10^LPP^. LOG (PP); 11^LNBPL = LOG (NBPL); 12^LNBPH . LOG (NBPH); 13^LNBPP = LOG (NBPP); 14^LTCL = LOG (TCL); 15^ITCH = LOG ITCH); 18^LTCP = LOG (TCP); 17^LLIBOR= LOG (LIBOR); 18^LLUL = LOG (LUL); 19^LLUH = LOG (LUH); 20^LLUP = LOG (LUP); NOTE: INFILE FILE1 IS: DSNAME=SHIPSAS1, UNIT=DISK,VOL=SER.MTS026,DISP.OLD, DCB=(BLKSIZE=32760.LRECL.32760,RECFM.U) NOTE: INFILE FILE2 IS: OSNAME=SHIPSAS2, UNIT=DISK,VOL=SER=MTS013,DISP=OLD, DCB.(BLKSIZE=32760,LRECL=32760,RECFM.U) NOTE: SAS WENT TO A NEW LINE WHEN INPUT STATEMENT REACHED PAST THE END OF A LINE. NOTE: 818 LINES WERE READ FROM INFILE FILE1. THE MINIMUM LINE LENGTH IS 1, THE MAXIMUM LINE LENGTH IS 79. NOTE: 688 LINES WERE READ FROM INFILE FILE2. THE MINIMUM LINE LENGTH IS 1. THE MAXIMUM LINE LENGTH IS 125. NOTE: DATA SET WORK.TEMP HAS 615 OBSERVATIONS AND 38 VARIABLES. 42 OBS/TRK. NOTE: THE DATA STATEMENT USED 1.66 SECONDS AND 280K. 21^PROC CONTENTS; NOTE: THE PROCEDURE CONTENTS USED 0.65 SECONDS AND 434K AND PRINTED PAGES 1 TO 2. 22^• PROC MEANS; NOTE: THE PROCEDURE MEANS USED 1.00 SECONDS AND 426K AND PRINTED PAGE 3. 23^PROC AUTOREG;  14:08 MONDAY, MARCH 8, ^1993  SHIPS SAS PROGRAM^ AUTOREG^PROCEDURE VARIABLE DF^B VALUE^STD ERROR^I RATIO APPROX PROB LTCL LTCH LTCP LLIBOR LLUL LLUH LLUP DI D2 D3 D4 D5 D6 07 08 D9 D10  -0.05446431 0.44788756 0.45913515 -0.13820191 0.08484224 0.02600700 -0.00730757 -0.98773674 -0.85501874 -0.91082481 -0.87279346 -1.06296073 -0.76518323 -0.70786053 -0.77397072 -0.72786838 -0.97474169  .47360983^-0.115 .53578416^0.836 .34305318^1.338 .22377602^-0.618 .10784814^0.788 .08305995^0.313 .05139742^-0.142 .39360337^-2.509 .41038333^-2.083 .42216378^-2.158 .44588037^-1.958 .45367023^-2.343 .44354347^-1.725 .40838746^-1.733 .38514813^-2.010 .37742452^-1.929 .39373082^-2.476  .9085 .4036 .1813 .5371 .4309 .7543 .8870 .0124 .0376 .0314 .0508 .0195 .0850 .0836 .0449 .0543 .0136  11  ^  SHIPS SAS PROGRAM AUTOREG  ^  PROCEDURE  ESTIMATES OF AUTOCORRELATIONS LAG COVARIANCE CORRELATION ^1 9 8 7 6 5 4 3 2 1 0 1 2 3 4 5 6 7 8 9 1 0^0.124352 1.000000 1 0.00355163 0.028581 2 0.00974449 0.078382 3 -.00223853 -0.017986 4 0.00418543 0.033658 5 0.00484644 0.038974 8 0.00884074 0.071095 7^.000557991 0.004487 8 -.00536708 -0.043180 •• 9 -.00948228 -0.078254 •• 10^-0.0101487 -0.081613 11^-.00107887 -0.008674 12 .000047723 0.000384 PRELIMINARY MSE=^0.120 525 ESTIMATES OF THE AUTOREGR LAG^COEFFICIENT 1^-0.01512982 2^-0.07502589 3^0.01657736 4^-0.03228849 5^-0.04285336 6^-0.06827949 7^-0.00183554 8^0.04571747 9^0.07255588 10^0.07831513 11^0.00147192 12^-0.00806918  SSIVE PARAMETERS STD ERROR^T RATIO .04155723 -0.364067 .04158194 -1.805159 .04155148 0.398980 .04144783 -0.778538 .04142577 -1.034461 .04148398 -1.646718 .04148398 -0.039445 .04142577 1.103600 .04144783 1.750543 .04155146 1.884774 .04156194 0.035415 .04155723 -0.194170  YULE-WALKER ESTIMATES SSE^74.08571^DFE^579 MSE^0.1279548^ROOT MSE^0.3577074 SBC^875.0722^AIC^515.8938 REG RSO^0.8302^TOTAL RSO^0.8352  VARIABLE DF^B VALUE^STD ERROR^T RATIO APPROX PROF) INTERCPT 1^0.02688923^2.95548755^0.009^0.9927 LAGEH^1^-0.71433434^0.02075082^-34.424^0.0001 LNBPH^1^0.58975090^0.72885808^0.784^0.4333 LNBPL^1^0.28505838^0.88780583^0.414^0.6787 LNBPP^1^0.00646741^0.87819109^0.010^0.9924 LPL^1^0.01454587^0.02475088^0.588^0.5570 LPP^1^0.05789171^0.02818882^2.205^0.0279  14:08 MONDAY, MARCH 8, 1993 ^10  SHIPS SAS PROGRAM AUTOREG  ^  14:08 MONDAY, MARCH 8. 1993^9  PROCEDURE  DEPENDENT VARIABLE = LPH  ORDINARY LEAST SQUARES ESTIMATES SSE^78.47837^DFE 591 MSE^0.1294016^ROOT MSE^0.3597244 511.2405 SBC^617.3595^AIC REG RSO^0.8299^TOTAL RSO^0.8299 DURBIN-WATSON^1.9421  VARIABLE DF^B^VALUE INTERCPT LAGEH LNBPH LNBPL LNBPP LPL LPP LTCL LTCH LTCP LLIBOR LLUL LLUH LLUP 01 D2 D3 D4 05 D6 07 D8 09 D10  -0.07449956 -0.71472702 0.62175712 0.15293219 0.09776230 0.02805068 0.05330371 0.12810540 0.27192914 0.45040778 -0.13837878 0.15000049 -0.04894518 -0.00541538 -0.99605493 -0.83109398 -0.86163336 -0.83843484 -1.01818471 -0.71622669 -0.69131791 -0.74622943 -0.72333138 -0.94145421  STD ERROR^T RATIO APPROX PROB 2.79873568 0.02108982 0.87502546 0.64102042 0.62832223 0.02533495 0.02662562 0.44232837 0.50757690 0.32142807 0.20938101 0.10320461 0.07707106 0.04886637 0.39289297 0.40677263 0.41718704 0.43777835 0.44550342 0.43595571 0.40502871 0.38495013 0.37818400 0.39286318  -0.027 -33.890 0.921 0.239 0.156 1.107 2.002 0.285 0.536 1.401 -0.681 1.453 -0.635 -0.111 -2.536 -2.043 -2.065 -1.915 -2.281 -1.643 -1.707 -1.939 -1.913 -2.398  0.9788 0.0001 0.3574 0.8115 0.8784 0.2887 0.0457 0.7757 0.5923 0.1817 0.5089 0.1466 0.5256 0.9118 0.0115 0.0415 0.0393 0.0559 0.0229 0.1009 0.0884 0.0530 0.0563 0.0168  SHIPS SAS PROGRAM  14:08 MONDAY,^MARCH 8,^1993  AUTOREG^PROCEDURE 3 4 5 6 7 8 9 10 11 12  -0.00446778 -0.04663396 -0.04813403 -0.07927815 -0.01091641 0.04883142 0.07087797 0.07087375 -0.01708820 -0.01277502  0.04127253 0.04117007 0.04116574 0.04120715 0.04120715 0.04116574 0.04117007 0.04127253 0.04121373 0.04120092  -0.108251 -1.132715 -1.120690 -1.923893 -0.264918 1.186215 1.716732 1.719837 -0.414090 -0.310068  YULE-WALKER ESTIMATES SSE^75.33294^DFE^ 589 MSE^0.1278997^ROOT MSE^0.3516307 SBC^621.1349^AIC^508.1727 REG RSO^0.8075^TOTAL RSO^0.8324  VARIABLE DF INTERCPT LAGEH LNBPH LTCH 01 D2 D3 D4 05 De 07 D8 D9 D10  1 1 1 1 1 1 1 1 1 1 1 1 1 1  B VALUE  STD ERROR  1.21511804 -0.71762105 0.76186113 0.88616451 -0.99287707 -0.76238887 -0.75724382 -0.79229428 -0.95087088 -0.66831515 -0.57905174 -0.69232864 -0.62147823 -0.77134932  1.75390942 0.02048948 0.27548543 0.18080302 0.38298441 0.39360482 0.39386436 0.42881004 0.42636804 0.41461771 0.38482215 0.37368278 0.36285521 0.38480786  T RATIO APPROX PROB 0.693 -35.024 2.765 5.523 -2.592 -1.937 -1.923 -1.849 -2.230 -1.612 -1.505 -1.853 -1.713 -2.114  0.4887 0.0001 0.0059 0.0001 0.0098 0.0532 0.0550 0.0650 0.0281 0.1075 0.1329 0.0844 0.0873 0.0349  8  SHIPS SAS PROGRAM AUTOREG  ^  14:08 MONDAY, MARCH 8, 1993^7  PROCEDURE  DEPENDENT VARIABLE = LPH  ORDINARY LEAST SQUARES ESTIMATES SSE^78.00066^DFE^ 601 MSE^0.1297848^ROOT MSE^0.3602586 SBC^565.2806^AIC^503.3779 REG RSO^0.8265^TOTAL RSO^0.8285 DURBIN-WATSON^1.9028  VARIABLE DF^B^VALUE^STD ERROR^T RATIO APPROX PROB INTERCPT LAGEH LNBPH LTCH D1 D2 D3 D4 D5 D6 D7 D8 D9 D10  1.02310945^1.52880638^0.670 -0.71661744^0.02085754^-34.358 0.78534511^0.23771287^3.304 0.90866149^0.13886804^8.543 -0.99208251^0.38054084^-2.807 -0.73554081^0.38872541^-1.892 -0.73843760^0.38918458^-1.897 -0.75982098^0.41558121^-1.828 -0.90933944^0.41398110^-2.197 -0.64887313^0.40492531^-1.602 -0.57272250^0.38235448^-1.498 -0.69188500^0.37407532^-1.850 -0.82124197^0.38620504^-1.898 -0.77389528^0.38652175^-2.111  0.5031 0.0001 0.0010 0.0001 0.0094 0.0589 0.0582 0.0880 0.0284 0.1096 0.1347 0.0649 0.0903 0.0351  ESTIMATES OF AUTOCORRELATIONS LAG^COVARIANCE^CORRELATION^1 9 8 7 6 5 4 3 2^1 0 1 2 3 4 5 6 7 8 9^1 0^0.12683^1.000000 • 1^0.00610602^0.048143 • 2^0.0120837^0.095275 3^0.00103293^0.008144 4^0.00872718^0.053041 5^0.00834074^0.049994 •• 8^.0.0108872^0.085683 7^0.00238833^0.018857 8^-.00492757^-0.038852 9^-.00854849^-0.087385 10^-.00910827^-0.071799 11^0.00103594^0.008188 12^.000459508^0.003823 PRELIMINARY MSE=^0.1225595 ESTIMATES OF THE AUTOREGRESSIVE PARAMETERS LAG^COEFFICIENT^STD ERROR^T RATIO 1^-0.03021811^0.04120092^-0.733433 2^-0.08884797^0.04121373^-2.155788  1■3  O  SHIPS SAS PROGRAM^  14:08 MONDAY, MARCH 8, 1993^1  CONTENTS PROCEDURE CONTENTS OF SAS MEMBER WORK.TEMP CREATED BY OS JOB 81146CINS^ON CPUID 83.3081-020804 ^AT 14:08 MONDAY, MARCH 8, 1993^BY SAS RELEASE 5.08^INFILE(DSN=SHIPSASI VOL=SER=MTS028/^INFILEIDSN=SHIPSAS2^VOL=SER=MTS013/^DSNAME=-WORK00^OBSERVATIONS PER TRACK =42 ^BLKSIZE=12940^LRECL=308 GENERATED BY DATA NUMBER OF OBSERVATIONS: 815^NUMBER OF VARIABLES: ^38 MEMTYPE:^DATA ----ALPHABETIC LIST OF VARIABL S AND ATTRIBUTES ^ VARIABLE TYPE LENGT POSITION FORMAT 2 AGEH^NUM 12 15 DI^NUM 118 18 02^NUM 124 17 03^NUM 132 18 04^NUM 140 19 05^NUM 148 20 D6^NUM 158 21 07^NUM 164 22 D8^NUM 1/2 23 09^NUM 180 24 D10^NUM 188 28 LAGEH^NUM 204 11 LIBOR^NUM 84 35 LLIBOR^NUM 278 37 LLUH^NUM 292 36 LLUL^NUM 284 38 LLUP^NUM 300 30 LNBPH^NUM 238 29 LNBPL^NUM 228 31 LNBPP^NUM 244 25 LPH^NUM 198 27 LPL^NUM 212 28 LPP^NUM 220 33 LICH^NUM 280 32 LIU^NUM 252 34 LTCP^NUM 288 13 LUH^NUM 100 12 LUL^NUM 92 14 LUP^NUM 108 4 NBPH^NUM 28 3 NBPL^NUM 20 5 NBPP^NUM 38 1 PH^NUM 4 6 PL^NUM 44 7 PP^NUM 52 9 TCH^NUM 88 8 TCL^NUM 80 10 TCP^NUM 78 SOURCE RECORDS DATA TEMP; Infile Mel; Input PH AGEH NBPL NBPH NBPP PL PP TCL TCH TCP; 1nf1le f1162;  INFORMAT^LABEL  SHIPS SAS PROGRAM^ Input LIBOR LUL LU11 LUP 01 D2 03 D4 05 08 07 08 09 010; LPH^= LOG (PH); LAGER = LOG (AGEH); LPL^= LOG (PL); LPP^= LOG (PP); LNBPL = LOG (NBPL); LNBPH = LOG (NBPH); LNBPP = LOG (NBPP); LTCL = LOG (ICU; LICH . LOG ITCH); LTCP = LOG (TCP); LLIBOR= LOG (LIBOR); LLUL . LOG (LUL); LLUN = LOG (LUH); LLUP . LOG (LUP);  14:08 MONDAY, MARCH 8, 1993^2  SHIPS SAS PROGRAM VARIABLE  PH AGEH NBPL NBPH NBPP PL PP 7CL ICH TCP LIBOR LUL LUH LUP DI D2 D3 D4 D5 D6 D7 08 09 DIO LPH LAGEH LPL LPP LNBPL LNBPH INBPP LTCL LICH LTCP LLIBOR LLUL LLUH LLUP  14:08 MONDAY,^MARCH 8, MEAN  15 15 15 15 15 15 15 15 15 15 15 15 15 15 15 15 15 15 15 15 15 15 15 15 15 15 15 15 15 15 15 15 15 15 15 15 15 15  193.58721951 10.58373984 538.97235172 445.99186992 376.58264065 165.59692683 138.39489431 6.58406504 5.29300813 4.25609756 9.29918699 2259.83252033 1510.54308943 1181.53008130 0.05365854 0.05691057 0.10081301 0.07642276 0.10731707 0.13333333 0.14798748 0.12032520 0.08455285 0.12032520 4.95552489 2.16284015 4.82716669 4.88650287 6.25806138 6.05393800 5.89498042 1.82281039 1.59934552 1.36244899 2.19893793 7.53277693 7.18173298 6.80848219  STANDARD DEVIATION  MINIMUM VALUE  MAXIMUM VALUE  STD ERROR OF MEAN  139.27503825 5.57221431 134.70289741 134.29122335 99.82501476 122.52249720 91.47332095 2.27110580 1.87233412 1.69099877 2.61518044 1437.60926890 801.82957565 758.03734762 0.22552605 0.23186024 0.30132604 0.28588960 0.30976789 0.34021134 0.35535677 0.32560626 0.27844160 0.32580626 0.85589374 0.71900844 0.79228020 0.74787851 0.25257220 0.30869713 0.27159042 0.35595628 0.37320953 0.42334032 0.24757585 0.61125803 0.53033132 0.75425747  9.70000000 0.50000000 348.00000000 272.50000000 233.33300000 11.24000000 6.27000000 3.40000000 2.50000000 1.80000000 4.20000000 118.00000000 344.00000000 176.00000000 0.00000000 0.00000000 0.00000000 0.00000000 0.00000000 0.00000000 0.00000000 0.00000000 0.00000000 0.00000000 2.27212589 -0.69314718 2.41947884 1.83577635 5.85220248 5.60763861 5.45246662 1.22377543 0.91629073 0.58778666 1.43508453 4.77088462 5.84064188 5.17048400  738.7900000 28.0000000 784.0000000 662.5000000 533.3330000 721.3100000 439.0800000 10.6000000 8.5000000 6.8000000 18.5000000 6186.0000000 3887.0000000 3184.0000000 1.0000000 1.0000000 1.0000000 1.0000000 1.0000000 1.0000000 1.0000000 1.0000000 1.0000000 1.0000000 8.6023029 3.3322045 8.5810890 6.0846361 8.8385678 6.4980206 6.2791460 2.3608540 2.1400662 1.9169226 2.9177707 8.7300440 8.2653929 8.0658935  5.61611159 0.22469337 5.43174508 5.41514478 4.01726833 4.94058394 3.68856031 0.09157982 0.07549980 0.06818757 0.10545425 57.97000079 32.32482437 30.48636830 0.00909409 0.00934951 0.01215084 0.01072170 0.01249105 0.01371865 0.01432937 0.01312971 0.01122785 0.01312971 0.03450490 0.02899322 0.03194701 0.03014931 0.01016469 0.01238722 0.01095158 0.01435354 0.01504928 0.01707073 0.00998322 0.02464822 0.02138502 0.03041460  SUM^VARIANCE  119058.1400 6509.0000 331468.0000 274285.0000 231598.3240 101842.1100 85112.8600 4049.2000 3255.2000 2617.5000 5719.0000 1389797.0000 928984.0000 714341.0000 33.0000 35.0000 82.0000 47.0000 66.0000 82.0000 91.0000 74.0000 52.0000 74.0000 3047.6477 1330.1467 2968.7075 2882.1993 3848.7077 3723.1725 3825.4007 1121.0284 983.5975 837.9061 1351.1168 4832.6578 4416.7658 4187.2042  19397.5363 31.0498 18144.8706 18034.1327 9925.1436 15011.7623 8367.3684 5.1579 3.5056 2.8595 8.8392 2066720.4100 642609.9766 571592.4710 0.0509 0.0538 0.0908 0.0707 0.0960 0.1157 0.1263 0.1060 0.0775 0.1060 0.7322 0.5170 0.6277 0.5590 0.0638 0.0941 0.0738 0.1287 0.1393 0.1792 0.0613 0.3738 0.2813 0.5889  1993^3  C.V 71.944 52.649 24.993 30.111 26.455 73.988 66.096 34.494 35.374 39.731 28.123 63.616 53.069 65.090 420.299 407.412 298.898 347.919 288.647 255.159 240.159 270.805 329.311 270.805 17.267 33.244 16.413 15.954 4.038 5.066 4.607 19.528 23.335 31.072 11.269 8.115 7.384 11.078  ^ ^  SHIPS SAS PROGRAM AUTOREG  ^  14:08 MONDAY, MARCH 8, 1993 ^4  PROCEDURE  DEPENDENT VARIABLE = LPH  ORDINARY LEAST SQUARES ESTIMATES SSE^82.12043^DFE^ 611 MSE^0.1344033^ROOT MSE^0.3666106 SBC^532.1181^AIC^515.0316 REG RSO^0.8173^TOTAL RSQ^0.8113 DURBIN-WATSON 1.8218  VARIABLE DF  ^  B VALUE^STD ERROR^I RATIO APPROX PROB  INTERCPT^1^1.19681777^0.343893234^3.480^0.0005 LAGEH^1^-0.71404302^0.020721359^-34.459^0.0001 LNBPH^I^0.55497803^0.065777949^8.437^0.0001 LTCH^I^1.21503886^0.053921347^22.534^0.0001 ESTIMATES OF AUTOCORRELATIONS LAG COVARIANCE CORRELATION 0^0.133529^1.000000 1^0.0113835^0.085251 2^0.0163304^0.122299 3^0.0048285^0.036181 4^0.0115278^0.088330 5^0.0100891^0.075557 6^0.014^0.104848 1 0.00494735^0.037051 8^.00293882^-0.022009 9^-.00811757^-0.045814 10 -.00769959^-0.057862 11 0.00176116^0.013189 12 0.00152568^0.011428  1 9 8 7 6 5 4 3 2 1 0 1 2 3 4 5 6 7 8 9 1  PRELIMINARY MSE=^0.1273627 ESTIMATES OF THE AUTOREGRESSIVE PARAMETERS LAG^COEFFICIENT^STD ERROR^I RATIO 1^-0.05881484^0.04085135^-1.390785 2^-0.10430629^0.04091074^-2.549607 3^-0.01610908^0.04104264^-0.392518 4^-0.06910053^0.04097764^-1.888298 5^-0.05894918^0.04103291^-1.436832 6^-0.08823231^0.04109883^-2.098169 7^-0.01522339^0.04109883^-0.370409 ^a ^0.04538884^0.04103291^1.106108 ^9 ^0.05878154^0.04097764^1.433990 10^0.08638749^0.04104264^1.817038 11^-0.01787602^0.04091074^-0.436952 12^-0.01921021^0.04085135^-0.470247  SHIPS SAS PROGRAM AUTOREG  ^  14;08 MONDAY, MARCH 8, 1993^5  PROCEDURE  YULE•WALKER ESTIMATES SSE^78.27552^DFE^599 MSE^0.130877^ROOT MSE^0.3814927 SBC^580.5031^AIC^509.7571 REG RSO^0.7782^TOTAL RSO^0.8269  VARIABLE OF  B VALUE  STD ERROR  1 1 1 I  1.18956158 -0.71432178 0.55825281 1.20719181  0.463847819 0.020358053 0.088375080 0.072482699  INTERCPT LAGEH LN13PH LTCH  T RATIO APPROX PROS 2.585 -35.088 8.317 18.855  0.0108 0.0001 0.0001 0.0001  ^ ^  14:08 MONDAY, MARCH 8, 1883^6  SHIPS SAS PROGRAM^ PLOT OF LPHIRES'LP111PRED^SYMBOL USED IS a LPH1RES 1.25^  1.00^  •  I  a  a a a^a^a 8 a 0.75^•^ A^a a a a^a a a aaaaaa a a^ a a^ aa ^ a aa aa a a^aa 0.50 1^ a a^a a a^as^a a^aaa a a a^aaa aa^a a aaaa a^a a a a^aa aa^aaa aaaaaa^aaaaa a aaaaaa^a aa aaaa aa aa aa a aaaaaa ^aa aa a a^ 0.25 1^ aaaaaaaaaaa^aa aaaaa aaa aaaaaaaaaaa a aa a a a aa a^a a aaaaaa aaa a a aaaaaaaa a ^aaa^aa aaa a^a a aaaaaa aaa aa aa aa aaaa a a 0 00 •^a^ aa^aaaaaa^aa^aaaaa a aaaaaa aa aaa^aaa a a a aa aaaaa^aa^aaaaaa aaaaa aa a a^aaaaa a a a as as a a a aa a aa^aa a a a^a aa a a^a a as aa a aaaaaa^aa a aaa aaa^a aaaa^a a^ aa a a a aa 8 8^a^a a^8 add 8^a^a -0.25^• a^a :^ aaa aa^a aaa a a a aa aa a a aaa a a a as^a^88^ a a a a^ a^aa a a a a^ -0.50^•^ a^ a^aaa a a a a a a 8^& A a^a^a a a aa a^ a aa^a a^ A^ a^ a^ a a -0.75^• aa a a a^ -^ 1.00^•^  a a  a a a  a  a a a a^  -1.25^•^  a^  a  a^  -1.50^• ♦^  3.0  •^ 3.4  a^  3.8  •^ 4.2  •^ 4.8  •^ 5.0  •^ 5.4  LPH1PRED NOTE:^160 OBS HIDDEN  •^ 5.8  •^ 8.2  •^ 8.6  •^ 7.0  •^ 7.4  •7.8  127  APPENDIX III TEST OF SECOND-HAND SHIP PRICE FUNCTION CHANGE OVER TIME  128 PART I. INTRODUCTION  As analyzed in Chapter IV and Chapter V, the ship sale and purchase market performance is dominated by various related market conditions which also determine the function of the second-hand prices of ships. The ship price function might change over time due to substantial changes of one or more important market conditions. For example, a revolution in the technology of shipbuilding which leads to substantial savings in building new ships might result in a surge for newbuildings. As a result, the second-hand ship prices might drop substantially as buyers would prefer to order new ships, and the second-hand price function structure might be affected.  The structure of a market is a defined as a description of the behaviour of buyers and sellers in that market. The ship sale and purchase market has been regarded competitive since the existence of the modern telecommunication techniques. The second-hand ship price function might change if the sale and purchase market structure changes.  For the last 15 years or so, we have not observed neither any revolutionary changes in the related market conditions nor any market structure change in the ship sale and purchase market.  Therefore, qualitatively speaking, the assumption that the second-  129  hand price function has remained unchanged over time might be realistic.  130 PART II. STATISTICAL TESTING  Suppose the whole sample with time series data is divided into two sub samples by time on the variables Y and X, the one containing n observations and the other m observations, and the two sub samples are separately used for the estimation of the relationship between Y and X. Two estimates of the same relationship for two different periods of time can be obtained.  If the two estimated relationships differ significantly, it is concluded that the relationship is changing from one sample to the other.  Suppose Y represents the second-hand prices, X the related independent variables. The whole sample (from 1982 to 1991) is divided into two samples, one from 1982 to 1986, another from 1987 to 1991. The following estimated second-hand price functions can be obtained using Ordinary Least Squares:  Yi = a l + b 1 X 1 and Y 2 = a 2 + b2 X2  Are the tow estimated functions significantly different? Does the price function change ship over time (b 1 * b 2 )? Or, is the  131 difference insignificant so that it may be attributed to chance, in which case it may be concluded that the price function is stable over time?  To answer these questions an F test is suggested. 1 There are following steps:  Step 1.^Estimate the whole (called "pooled") sample covering 1982-1991 and obtain the pooled unexplained variation  (SSE) with (n+m-K) degrees of freedom. (K is the total number of coefficients.)  Step 2.^Estimate the two sub samples separately and obtain the unexplained variations (SSE, and SSE 2 ) with (n-K) degrees of freedom for the first sample and (m-K) for the second sample respectively.  Step 3.^Add the two unexplained variations to form the total unexplained variation with (n-K) + (m-K) degrees of freedom.  Step 4.^Subtract the above sum of residual variations from the pooled residual variance of Step 1, and obtain the difference SSE. = SSE - (SSE 1 + SSE 2 ) with (n+m-K) - (n+m1^  Koutsoyiannis A. (1977) Theory of Econometrics, The MacMillan Press Ltd, London, England, 164-167  132 2K) = K degrees of freedom.  Step 5.^Form the ratio SSE* / K  F* - ^ (SSE i +SSE 2 ) / (n+m-2K) Step 6.^Compare the observed F * with the theoretical value of F 0.05 (or other level of significance) with v 1 = K and v2 = (n+m-2K) degrees of freedom.  If F * > F 0.05 accept that the two price functions differ significantly, or the price functions change over time. Vice, versa.  For the regression model selected in Chapter VI, the following values are available:  SSE^= 82.12043 SSE^= 30.09181 SSE 2^= 49.01782 SSE *^= 1.01080 n^= 253 m^= 362 n+m^= 615 (total observations)  K^= 4  133  F e = 1.8849 < F 0  = 2.37.  Therefore, the two relationships estimated by two sub samples do not differ significantly. In other words, the second-hand ship price function did not change between the two periods, 1982-1986 and 1987-1991. This result is consistent with the assumption made in Chapter VI. Similar tests can also be done for other different two periods, and similar results are expected.  134 REFERENCES Ademuni-Odeke (1988) Shipping in International Trade Relations, Avebury Gower Publishing Company Limited, England. Beenstock M. (1985) "A Theory of Ship Prices," Maritime Policy and Management 1985 Vol 12, No. 3, 215-225. Branch, A.E. (1988) Economics of Shipping Practice and Management (Second Edition), Chapman and Hall, London. Brauner A. (1992) "Fleet Replacement - the Key Issue is Credit," Seatrade Business Review, January 1992, 41. Bross S.R. (1956) Ocean Shipping, Cornell Maritime Press, Cambridge, Maryland, USA. Charemza W. and M. Gronicki (1981) "An Econometric Model of World Shipping and Shipbuilding," Maritime Ploicty and Management, 1981 Vol 8 No. 1, 21-30. Chrzanowski I. (1985) An Introduction to Shipping Economics, Fairplay Publications, Mayhew McCrimmon Printers Ltd., London England. Cockburn I. and M. Frank (1991) "Market Conditions and Retirement of Physical Capital: Evidence From Oil Tankers", Seminar, November 24, 1991, UBC. Drewry Shipping Consultants Ltd. (1991) Forecast Dry Bulk Carrier Profitability, Drewry Shipping Publications Ltd., England. Drewry Shipping Consultants Ltd. (1990) Bulk Fleet Growth, Drewry Shipping Publications Ltd., England.  Drewry Shipping Consultants Ltd. (1990) Handymax Bulk Carriers, Drewry Shipping Publications Ltd., England. Drewry Shipping Consultants Ltd. (1991) Shipping Statistics and Economics, Monthly 1981-1992, Drewry Shipping Publications Ltd., England. Dunlop A. (1992) "Sale and Purchase Activity Reflects Weak Charter Markets," International Marine Business, January/February 1992, 6. Frankel E.G. (1987) The World Shipping Industry, Croom Helm Publishers Ltd, Kent, England. Freund R.J. and R.C. Littel (1981) SAS For Linear Models - A Guide to the ANOVA and GLM Procedures, SAS Institute Inc., NC, USA.  135 Gaunt I. (1992) "Reducing Reliance on Bank Finance," Seatrade Business Review, January 1992, 27-29. Goss R.O. (1977) Advances in Maritime Economics, Cambridge University Press, Cambridge, England. Hale C. and A. Vanags "The Market for Secondhand. Ships: Some Results on Efficiency Using Cointegration," Maritime Policy and Management 1992, Vol 19, No. 1, 31-39. Koutsoyiannis A. (1983) Theory of Econometrics - An Introductory Exposition of Econometric Methods, The Macmillan Press Ltd, London England. Lyons C. (1992) "Investment Strategy - The Making of Meyer's Millions," Seatrade Business Review, January 1992, 35-37. Marlow P.B. (1991) "Shipping and Investment Incentives: a Trilogy. Part 2. Investment Incentives for Shipping," Maritime Policy and Management 1991 Vol 18 No.3, 201-216. Metaxas B.N. (1971) The Economics of Tramp Shipping, The Athlone Press of the University of London, England.  Muth J.F. (1961) "Rational Expectations and the Theory of Price Movements," Econometrica, Vol 29 No. 3 (July 1961), 315-335 Neter J., W. Wasserman and M. H. Kutner (1990) Applied Statistical Models, Richard D. Irwin, Inc. Boston, MA Oum T.H. (1989) "Alternative Demand Models and Their Elasticity Estimates," Journal of Transportation and Economic Policy, May 1989, 163-187. Proctor I.L. (1970) A Statistical Investigation of the Ocean Charter Market, Master's Thesis, Commerce UBC. Rinman T. and R. Brodefors (1983) The Commercial History of Shipping, Elanders Boktryckeri AB, Kungsbacka Sweden. Shimojo T. (1979) Economic Analysis of Shipping Freights, Hayashi Obundo Printing Co., Ltd, Kobe Japan. Sloggett J.E. (1984) Shipping Finance, Fairplay Publications, Mayhew McCrimmon Printers Ltd., London England. Stopford, M. (1988) Maritime Economics, Allen & Unwin Inc., Winchester Mass. Stopford M. (1990) "Analysis - Ship Finance," Seatrade Business Review, May/June 1990, 23-29.  136  -  Zannetos Z.S. (1966) The Theory of Oil Tankshio Rates, the MIT Press, Cambridge Mass.  


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