Open Collections

UBC Theses and Dissertations

UBC Theses Logo

UBC Theses and Dissertations

Redirecting or reducing demand? the efficacy of Chinese government efforts to control the housing market Yang, Yang 2014

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

Item Metadata

Download

Media
24-ubc_2014_september_yang_yang.pdf [ 30.5MB ]
Metadata
JSON: 24-1.0102479.json
JSON-LD: 24-1.0102479-ld.json
RDF/XML (Pretty): 24-1.0102479-rdf.xml
RDF/JSON: 24-1.0102479-rdf.json
Turtle: 24-1.0102479-turtle.txt
N-Triples: 24-1.0102479-rdf-ntriples.txt
Original Record: 24-1.0102479-source.json
Full Text
24-1.0102479-fulltext.txt
Citation
24-1.0102479.ris

Full Text

 Redirecting or Reducing Demand? The Efficacy of Chinese Government Efforts to Control the Housing Market   by  Yang Yang  B.Comm., University of British Columbia, 2014     A THESIS SUBMITTED PARTIAL FULFILLMENT OF THE REQUIREMENTS FOR THE DEGREE OF  MASTER OF SCIENCE IN BUSINESS ADMINISTRATION  in  The Faculty of Graduate and Postdoctoral Studies  (Urban Land Economics)    THE UNIVERSITY OF BRITISH COLMBIA  (Vancouver)  July 2014  © Yang Yang, 2014   	   ii	  Abstract  This paper examines the success of Chinese government policies in restricting housing market activity. The focus is whether policies designed to restrict speculation and cool the higher end of the housing market, succeeded in doing so, and if successful, did they result in an overall reduction in activity. The particular government interventions we examine are the restrictions on lending imposed from late 2010 to early 2011 that targeted the higher end market by only affecting units over 90 square meters in size. As well as targeting only one segment of the market, these policies were not implemented uniformly across cities. Our identification strategy is to exploit the differential restrictions across unit size to assess the effects of these policies on overall activity, the targeted high-end market, and housing sub-markets less affected by restrictions. The empirical analysis uses a panel data from nine Chinese cities that aggregates activity by sub-markets from 2006Q1 to 2013Q4. The detail in the data allow us to look at relative activity by housing sub-market within individual cities, the difference that we exploit to identify the effects of the government restrictions.     	   iii	  Preface 	  The	  dissertation	  is	  original,	  unpublished,	  independent	  work	  by	  Yang	  Yang.	  The	  research	  is	  under	  supervision	  of	  Prof.	  Tsur	  Somerville.	   	  	  	    	   iv	  Table of Contents ABSTRACT....................................................................................................................................ii PREFACE.......................................................................................................................................iii TABLE OF CONTENTS................................................................................................................iv LIST OF FIGURES........................................................................................................................vi LIST OF TABLES.........................................................................................................................vii 1. INTRODUCTION .................................................................................................................................... 1 2. RELATED LITERATURE ....................................................................................................................... 7 3. BACKGROUND OF CHINESE REAL ESTATE MARKET ............................................................... 12 3.1 THE EVOLUTION OF CURRENT CHINESE REAL ESTATE POLICIES ................................................... 12 3.1.1. Financial Policies ..................................................................................................................... 12 3.1.2. Housing Finance Options ......................................................................................................... 17 3.1.3. Tax Policies .............................................................................................................................. 20 3.1.4. Land Policies ............................................................................................................................ 21 3.1.5. Additional Restrictions and Regulations .................................................................................. 22 3.2. REAL ESTATE DEVELOPMENT IN CHINA .......................................................................................... 24 3.2.1. Decision-making Phase ............................................................................................................ 24 3.2.2. Land Use Rights Acquisition Phase ......................................................................................... 25 3.2.3. Demolition Phase ..................................................................................................................... 28 3.2.4. Land Development .................................................................................................................... 29 	   v	  3.2.5. Construction and Sale Phases .................................................................................................. 30 3.3. BACKGROUND OF HOUSING POLICY IN CHINA ................................................................................ 31 3.3.1. Pre-reform Stage: 1949-1977 .................................................................................................. 31 3.3.2. First Stage: 1978-1987 ............................................................................................................. 32 3.3.3. Second Stage: 1988 to 1997 ..................................................................................................... 35 3.3.4. Third Stage: 1998 to 2002 ........................................................................................................ 38 3.3.5. Fourth Stage: 2003 to Present ................................................................................................. 41 4. HOUSING MARKET DATA ................................................................................................................. 48 4.1 DATA ................................................................................................................................................. 48 4.1.1 Housing Price and Sales Volume .............................................................................................. 51 4.1.2 Macroeconomic Data ................................................................................................................ 52 4.1.3 Regulatory Variables ................................................................................................................. 52 4.2 DESCRIPTIVE STATISTICS .................................................................................................................. 54 5. EMPIRICAL ANALYSIS ...................................................................................................................... 69 6. CONCLUSION ....................................................................................................................................... 81 BIBLIOGRAPHY ....................................................................................................................................... 84 APPENDICES ............................................................................................................................................ 88 APPENDIX 1: SUMMARY OF POLICIES AT CITY AND NATIONAL LEVEL FROM 2007 TO 2013 .................. 88 APPENDIX 2: SUMMARY OF NATIONAL POLICIES FROM 1978 TO 2013 ................................................... 92 	  	   vi	  List of Figures   	   Figure 3.1: Required Minimum Down Payment ................................................................... 13 Figure 3.2: Long Term Benchmark Interest Rate from 2006 to 2012 .................................. 14 Figure 3.3: Impact of Restrictions on Demand and Supply .................................................. 15 Figure 3.4: Change of Required Reserved Ratio From 2004 to 2012 .................................. 16 Figure 3.6: Average Land Transaction Price from 2000 to 2012 ......................................... 27 Figure 3.7: Floor Space of Newly Built Residential Buildings in Urban and Rural Areas from 1978 to 1997 ......................................................................................................... 37 Figure 3.8: Per Capita Living Spaces in Urban and Rural Areas from 1978 to 1997 .......... 38 Figure 3.9: Investment in Real Estate from 1998 to 2002 .................................................... 39 Figure 3.10: Floor Space of Building Completed and Sold from 1998 to 2002 ................... 40 Figure 3.11: Growth Rate of Investment in Real Estate Development from 1999 to 2011 .. 42 Figure 3.12:Annual Growth in Average Selling Price of Residential Housing .................... 46 Figure 4.1: Location of Nine Selected Cities ........................................................................ 50 Figure 4.2: Changes in Nominal Housing Prices .................................................................. 57    	   vii	  List of Tables  Table	  3.1:	  Information	  of	  Minimum	  Down	  Payment	  and	  Interest	  Rate	  for	  Major	  Banks	  in	  2014	  ....................................................................................................................................................	  18	  Table	  3.2:	  The	  lowest	  Bound	  of	  Interest	  Rate	  Relative	  to	  the	  Benchmark	  Rate	  ..............	  18	  Table 3.3: Major Real Estate Tax Changes	  ..........................................................................................	  21	  Table 3.4: Land Use Right Time Limits By Use	  .........................................................................	  25	  Table	  4.1:	  City	  Facts	  ....................................................................................................................................	  49	  Table	  4.2: Summary of the POLICY Variable	  ....................................................................................	  53	  Table 4.3: Descriptive Statistics of the Panel Data	  ....................................................................	  56	  Table 4.4: Descriptive Statistics of Year on Year % Price Change from 2006 to 2013	  ....................................................................................................................................................................	  58	  Table	  4.5: Descriptive Statistics of Key Variables by City	  ..................................................	  61	  Table	  4.6: Descriptive Statistics of Key Variables by Size Category	  .............................	  66	  Table 5.1: Housing demand, Dependent variable: log (average housing price changes).......75 Table 5.2: Housing demand, Dependent variable: log (unit sold)..........................................78 Table 5.3: Housing demand, Dependent variable: log (total floor space sold).......................80 	  	   1	  1. Introduction Since 2009, the exponential growth of the real estate market in China has resulted in considerable attention from the Chinese public and government. High price- to-income and price-to-rent ratios, along with high vacancy rates for completed buildings have caused an uproar from average citizens, increasing the pressure on the Chinese government to reinforce and implement regulations in order to stabilize the housing market. The government has responded with several waves of tightened housing policies since early 2010 to combat soaring of housing price without jeopardizing market stability.  This thesis investigates the effectiveness of government policies in controlling local housing markets.  The focus is not just on aggregate affects on price levels, but whether the policies successfully slowed the higher end of the market while preserving stability in the sub-markets more accessible to marginal homebuyers.  Government interventions include restrictions on the number of homes a buyer may purchase and differences in the minimum lending and down payment rates based on the number and size of homes.  These favor households who own just one unit and of a size under 90 square meters. The central government’s concern is with an overheated real estate market, so the policies aim to reduce house price appreciation and the volume of new construction as a means, to reach a “reasonable” price-to-income ratio with adequate unit occupancy rates. The identification of the effects of these policies imposed from late 2010 to early 2011 to restricting demand for bigger 	   2	  units and second and third homes, comes both from the time series around their implementation and exploits the variation in when and to what extent they were imposed by different local governments. Different attitudes of developers, homebuyers and government towards the macro-economic control policies has been brought to the public’s attention and raises several questions. Do the control polices have a negative impact on the housing market? Do the policies cause a temporary or permanent impact on the housing market? Also, if the policies brought a significant negative shock to the housing market, would this result in a severe downturn? More importantly, do the policies divert demand from one segment to another in a way that undermines government expectations? These questions have provided the motivation for the research in this paper.  The key question to be addressed concerns the differential impact of restrictive policies across groups. The empirical work in this paper uses a panel data from nine cities that segments aggregate market activities into three sub-markets (home size below and equal to 90m2, 90 to 140 m2, and above 140 m2) from 2006Q1 to 2013Q4. Average housing prices, sales volume, and total floor space sold in each sub-market provide an opportunity to assess the market activities within individual cities, and identify the impact of restrictive polices. I examine new construction, but this variable is not segmented. To preview the results, the econometric analysis indicates that real estate control policies during this period have a clear negative correlation with on housing price changes and quantity demanded. Households respond to the implementation of new down payment policies before 	   3	  they take effect. Minimum down payments required by bank loans have a statistically different from zero negative impact on the quarterly change of housing price, unit sales and total floor space sold over two quarters. In addition, the negative correlation of the increases in the minimum down payment required by bank loans is larger than the correlation of the same changes for loans from a household’s Housing Provident Fund. Surprisingly, restrictions on the purchase of a third home positively contributed to housing price increases, sales volume and total floor space sold. More specifically, even though the correlation of purchase restriction with price changes is not statistically significant different from zero, the implementation of the purchase restriction results in a 24.1% increase in quarterly sales volume and a 60.8% increase in total floor space sold. The results suggest that the differential policy effects across sub-markets are significant. While the average price changes for bigger units are minimal, housing demand is diverted from medium- to-big homes to smaller units, though not in sufficient volume to offset the overall dampening effect of the policies.  This study provides valuable insight about the impact of lending supply shocks and purchase restrictive policies on the housing market. From a policy perspective, the implications of this study could help policy makers with their decision-making regarding future policy adjustment and setting. From a developers’ perspective, this research could help them to better predict future housing market movements and to develop strategies accordingly.  The experiences of China’s effort to control their real estate market in the last four years offers a useful case study of the nexus between real estate restraining policy and the dynamic of the 	   4	  housing market for countries experiencing immigration, particularly of wealthy households. For instance, in cities like Vancouver, San Francisco, and Sydney where a large number of immigrants and foreign capital appear to have had an effect on demand for higher end properties, there is a warning that effects on one segment can result in change son other segments that may not be consistent with overall objectives.  The major contribution of this research is that it provides sub-market aggregated transaction data, which allows us to observe differential policy effects on each segment and suggestions regarding policy cross-elasticity of substitution. This will in turn allow for the assessment of policy effectiveness, by examining whether the policy successfully cooled down the higher end of the housing market and encouraged consumers to purchase on average smaller units. In addition, while considerable literatures have been focused on the impact of monetary policies on the housing market in China (Ye and Wu, 2008; Cao and Keivani, 2013; Gerlach and Pang, 2003), the macroeconomic control policies introduced in 2010 have been largely neglected. Therefore, this research attempts to fill the gap by providing an updated impact of macroeconomic control policies on China’s housing market in the past five years. Also, while many Chinese journals that study policy effects rely on national aggregated data in their statistical analysis of current trends (Yu, 2010; Wei et al., 2013), the detailed city data used in this paper will provide a more accurate assessment of policy impact to individual cities, which captures heterogeneity in each city. Moreover, the available literatures in China in the same field provide subjective analyses without econometric models. This paper uses regression models to show a more detailed analysis 	   5	  of policy effects in specific cities and across market segments.  Even though purchase and financing restrictions were imposed to restrict quantity demanded, in reality, households are astute enough to evade these restrictions on the purchase of the third or more homes. The most pervasive method to avoid the purchase restriction is to fake marriage or divorce. Since the purchase restriction is imposed on a family instead of an individual person, one couple becomes two families after divorce, therefore the total number of homes allowed to buy increases from 2 units to 4 units. In addition, the purchase restriction in most cities allows non-citizens to purchase one home and forbids non-residents to purchase homes. In response to the rule, a non-citizen can increase the quota through a fake marriage and a non-resident can buy homes through a forged residence permit or a fake personal income tax certificate. Also, commercial banks have created the Personal Housing Relay Loan, which allows parents and the child to act as co-borrowers and co-owners of the purchased home, in order to avoid purchase and financing restrictions.  In this way, a considerable number of arranged transactions are undertaken given the backdrop of purchase and financing restrictions, indicating that the actual quantity demanded has been underestimated. Data regarding the arranged transactions is not available. To be conservative, the empirical analysis in this research does not control for the arranged transactions. In this sense, the estimate of policy effects in the empirical work undervalues the actual policy impacts on housing prices and quantities. 	  The paper is structured as follows. Section 2 introduces the current restrictive policies and real 	   6	  estate development process in China. It also reviews developments in Chinese housing markets from 1949 to present. Section 3 discusses the previous literature related to our topic. Summary statistics will be presented in Section 4 and Section 5 provides the econometric model and empirical analysis. Finally, the paper looks at possible explanations for why the micro-control policies are redirecting housing demand and provides concluding remarks in Section 6. 	  	   	  	   7	  2. Related Literature The impacts of policy constraints on house price that arise in this study have been discussed in previous literature. Ortalo-Magne and Rady (2006) provide a theoretical framework, showing that credit constraints on first-time homebuyers can have a dramatic impact on the overall housing market. Duca, Muellbauer and Murphy (2011) show that an easing of mortgage credit standards has a significant short-term effect on home prices. Mayer and Somerville (2000) find that land use regulation lowers the level of new construction. They provide evidence for the importance of using house price changes, rather than levels of house price, to explain new starts, and to eliminate the issue of serial correlation. Part of my regression is in fact based on this idea; by regressing log price change on its lagged values, the relationship between housing price and policy constraints can be further observed. To reduce the possibility of endogeneity problem, the regression model looking at new construction presented in this paper takes an approach similar to that taken by a subsequent research conducted by Mayer and Somerville in 2000. My regression model is derived from their empirical model of housing supply. In their analysis, new housing construction is a function of lagged changes in house prices and costs. Since housing price may be simultaneously determined with new construction, using a leading measure of new construction and lagging measure of house price in the analysis helps to eliminate endogenous issues.  A large body of literature on the topics of impacts of financial policies on housing price using cointegration approaches such as two-step residual-based technique (Engle and Granger, 1987), 	   8	  system-based reduced rank regression (Johansen, 1991), or autoregressive distributed lag model (Pesaran, Shin, and Smith, 2001; Qi and Cao, 2006). Qi and Cao (2006) examine the causality and causal relationship between property prices and bank lending in China during the period of 1999Q1 to 2006Q2. The percentage of literatures in this category suggests that government may control housing prices through tightening of credits.  The relationship between real estate policies and housing market differs across regions. Gerlach and Peng (2003) elaborate the two-way causality between housing prices and bank lending in Hong Kong by using a multivariate cointegration framework. Their empirical research provides evidence that bank lending adjusted to property prices does not have a significant influence on property prices; however, property prices do affect bank-lending policies. In particular, they elucidate that the sensitivity of credit to property prices decline as banks tightened credit standards. In a similar research by Collyns and Senhadji (2002), however, the empirical results indicate a credit shock has a significant impact on property prices for over 6 quarters in certain Asian countries. Unlike Gerlach and Pang (2003), Collyns and Senhadji’s research proves that the causality between credit lending and real estate prices are mutually dependent. Vargas-Silva (2007) also investigates the relationship between monetary policy and housing market in the US by using a VAR analysis. The empirical results indicate a negative impact of tightened monetary policy shocks on housing starts and residential investment. These contradicting results are indeed very interesting. The contradiction could be due to the dissimilarities in economic and geographical fundamentals across the market regions. On this basis, the impact of foreign 	   9	  policies in its local soil should only be taken with a grain of salt during the real estate policy-making process in China.  In the Chinese real estate industry, the government acts as a moderator by imposing policies and restrictions on financing and taxation to control the housing market. Many literatures point that the underlying economic factors are important determinants in explaining housing price movements.  Bhattacharya and Kim (2010) conducted a panel study for the period of 1990 to 2009 and concluded that economic fundamentals, such as employment, real construction cost and user costs significantly impact real housing prices in the US.  While certain theoretical and empirical literature argued that house price can reflect economic fundamentals in the long run (Malpezzi 1999, Capozza et al. 2002, Meen 2002), both national level data and panel MSA data in the US do not support the evidence of codependency between economic fundamentals and housing prices (Galin, 2002). This should not be interpreted as economic fundamentals having no hold over house prices; more specifically, it means that the level of house prices does not have a stable long run relationship with the level of fundamentals. Subsequently, Yu (2010) also shows that there is no stable relationship between housing price and economic fundamentals in the Chinese context. By using China’s panel data of 35 cities from 1998 to 2007, Yu provides evidence supporting the fact that real estate policies are an important factor in explaining the changes of China’s housing price. In light of this, the regression work in my paper uses economic fundamentals such as regional GDP per capita, 	   10	  investment in real estate and interest rate, together with various policy variables to further explore housing market movements.  It is a common consensus that policies implemented in earlier periods (from the 1990s to 2000s) did not achieve their desired results. Ye and Wu (2008) review the Chinese urban housing policies from 1998 to 2007 and provides an intensified analysis of urban housing policies from 2004 to 2007, pointing to the importance of continuous monitoring and revising of policies in Chinese housing market. Cao and Keivani (2013) examine the housing policies in China from 1998 to 2011 and argue that these policies have not only failed to achieve a satisfactory outcome, they have contributed to the declination of housing affordability.  Wei et al. (2013) elaborates on the topic of China’s credit control on real estate industry from an institutional perspective. Their empirical work builds on a series of interviews from properties practitioners on the effectiveness of monetary control measures during the period of August 2010 to August 2011. In particular, Wei et al. (2013)’s evaluation of China’s credit control on the supply side indicates that credit control policies are implemented in isolation, providing opportunities for developers to earn extraordinary profits. While Wei et al. (2013) does not provide empirical statistics analysis to prove that the market response was less than expected; their research provides general views from the supplier side for the credit control before 2011. While considerable literatures have been focused on the impact of monetary policies on the housing market in China, the macroeconomic control policies introduced in 2010 have been largely neglected. 	   11	  Literatures that discuss current policies tend to focus either on the supply side or the demand side; in contrary, my paper not only analyzes the impact of Chinese policy control from both supply and demand side, it also provides an updated impact of macroeconomic control policies on China’s housing market in recent five years. Unlike most journals published in China, which offers a discussion of China’s policies and markets trends, this paper uses regression models to provide a more detailed analysis of the impacts of the policies in specific cities and across market segments. Moreover, while many Chinese journals rely on national data in their statistical analysis of current trends, the addition of city data used in this paper will provide a more accurate assessment.    	   12	  3. Background of Chinese Real Estate Market This paper’s focus is on the Chinese urban housing market, and when referring to the real estate market in China, the intention is to the urban rather than rural sector. This chapter is structured as follows. The Chinese government uses four main methods to control the real estate market: financial policy, tax policy, land policy, and additional regulations, which are summarized in Section 1. Section 2 describes the general process of real estate development in China.   Section 3 provides a brief history of the five-stages of China’s housing policy reform in chronological order for future references.   3.1 The Evolution Of Current Chinese Real Estate Policies     This section reviews the evolution of specific Chinese government policies designed to regulate demand for residential real estate. A list of the financial, tax, land, and demand policies and regulations implemented at the national and municipal level from 2007 to 2013 are listed in Appendix 2.  3.1.1. FINANCIAL POLICIES  The central bank in China, the People’s Bank of China, changes lending rates to effect demand for housing, and savings rates and the required reserve ratios to control currency circulation, all of which affect real estate investment. Figure 3.1 illustrates the change in the required minimum down payment for home purchase since 1998 and Figure 3.2 tracks the changes of benchmark 	   13	  loan and benchmark rates, which are set by the PBOC. Figure 3.1: Required Minimum Down Payment 	   Banks in China adjust mortgage rates based on the benchmark interest rate. Minimum mortgage rates are set by the PBOC as a percentage of the benchmark rate. From 1998 to 2002, the mortgage rate dropped from 10.53% to 5.76%. From 2003 to 2005, the mortgage rate was raised eight times, increasing from 5.76% to 7.83%. In response to financial crisis in 2008, the PBOC reduced the long-term benchmark rate five times from 7.83% to 5.94%, and first home purchasers could attain a mortgage loan at 70% of the benchmark rate. The central bank raised the mortgage interest rate to 1.1 times of benchmark rate in January 2010. In October 2010, the People’s Bank of China increased the one-year benchmark interest rate from 2.25% to 2.5% and 0%	  10%	  20%	  30%	  40%	  50%	  60%	  70%	  80%	  %	  of	  price	  First	  Home	  Second	  Home	  	   14	  raised the one-year mortgage benchmark rate from 5.31% to 5.56%.  Figure 3.2: Long Term Benchmark Interest Rate from 2006 to 2012   (Soucre: Data generated from the Monetary Policy Department of the People's Bank of China, and Ministry of Housing and Urban-Rural Development of the People's Republic of China (MOUHURD) When commercial financing rates increase the supply of capital to developers is reduced, or developer are required to provide more equity cost of development increases.  Other national policies that increase the cost of development include policies such as the 2004 requirement that real estate companies hold a 35% capital ratio for new projects. These polices to restrict development or make it more costly contradicted the government’s original aim to lower housing prices. The net impact on prices depends on the relative size of the cooling of demand (inward shift in the demand curve) and the increased costs imposed on developers (inward shift in the 2006.08.19	  2007.03.18	  2007.05.19	  2007.07.21	  2007.08.22	  2007.09.15	  2007.12.21	  2008.09.16	  2008.10.09	  2008.10.30	  2008.11.27	  2008.12.23	  2010.10.20	  2010.12.26	  2011.02.09	  2011.04.06	  2011.07.07	  2012.06.08	  2012.07.06	  Loan	  Rate	   0.45	  	   0.27	  	   0.09	  	   0.18	  	   0.18	  	   0.27	  	   0.00	  	   (0.09)	   (0.27)	   (0.27)	   (1.08)	   (0.18)	   0.20	  	   0.26	  	   0.20	  	   0.20	  	   0.25	  	   (0.25)	   (0.25)	  Deposite	  Rate	   0.54	  	   0.27	  	   0.54	  	   0.27	  	   0.27	  	   0.27	  	   0.09	  	   0.00	  	   (0.27)	   (0.45)	   (1.26)	   (0.27)	   0.60	  	   0.35	  	   0.45	  	   0.25	  	   0.25	  	   (0.40)	   (0.35)	  (1.50)	  (1.00)	  (0.50)	  0.00	  	  0.50	  	  1.00	  	  %	  Change	  	   15	  supply curve) as shown in Figure 3.3. Figure 3.3: Impact of Restrictions on Demand and Supply   While	  demand-­‐side	  interventions	  try	  to	  restrict	  quantity	  demanded	  through	  increasing	  minimum	  down	  payments	  and	  imposing	  quotas	  on	  new	  homes	  purchased,	  supply-­‐side	  interventions	  focus	  on	  developers,	  by	  affecting	  the	  supply	  of	  land	  and	  capital	  through	  bank	  lending	  constraints,	  credit	  control,	  and	  changes	  in	  land	  transaction	  systems.	  The	  regression	  in	  later	  sections	  focuses	  on	  the	  demand-­‐side	  restriction.	  Identification	  of	  the	  demand	  side	  restrictions	  includes	  the	  unobservable	  supply-­‐side	  effects.	  However,	  the	  simultaneous	  application	  of	  supply-­‐side	  restrictions	  that	  are	  not	  controlled	  in	  the	  analysis	  suggests	  that	  we	  cannot	  distinguish	  between	  the	  supply	  and	  demand-­‐side	  mechanism	  in	  driving	  market	  quantity	  outcomes. When the central bank increases the required reserve ratio, the money multiplier becomes 	   16	  smaller, resulting in less circulation in currency. And higher interest rates, which should slow activity in the real estate market and stabilize house prices. Figure 3.4 shows the changes in the required reserved ratio from 2004 to 2012. The required reserve ratio increased substantially over this period, only being reduced at the end of 2008 and in early 2012, in response to concerns about excessive weakness in the economy.       Figure 3.4: Change of Required Reserved Ratio From 2004 to 2012  (Source: People’s Bank of China)  0.00%	  5.00%	  10.00%	  15.00%	  20.00%	  25.00%	  Mar-­‐04	  Jul-­‐04	  Nov-­‐04	  Mar-­‐05	  Jul-­‐05	  Nov-­‐05	  Mar-­‐06	  Jul-­‐06	  Nov-­‐06	  Mar-­‐07	  Jul-­‐07	  Nov-­‐07	  Mar-­‐08	  Jul-­‐08	  Nov-­‐08	  Mar-­‐09	  Jul-­‐09	  Nov-­‐09	  Mar-­‐10	  Jul-­‐10	  Nov-­‐10	  Mar-­‐11	  Jul-­‐11	  Nov-­‐11	  Mar-­‐12	  	   17	  3.1.2. HOUSING FINANCE OPTIONS  • Home mortgage loans from commercial banks Individual buyers apply for housing loans from commercial banks for the purchase of housing. Two types of housing loans are currently issued by commercial banks: Personal New Housing Loan, which refers to “commercial RMB loan issued to individual buyers for first home purchase with bank credit funds as resource... customers can enjoy preferential home loan interest rate and one-time preferential payment”(Bank of China, 2014) and Personal Second-hand Housing Loan, which is issued by commercial banks for the purchase of the second-hand housing.  All commercial banks are required to follow the regulations set by the PBOC. Each bank will adjust the minimum down payment, mortgage interest rate and maximum term of loan accordingly. Table 3.1 presents the down payment requirements and mortgage interest rate for major banks in China in 2014. Table 3.2 tracks the lower bound of interest rates relative to the benchmark loan rate required by the PBOC. In July 2013, the People’s Bank of China removed the floor on lending rates, which was 70 percent of benchmark lending rate. Henceforth, commercial banks could independently determine the interest rate on a market basis. As can be seen from Table 3.1, all commercial banks are strictly enforcing the minimum down payment rules for the first and second homes set by central bank and the State Council, and reducing preferential treatment of loan interest rate on the first home. Only six banks offer discounted interest rates and most banks in China set the loan interest rate same or higher than the benchmark interest rate.  	   18	   Table 3.1: Information of Minimum Down Payment and Interest Rate for Major Banks in 2014 	    Table 3.2: The lowest Bound of Interest Rate Relative to the Benchmark Rate    2006 2007 2008 2009 2010 2011 2012 2013 2014 First Home 0.9 0.9 0.9 0.9 0.9 0.9 0.8 0.7 N/A Second Home 1.1 1.1 0.7 0.7 1.1 1.1 1.1 1.1 1.1  • Personal Housing Provident Fund  Since 1991, the Personal Housing Provident Fund (HPF) provides employees with an alternative source for mortgage loans to borrowing from commercial banks. Under the HPF scheme, all employers and employees are required to contribute a portion of their monthly salaries to the 	   19	  fund. This contribution is based on locality and changes over time. In April 2013, Beijing Housing Provident Fund Management Center increased the minimum percentage from 5% to 8%, and set the maximum percentage contribution to 12%. Currently, the minimum percentage contribution in Shanghai is 7%. Individuals can apply to use HPF monies for the purchase, construction or overhauling of housing. Local administrators of housing provident fund follow regulations set by the Ministry of Finance and the People’s Bank of China, and determine loan limits, terms and down payment requirements.  Unlike a mortgage loan from commercial banks that set a maximum percentage an individual can borrow based on property price, the HPF loan sets its upper limit in a monetary amount. The maximum amount of an HPF loan that an individual can apply for is location specific and depends on the individual’s contribution to date to the HPF, credit rating, the term of loan, minimum down payment and the floor space of the home purchased. The maximum HPF loan an individual can borrow is ¥1,040,000 in Beijing,¥600,000 in Shanghai, and ¥900,000 in Shenzhen. In order to help medium to low income families to purchase housing, the HPF loan sets different minimum down payment requirements for properties smaller than 90 square meters, properties greater than 90 square meters, and the second property purchased, favoring the smaller or first property. Since 2010 because of macroeconomic controls on the real estate industry, HPF loans have been marginally better than the bank loans for home purchases. Appendix 1 illustrates the minimum down payment requirements for nine cities in China.  These requirements suggest 	   20	  that the local housing provident fund management centers have gradually increasing minimum down payment.  • Personal housing mixed loan A second type of loan is the Personal Housing Mixed Loan, which is a combination of commercial bank and the HPF loans. For many households the HPF loan is not large enough to enable them to purchase a home.  The Personal Housing Mixed Loan allows an individual to borrow from commercial banks to get the remaining funding.   3.1.3. TAX POLICIES In China, there are four main types of real estate taxes: sales, deed transfer, land appreciation, and property taxes. Sales tax applies to all real estate transactions. Recently, a higher tax rate has been implemented for previously owned properties to reduce speculation in the housing market. A different tax rate is set for deed transfers to impact the demand for different types of housing. To discourage land rights holders from transferring land use rights for profit, the government charges a high land appreciation tax. Property owners are required to pay an annual property tax, which is higher for larger properties. Table 3.3 illustrates major real estate tax changes since 2001.  	   21	  Table 3.3: Major Real Estate Tax Changes Policy Initiation Time  Major Changes  January 2001  Sales Tax: Decrease from 5% to 3%    Property Tax: Decrease from 12% to 10%    Income Tax: Decrease from 20% to 10%    Tax for Rent Profit: Decrease from 7.3% to 14.57%  May 2005 Sales Tax: be exempted within two years  October 2005  Income Tax: no more exemption for transactions in the secondary market  May 2006 Sales Tax: be exempted within five years June 2006 Income Tax: Full taxation on homes less than five years July 2006 Income Tax: 20% of additional home transfer February 2007 Deed Tax: 3%~6% November 2008 Deed Tax: 1%    Land Appreciation Tax: exemption October 2009 Sales Tax: No exemption  February 2013  Income Tax: 20% of all transactions in the secondary housing market  (Source: State Administration of Taxation) 3.1.4. LAND POLICIES The land policy in China has four main objectives: land supply, allocation, transaction and development. The Chinese government may stabilize housing prices by controlling the amount of land available for development at a given time. When the government increases the land supply, developers will pay less for land. If the housing demand remains the same, developers will likely lower prices in a competitive market. The government currently allocates more land to low-priced housing developments, including low-rent buildings for residences with low and medium income. By allocating sufficient land for this purpose, the government intends to increase affordability for low and medium income 	   22	  households. The “9070” policy announced in May 2006 required 70% of newly constructed homes to be smaller than 90 m2. Land allocation methods have changed over time to ensure fairness. While all land use rights were previously allocated by the government, most land use rights are auctioned today. All parties could only acquire land through auctions in the open market after August 31, 2004. To eliminate issues surrounding land hoarding, the government requires developers to pay a penalty if the land is undeveloped for one year. If the land remains undeveloped for two years or more, the land use rights will be deemed invalid.  3.1.5. ADDITIONAL RESTRICTIONS AND REGULATIONS Currently in most cities households are restricted to owning no more than one property at any given time. Moreover, financing may be unavailable, or mortgage rates may be much higher, for buyers of additional homes. The government stopped encouraging foreign investment in the real estate industry in 2007. Since April 2011, the government has implemented internal regulations for the proper enforcement of housing policies by government officials.   The central government currently uses regulations and laws to set a general direction of the real estate market and land development for China. These are carried out by the local governments in their respective provinces. The provincial government forwards the messages from the central government to the municipal and lower level governments. There will be a time lag between the 	   23	  implementation of national policies and local policies, as it is up to local governments’ discretion to customize these policies and determine the timeline based on local economic conditions. Appendix 2 illustrates the difference in time between the announcement of national policies and its actual implementation in the selected cities.  Since taxation for different sectors are allocated to different levels of government, conflicts of interest may arise in the implementation of certain regulations and laws. For instance, tax revenue from housing and land-related sectors is allocated to the local government, and tax revenue from construction-related industries is allocated to the central government. The impact of selling land use rights on the municipal government is substantial as the transactions account for 25% to 50% of the cities’ revenue (Ding and Knaap, 2005). As a result, local governments are primarily interested in increasing housing and land transactions, while the central government is primarily focused on slowing down the increase in housing prices to stabilize the housing market.  While private companies pay only taxes, government-owned companies contribute both taxes and a large portion of profits to the government. This means that the government benefits from the growth of government-owned companies. In 2009, it was reported that 8 of the 10 highest-priced lands were owned by state-owned or municipal-owned companies. In 2009, a major natural resources company in China, Sinochem, purchased usage rights to a piece of land approximately 26 km2 for more than 4 billion RMB. The cost of usage rights to this property was only approximately 1.8 billion RMB in 2004. The sales revenue of state-owned real estate 	   24	  companies was 220.9 billion RMB in 2009, accounting for 5% of real estate sales nationwide (National Bureau of Statistics of China, 2010). In auctions involving government-owned companies, the cost of land use rights tends to reach staggering new heights. Since only a handful of large companies, most of which are government-owned, are able to afford the high land usage costs. In 2010, the State Council issued a new regulation requiring non-real estate associated government-owned companies to stay out of the real estate market in order to promote a healthier market. However, many government-owned companies are still active in the market.   3.2. Real Estate development in China This section provides a brief introduction of real estate development process in China. Due to the development of the real estate related policies over time, the property development process in China has undergone great changes. The process of land development generally follows the phases outlined below:      3.2.1. DECISION-MAKING PHASE      In the initial phase, land developers generally focus on feasibility studies on technical, financial, and economic factors to determine whether they should move forward with a project. The law in China requires developers to acquire an official evaluation report from government-assigned agencies before construction. Acquiring the official evaluation report is costless but usually takes 3 months to process. Developers will be ranked accordingly, where higher ranked companies 	   25	  will face fewer difficulties while seeking for alternative financing from banks and other sources.  3.2.2. LAND USE RIGHTS ACQUISITION PHASE      After determining that the project is worthwhile, land developers need to acquire land use rights. Unlike in many countries, a fee-simple interest in land cannot be purchased in China as the state retains ownership.  Instead, there is a right to use the land for a defined period, analogous to a land lease. The government at the municipal and county levels is in charge of assigning land use rights. Land users should pay the total amount of the land rights assignment fee within 60 days of signing the contract to acquire the Certificate for the Right of Land Usage. Table 3.4 below shows land use right time limits.  Table 3.4: Land Use Right Time Limits By Use  Land Use Purposes  Time Limit  Residential  70 years  Industrial  50 years  Education, science, culture, public health and physical education  50 years  Comprehensive utilization or other purposes  50 years  Commercial, tourist and recreational purposes  40 years  According to the regulations governing land usage issued by the State Council on May 19, 1990, assignment of land use rights can only be carried out by reaching an agreement with another land use rights holder, by acquiring land use rights from the local government, or by winning a land 	   26	  use rights public auction. They can also bid for land use rights in redevelopment areas in China or partner with a company holding land use rights. In some cities, the local government may offer developers land use rights at a lower than average cost in exchange for the construction of public infrastructure. Additionally, the local government may reassign land use to attract major companies in exchange for higher tax revenue.  The process of land use rights assignment has undergone drastic changes since 2001. The State Council issued a notice in 2001 to strengthen the management of state-owned land, which, for the first time, clearly defined the scope and limitations of land acquisition and usage. Before the implementation of this “831 Deadline,” the state government owned all the land and had the right to sell or allocate land use rights to third parties. Effective as of August 31, 2004, developers wishing to acquire land use rights must participate in a public auction, which means that developers can no longer get land solely from negotiating with the government.    Before the public auctions of land use rights, the government was the only active player in assessment and assignment of land use. This meant that the government could be flexible with land lease terms, which could lead to bribery. The introduction of public auctions helped to improve fairness in land transactions and intensified competition among land developers. Data released by the National Bureau of Statistics of China shows that housing transactions in Tianjin in the first quarter of 2004 increased 14.2% compared to the first quarter in 2003, and the land transaction price increased 18.1% compared to the first quarter in 2003 (National Bureau of Statistics of China, 2004).  The following chart shows a steady increasing trend of the land 	   27	  transaction price from 2000 to 2012.  Figure 3.5: Average Land Transaction Price from 2000 to 2012 	    (Source: China Statistical Yearbook, 2013)  Although the implementation of this policy was to increase fairness in land distribution, local governments were able to manipulate regulations by setting specific bidding qualifications in favor of certain companies. For instance, a local government might require all bidding companies to have a certain amount of capital; thus, companies with deeper pockets and closer relationship with local governments have more opportunities, while companies with less capital, which cannot afford the high land use costs, are less likely to stay competitive and are eventually forced out of the market. This phenomenon happens in most cities and has become an unspoken business culture.  434	   444	   461	   576	  647	   759	  1,043	   1,211	  1,524	  1,888	  2,503	   2,600	  3,393	  1,948	   2,017	   2,092	  2,197	  2,608	  2,937	  3,119	  3,645	   3,576	  4,459	  4,725	  4,993	  5,430	  0	  1,000	  2,000	  3,000	  4,000	  5,000	  6,000	  2000	   2001	   2002	   2003	   2004	   2005	   2006	   2007	   2008	   2009	   2010	   2011	   2012	  Land Transaction Price (Yuan/Square Meter)  Land	  Transaction	  Price	   Average	  Housing	  Prices	  	   28	  3.2.3. DEMOLITION PHASE  Public housing constructed after 1949 are not up to the standard of 70 years of residential usage based on its materials and technologies used at the time. Local government then sell the land use to developers and established regulations to address this issue. Demolition is a necessary part of the land development process, which can legally be done by the local government, land developers, or companies specializing in housing demolition. After the Administrative Department of Demolition receives the application for housing demolition and related administrative fees from the developers, a Permit of Demolition will be issued. Unlike other countries, the developer must shoulder the cost of relocating existing site occupants.  Consequently, the demolition process may be the most important source of variation in a developer’s costs, with important effects on completion time as well.    Usually, developers offer three common types of compensation: monetary, an in-kind housing settlement, and a relocation settlement.  In practice, developers find it difficult to reach compensation and resettlement agreements with unit owners of a building in the demolition process. With the increasing number of development projects in China urban housing demolition and relocation disputes have become more common.  “Nail households,” which refers to unit owners who refuse to vacate a property undergoing demolition or those who repeatedly ask for higher compensation, have been regularly reported in recent years. Conflicts arise when these owners refuse to accept the compensation, and in some situations, developers may use illegal methods to intimidate the tenants to vacate the properties, 	   29	  sometimes resulting in casualty or property damage during the forced demolition process. A prolonged demolition process increases the costs and risks of the development project as construction times are postponed. Moreover, disputes can tarnish the involved parties’ reputations and negatively affect the future sales of the properties. Thus, establishing an efficient relocation compensation method is necessary and important in any land development project.  3.2.4. LAND DEVELOPMENT  A common type of land development in China is cooperative land development where one party shares land use rights with another party in exchange for capital, technology and labor. These two parties become partners in the real estate project. The parties involved in cooperative land development can be the local governments, who have the land use rights, and private developers, who have the capital and expertise. For instance, the Westlake Boulevard, the new central business district in Hangzhou, was developed cooperatively by the local government and Intime Retail Group Co Ltd.. Located in a central part of the city, the local government provided the land use rights and organized the demolition process, and Intime Retail Group was responsible for construction and management of the approximately 100,000m2 commercial retail space in the area (Uptown District of Hangzhou, 2013).    The most popular land development model in recent years is to use a holding company to control a series of smaller designated real estate project companies that each focus on the development of a specific project. In the case of insolvency, the subsidiary project companies could declare bankruptcy without affecting the parent company. However, in 2010, to cool down the housing 	   30	  market, the Central Bank of China, the People’s Bank of China (PBOC), forbade public banks from lending to small to medium real estate companies, forcing subsidiary companies to turn to private borrowers and trust funds, which usually required the mother company to be the guarantor of the debt. This phenomenon led to a large number of real estate companies going bankrupt in 2013. This situation accentuated the advantages of larger companies and resulted in the loss of smaller companies, potentially resulting in a less competitive marketplace.  3.2.5. CONSTRUCTION AND SALE PHASES      As in other countries, developers in China need to obtain various permits, such as land use and construction permits.  Once they receive a construction permit for the project, they can begin designing the structures and establish a supervision and construction team.  The developers have two options to sell the properties: pre- and post-completion sale (pre-sale and post-sale). In recent years, the pre-sale of residential housing has become the primary real estate transaction. For pre-sale of properties, developers must acquire a pre-sale permit from the local government, which includes requirements such as the capital expenditure must have exceeded 25% of the total planned investment for the project. Some cities also have specific requirements as to what percentage of the construction must be completed prior to receiving the approval to pre-sell. After the project has been completed, buyers will register their ownership with the real estate board and legal ownership is transferred from the developers to the buyers. 	   31	  3.3. Background of Housing Policy in China In recent years, real estate macro-control policies in China have become an important subject of public concern. Over the last three decades, housing reform in China has led to the commercialization of the housing market. This section provides an introduction to these reforms, looking specifically at policies aimed at addressing issues pertaining to land acquisition and usage, credit and financing, and housing as a social benefit.  The development of China’s real estate market can be divided into four stages. During the first two stages, government reforms transformed real estate in China from a social benefit solely distributed by the State to a commodity good that can be traded in the open market. During the later two stages of the reform, properties have changed from being consumption goods to increasingly being investment goods. 3.3.1. PRE-REFORM STAGE: 1949-1977 The first stage took place from 1949 to 1977, where most urban residents relied on the government or public institutions for housing.  With the founding of the People’s Republic of China in 1949, the urban housing system was built on the country’s planned economy.  The construction, distribution, management, and pricing of urban housing were based on a socialist system. During this period, central government controlled housing investments, while local governments and government-owned institutions provided public housing, where the allocation of housing became intertwined with employment. Low rent and unlimited tenure are 	   32	  the principal features of this system for households, so that housing was supplied by the State as a kind of social welfare or employment-based benefit.  The system led to insufficient supply, poor built quality, and an unequal distribution of housing became major problems due to low investment, poor management, and poorly enforced regulations during the Cultural Revolution (1966-1976) (Wang & Murie 1999, p.67). Since artificially low government set rents could not cover the development and maintenance costs, the government had to invest more funds into the housing system, resulting in a heavy financial burden. The shortage of housing in major cities under this sytem is estimated to have reached 1 billion square meters by 1978 (Wang and Murie, 1999, p.101). The Third National City Works Conference in March 1978, “Strengthening Urban Construction Work” was launched to address different aspects of urban housing development. At the same time, commodity housing and land property rights began to be widely discussed within government and among scholars, leading to the second stage of the housing reform. 3.3.2. FIRST STAGE: 1978-1987 To address the existing problems in the public housing system, the Chinese central government began to explore new housing policies. This led to the second stage of the housing reform, which took place from 1978 to 1987, and consisted of three waves of housing reform experiments conducted in select cities in China.  	   33	  a). First Housing Reform Experiment: 1978 to 1982 The first housing reform experiment took place from 1978 to 1982 in Xian and Nanning, The government allowed households to purchase housing, setting the price equal to construction costs, without mark-up. However, high price per unit, rigid payment methods, and low rents for public housing discouraged residents from purchasing houses, resulting in insufficient demand for housing during the first period of the housing reform experiment.  b). Second Housing Reform Experiment: 1982 to 1985 The second experiment took place from 1982 to 1985. Under this regime, home purchasers were required to pay only one third of the sales price; local governments and employers of the purchasers were responsible for the rest of the cost. The policy was initiated in 1982 in 4 cities, Zhengzhou, Changzhou, Siping, and Shashi. The new policy effectively increased the housing demand. 83,200 m2 of housing were sold in 1983 (Xue, 1985, p.161).  In the following years, housing investment increased substantially. Between 1981 and 1985 the central government tripled its housing investment relative to the previous five year period, to 100 billion RMB to construct 648 million m2 of urban housing, compared with approximately 30 billion RMB to construct 267 million m2 of urban housing from 1976 to 1980 (Wang & Murie, 1999, p.104). By the end of 1985, the experimental policy was implemented in 160 cities and 300 towns. Although home ownership was now possible for individuals, owners had no right to sell their 	   34	  properties. They were only allowed to transfer home ownership to within their family. Housing reform policies during the second period transformed urban housing from a “free good” to a “subsidy good” (Chiu, 1996), but built the foundation for the future transition to a “commodity” housing market. Even though the national average living floor space per capita increased from 3.6 m2 in 1976 to 5.2 m2 in 1985 (Ministry of Construction, 1997), existing problems, such as the lack of construction standards and unfair distribution of housing to employees, remained. c). Third Housing Reform Experiment: 1986 to 1988 The third housing reform experiment took place from 1986 to 1988. The Housing Reform Steering Group at the central government level and the Housing Reform offices at the local government level were set up in 1986. The main policy change in this period was to increase rents to cover maintenance costs and introduce housing subsidies, to promote the sales of state-owned public housing to households. This new policy was first implemented in Yantai, and critically created a system for housing finance. Home were required to pay a 30% down payment, but could then pay off the outstanding balance over 15 years.  The three housing reform experiments that took place during the second housing reform period resulted in increased per capita living space in urban areas. According to China Statistic Yearbook, per capita living space in urban areas nearly doubled and floor space owned by individuals in urban areas increased from 33% in 1985 to 43% 1987. Even so, the majority of the housing stock remained in state sector, with most housing in China still owned by work units.  	   35	  3.3.3. SECOND STAGE: 1988 TO 1997 The third stage took place from 1988 to 1997, where a comprehensive housing reform was implemented nationwide. In January 1988, the first Nationwide Housing Reform Conference was held, leading to the subsequent release a month later by the State Council of the document, “Implementation Plan for a Gradual Housing System Reform in Cities and Towns.”. The new policy aimed to commercialize housing and create an operating housing market.  It urged the change in the low rent system and encouraged the circulation of funds in the housing market. As part of this process Shanghai introduced the Housing Provident Fund, in 1991 which provides employees with subsidized mortgages, and added housing coupons and home purchase discounts Details about the Housing Provident Fund will be discussed in a later section. In October 1991, the second national housing reform conference issued “Comments on the Comprehensive Reform of the Urban Housing System”, which updated the 1988 implementation plan and provided solid guidelines for all municipalities to carry out the national housing reform. This document not only suggests increasing rent, but also allowing properties to be traded in the market, as well as developing real estate financing channels.  Although the sales of public housing increased from 1988 to 1993, residents continued to purchase properties from their employers instead of from other individuals in the market. Therefore, the true real estate market was not completely formed until a new policy document, “Decisions on Intensifying the Urban Housing Reform”, was issued in 1994. The new policy aimed to change the old housing investment, management, and distribution system, and to 	   36	  develop a dual housing provision system, a public and private housing saving system, a new housing financing system, and a healthy and regulated market system. In addition, the first real estate law, “Urban Housing and Real Estate Management Act of the People’s Republic of China”, was established in In July 1994, to regulate the housing market. Before the 1990s, the main channel of real estate development financing was bank lending. The financing channels for funds for developers started to diversify in 1992, where companies could choose to raise funds through equity expansion by IPOs. In 1992 in Hainan province China experienced its first housing development boom.  Over 10,000 real estate companies were registered to develop commercial housing, overwhelmingly speculative development.  In response, in June 1993 the government implemented strict rules to prevent unqualified real estate companies from going public. From late 1993 to 1997, the government aimed to set the housing policies to better manage the housing market and to solve existing problems, such as an exponential rise in housing prices. During this nationwide comprehensive housing development period, the government strove to expand the commercial housing market. As a result of policies to encourage supply, commercial housing investment increased from 27% of total urban housing investment in 1991 to 56% in 1994 (Wang & Murie, 1999, p.174). However, the dynamics of the market were not that of well balanced matching of households with developers as most commercial housing was still purchased by companies and employers, with the intention of attracting employees. During this period, high commercial housing prices limited opportunities for home ownership. 	   37	  To address this problem, the government initiated a housing project for low-income households (The Anju Project) to support middle and low-income families by offering housing at low prices and with favourable tax treatment (Lau and Lee, 2001). Between 1995 and 1997, approximately 650,000 households benefited from the Anju project.  As shown in Figure 3.6, the building space per capita in urban areas gradually improved from 6.3 m2 in 1988 to 8.8 m2 in 1997. By June 1998, over 50% of the total housing stock was owned by individuals, compare to 43% in 1987. Figure 3.6 and Figure 3.7 compare the living condition between urban areas and rural areas. Even though urban living condition improved significantly during the period, rural residents continued to have more space per capita. Figure 3.6: Floor Space of Newly Built Residential Buildings in Urban and Rural Areas from 1978 to 1997       1978	  1980	  1985	  1986	  1987	  1988	  1989	  1990	  1991	  1992	  1993	  1994	  1995	  1996	  1997	  Urban	  Areas	   0.38	  0.92	  1.88	  1.93	  1.93	  2.03	  1.56	  1.73	  1.93	   2.4	   3.07	  3.57	  3.75	  3.94	  4.05	  Rural	  Areas	   1	   5	   7.22	  9.84	  8.84	  8.45	  6.76	  6.91	  7.54	  6.19	  4.81	  6.18	  6.99	  8.28	  8.06	  0	  2	  4	  6	  8	  10	  12	  Floor	  Space	  of	  Newly-­‐built	  	  Residential	  Buildings	  	  (100	  000	  000	  sq.m)	  Floor	  Space	  of	  Newly-­‐built	  Residential	  Buildings	  	  in	  Urban	  and	  Rural	  Areas	  from	  1978	  to	  1997	  	   38	  Figure 3.7: Per Capita Living Spaces in Urban and Rural Areas from 1978 to 1997   Source: China Statistical Yearbook, 1998. 3.3.4. THIRD STAGE: 1998 TO 2002 The fourth stage of the national housing reform took place from 1998 to 2002. Following the Asian financial crisis of 1997, the real estate industry entered into a period of contraction. Housing prices dropped rapidly, and sales for commercial housing slowed. The major stumbling block of development of real estate market was the interference of employers. Rongji Zhu, the former Premier of China, ruled that employers should no longer allocate subsidized housing after 1998. In April 1998, the People’s Bank of China announced a series of policies to eliminate the previous restrictions on mortgage loans to encourage individual property purchases, which led to a rapid growth in real estate market.  The government aimed to transform the real estate 1978	  1980	  1985	  1986	  1987	  1988	  1989	  1990	  1991	  1992	  1993	  1994	  1995	  1996	  1997	  Urban	  Areas	   3.6	   3.9	   5.2	   6	   6.1	   6.3	   6.6	   6.7	   6.9	   7.1	   7.5	   7.8	   8.1	   8.5	   8.8	  Rural	  Areas	   8.1	   9.4	   14.7	  15.3	   16	   16.6	  17.2	  17.8	  18.5	  18.9	  20.7	  20.2	   21	   21.7	  22.4	  0	  5	  10	  15	  20	  25	  Per	  Capita	  Living	  Spaces	  	  (sq.m)	  Per	  Capita	  Living	  Spaces	  in	  Urban	  and	  Rural	  Areas	  	  from	  1978	  to	  1997	  	   39	  industry into China’s economic pillar using these policies. This was matched with the abolition in late 1998 of the welfare housing allocation by employers. The subsidy to workers was subsequently monetized through the Housing Provident Fund, which could be used for the purchase of a property. The document, “Deeping the Urban Housing Reform and Housing Construction”, was released by the State Council in July 1998, making government distribution of housing a history. It was an important turning point in China’s housing reform history, representing the beginning of the commercialization of the real estate market.  As shown in the Figure 3.8 and Figure 3.9 below, real estate development and investments in residential urban areas have been increasing steadily since the implementation of the new policy in 1998. In addition, completed commercial buildings, which include all buildings that can be purchased, almost doubled from 1998 to 2002. Figure 3.8: Investment in Real Estate from 1998 to 2002  1998	   1999	   2000	   2001	   2002	  Residential	  Buildings	   6393.8	   7058.8	   7594.1	   8339.1	   9407.1	  Urban	  Areas	   4310.8	   5050.9	   5435.3	   6261.5	   7248.9	  Real	  Estate	  Development	   2081.6	   2638.5	   3312	   4216.7	   5227.8	  0	  1000	  2000	  3000	  4000	  5000	  6000	  7000	  8000	  9000	  10000	  Total	  Investment	  	  (Million	  Yuan)	  Investment	  in	  Real	  Estate	  from	  1998	  to	  2002	  	   40	  Figure 3.9: Floor Space of Building Completed and Sold from 1998 to 2002   (Source: China Statistical Yearbook, 2013) The set of new policies successfully stimulated demand for market (commodity) housing demand. Housing finance became substantially more favourable, from 1998 to 2002, as residents were allowed to pay a 20% down paymentment and were allowed mortgage terms up to 20 years. During this period, the annual mortgage rate dropped from 10.53% to 5.76%. The price for apartments measuring approximately150 m2 ranged from 2,000 to 9,000 RMB per m2 depending on the property location, floor level, and other amenities. Majority of the demand came from the lower end of the market, where prices were more affordable (Hook, 2003). By the end of 2002, individual property purchases represented 94.3% of the total real estate market activity, and per capita floor space of residential buildings in urban areas increased to 24.5 m2 (China Statistical Yearbook, 2013). 1998	   1999	   2000	   2001	   2002	  Floor	  Space	  of	  Building	  Completed	   70166.1	   79646.1	   80507.9	   85278.9	   93018.3	  Residential	  Buildings	  Completed	   47616.9	   55868.9	   54859.9	   57476.5	   59793.6	  Commercialized	  Housing	  completed	  	   17566.6	   21410.8	   25104.9	   29867.4	   34975.8	  Commercialized	  Housing	  Sold	  	   12185.3	   14556.5	   18637.1	   22411.9	   26808.3	  0	  10000	  20000	  30000	  40000	  50000	  60000	  70000	  80000	  90000	  100000	  Floor	  Space	  	  (10,000	  square	  meters)	  Floor	  Space	  of	  Building	  Completed	  and	  Sold	  from	  1998	  to	  2002	  	  	  	   41	  3.3.5. FOURTH STAGE: 2003 TO PRESENT Housing reform in the last decade focused on balancing the housing market. In many cities, such as Beijing, Shanghai, and Hangzhou, housing prices have been increasing rapidly since 2003. Average residential housing prices increased by 137.15% and luxury housing prices increased by 174% from 2003 to 2012 (China Statistic Bureau, 2014). To cool the overheated real estate market, the government implemented a series of policies to adjust and regulate the market; this period is commonly referred to as the intensified macro-regulation period.  The empirical analysis that follows studies the effects of government policies to control the market in this period,  In June 2003, the People’s Bank of China announced the No.121 Document to impose tighter debt control measures for real estate companies and stricter loan qualifications for individual property purchasers Many developers faced the risk of capital chain fracture due to the establishment of the new policy. As shown in Figure 3.10, the growth rate of investment in real estate development soared from 27.3% in 2002 to 30.4% in 2004. In 2005, the growth rate dropped to 22.9%, which was the lowest since 1998.        	   42	  Figure 3.10: Growth Rate of Investment in Real Estate Development from 1999 to 2011  (Source: China Statistical Yearbook, 2013) The new housing policies were aimed at stabilizing the housing market and continuing to provide affordable housing for middle to low-income households. Policies were not always consistent: while the People’s Bank of China imposed new restrictions on the financing process, the central government loosened certain housing restraining policies to encourage overall real estate development. In August 2003, the State Council released a new document “Notice on the Promotion of Sustained and Healthy Development of the Real Estate Market”, requiring banks to continue providing financial support to real estate developers who met specific financing requirements. These contrasting approaches revealed contradictory attitudes towards the development of the real estate market during this period. Despite the implementation of tightened monetary policies since 2003 and frequent adjustments to these policies, real estate prices grew dramatically. The annual growth rates of residential housing prices and high-end housing prices were 18.7% and 35.4% respectively in 2004 China Statistical Yearbook, 2013). In 2004, the central government retracted its decision to cool the housing market. To adjust the 26.8%	  25.5%	  27.3%	  24.0%	  29.6%	  30.4%	  22.9%	  25.6%	  32.0%	  24.6%	  14.1%	  32.8%	  30.3%	  1999	   2000	   2001	   2002	   2003	   2004	   2005	   2006	   2007	   2008	   2009	   2010	   2011	  Growth of Investment in Real Estate Development 	   43	  real estate market on both the supply and demand sides, China’s housing policies targeted four main areas: land acquisition, debt financing, market entrance prerequisites, and market demand. First, the famous “831 Deadline” was an important turning point in China’s land reform process. The policy stated that after August 31, 2004 companies could only acquire land through open market auctions. Prior to this, state entities, which owned all the land and had the right to sell or allocate land to third parties, could do so in whatever manner they deemed appropriate. The impact of this land policy reform on real estate development will be discussed in later section. Second, banks imposed strict debt financing requirements on real estate companies, requiring them to hold a 35% capital ratio for new projects. Third, in 2007, the government stopped encouraging foreign investment in the real estate industry, including real estate development, management, and purchase in the primary and secondary markets. Fourth, to keep housing prices from increasing uncontrollably, the government utilized credit and tax policies to increase the equity requirements for purchases in a manner to discourage property speculators: the minimum down payment for first homes below 90 m2 was raised to 20%, and that for first homes over 90 m2 was raised to 30%. The minimum down payment for second homes was raised to 40%. These macro-control policies differentiated treatment for small to medium size housing from that for larger units and investment properties.  The intention was to give priority to ensuring the supply of small size and lower-priced commodity housing. In March 2005, banks eliminated preferential mortgage policies to lower housing demand. During this period, the mortgage rate was raised eight times, increasing from 5.76% to 7.83%. In addition, the State Council imposed 	   44	  the first “Eight National Rules” to lower speculation and lower house price appreciation. For instance, sales tax for residential properties purchased less than 2 years ago needed to be fully charged, and pre-sale properties could not be resold before completion of the building. Finally, the “9070” policy announced May 2006 required 70% of newly constructed homes to be smaller than 90 m2. The macro-control housing policies did not appear to effectively calm the real estate market. The gap between average family income and average housing price remained huge. The house price –to- income ratio increased from 6.6 in 2003 to 8 in 2007. The high price to income ratio, especially for medium and low-income households, has been an important concern for government since 2007 (Yang & Zan, 2008). As a result of the 2008 financial crisis the real estate market in China retrenched placing great pressure on the Chinese government. Housing market experienced its first price drop since 2000. As shown in Figure 5, the growth rate of investment in real estate development dropped dramatically in 2008. In response, the Chinese government and commercial banks started to soften their home purchasing and financing policies to encourage home purchases. In December 2008, the State Council issued a new document “Promoting the Healthy Development of the Real Estate Market”, which introduced credit and tax preferential policies to encourage the purchasing of owner-occupied housing.  Encouraged by these benefits, more real estate speculators began purchasing multiple properties for investment In 2009, major cities, commercial and residential housing sales increased 86.6% 	   45	  comparing to 2008. (Sina News, 2009). By the third quarter of 2009, the real estate market in China was booming again, leading to rapidly increasing housing prices. Forbes described China’s real estate market as one of the seven looming financial bubbles in 2009 (Forbes, 2009). To stabilize housing prices, the central government in December 2009 again began to tighten real estate financing policies.  In January 2010, the central government implemented the “Eleven National Rules” for real estate market regulation. In April 2010, the central government replaced the “Eleven National Rules” with the “Ten National Rules”. The main difference between the two sets of rules was a new restriction on the first home purchased. For homes larger than 90 m2, first home purchasers were required to make a down payment of 30%. In addition, second home purchasers were required to make a 50% down payment. In some cities, the fast-growth in housing prices and lack of supply led the local government and commercial banks to impose a 100% full payment for additional home purchases. Since October 2010, the People’s Bank of China increased the one-year benchmark interest rate from 2.25% to 2.5% and raised the one-year mortgage benchmark rate from 5.31% to 5.56%. As shown in Figure3.11, the growth in the average selling price of residential buildings reached the bottom in 2008, with a price drop of 1.9%, and then suddenly increased by 24.9% in the following year. Similarly, high-end residential properties followed the same trend during this period; however, the price adjustment of high-end homes was less sensitive than other types of commercial residential buildings from 2007 to 2011. In the long run, price change of high-end 	   46	  housing was more volatile than other types of housing. Figure 3.11:Annual Growth in Average Selling Price of Residential Housing  (Source: China Statistical Yearbook, 2013) In January 2011, the central government introduced the “Eight National Rules”, which further increased the mandatory down payment percentage for the second home to 60%. In addition, the mortgage interest rate was set at 1.1 times higher than the benchmark mortgage rate and the Housing Provident Fund lending rate increased 25 basis points. More rigorous house purchase restrictions were also imposed on non-local residents.   In February 2013, the central government implemented the “Five National Rules”. On top of the previously introduced policies, down payment requirement for the second mortgage increased from 60% to 70% and a 20% personal income tax would be charged on all transactions in the secondary housing market.  1999	   2000	   2001	   2002	   2003	   2004	   2005	   2006	   2007	   2008	   2009	   2010	   2011	   2012	  Residential	  Buildings	   0.16%	   4.90%	   3.54%	   3.72%	   5.02%	   18.71%	  12.61%	   6.21%	   16.86%	  -­‐1.90%	  24.69%	   5.97%	   5.68%	   8.75%	  High	  End	  Housing	   -­‐2.02%	   -­‐4.77%	   1.40%	   -­‐4.46%	   -­‐0.22%	  34.52%	   4.63%	   12.87%	  13.46%	   4.41%	   23.86%	  13.16%	   0.55%	   4.24%	  -­‐10.00%	  -­‐5.00%	  0.00%	  5.00%	  10.00%	  15.00%	  20.00%	  25.00%	  30.00%	  35.00%	  40.00%	  Yuan/Square	  Meter	  	   47	  While financing policies were tightened to lower the housing demand and stabilize housing prices, housing prices continued to increase. Housing prices managed to keep a slightly increase even after the initial launch of the tightened policies in 2011 and then shot up again in early 2013. In January 2013, National Statistics reported that 53 of the 70 major cities in China experienced a growth in housing prices. For instance, housing prices in Shenzhen still increased by 2.2% from January 2012 to January 2013. Beijing, Guangzhou, and Shanghai also experienced a monthly growth rate of 2.1%, 2.0%, and 1.3%, respectively. It is noted that the annual economy growth rate at the time was 11%, which indicates that the growth of real estate industry was seized. Although the macro-control policy has undergone numerous adjustments, housing prices in China have continued to increase steadily, contrary to public expectations. Appendix 1 provides an overview of major real estate policies described above and implemented from 1978 to 2013.      	   48	  4. Housing Market Data  4.1 Data  The monthly macroeconomic and transaction data used here are from the Chinese Real Estate Index System (CREIS). As the major source of data in China, the CREIS records government statistics that are published by the central, provincial, and local governments on a weekly or monthly basis. To smooth the data, monthly data are aggregated to a quarterly frequency here. The data is a panel of nine cities - Beijing, Shanghai, Shenzhen, Wuhan, Dalian, Shenyang, Xi’an, Dongguan, and Nanchang, over the period from 2007Q3 to 2014Q1, with the exception of Shanghai, which is from 2006Q1. Since the data sets collected from the CREIS is incomplete, additional macroeconomic data are compiled directly from the National and Municipal Bureau of Statistics in China to supplement the analysis.  The cities are classified into three tiers by population size, economic growth, and political importance. Table 4.1 shows the city attributes to provide an overview of the city tiers. In the table, residents refer to individuals staying in the city for six months or longer, and citizens refer to individuals registered under the municipality. Cities differ in several aspects within and across tier levels.     	   49	  Table 4.1: City Facts  Tier one cities include Beijing, Shanghai, which are direct-controlled municipalities, and have the political status of provinces, and Shenzhen, which is a special economic zone. The real estate markets in these three cities are the most mature in China. Beijing and Shanghai are under the Central government’s direct administration. Located in northern China, Beijing is China’s political, cultural and educational center. Shanghai and Shenzhen, on the other hand, are the financial and economic pillars, located in the eastern and southern coasts of China, respectively. Compared to Beijing and Shanghai, where citizens account for more than half of the population, most people in Shenzhen are migrants. As seen in Table 4.1, the citizens of Shenzhen make up less than one third of the city’s population.      	   50	  Figure 4.1: Location of Nine Selected Cities   Tier two cities include Wuhan, Dalian, Shenyang, and Xi’an, which are sub-provincial cities. Wuhan, Shenyang, and Xi’an are all capital cities of their respective provinces, and Dalian is a coastal city in the same province as Shenyang. Tier three cities include Dongguan and Nanchang, which are prefecture-level cities. Nanchang is the capital city of Jiangxi Province, which is located in the interior part of China, while Dongguan is a coastal city in Guangdong Province, where local citizens only account for one fourth of the city’s residents. Figure 4.1 shows the location of the nine cities included in the data analysis. 	   51	  The variables used in this empirical analysis fall into the following categories: quarterly aggregated transaction data, macro economic data, and regulatory measures. 4.1.1 HOUSING PRICE AND SALES VOLUME Quality regional housing transaction data is not readily available in China. The data presented in this paper is information based on average selling price, and aggregate sales volume and total area of properties sold but categorized by size from the first quarter of 2007 to fourth quarter of 2013 for the selected nine cities. Although lending policies preferential treatment for properties smaller than 90 m2 and restraining treatment for properties greater than 140 m2, data in this paper are breakdown into three groups (≤ 90 m2, 90 – 140 m2, >140m2), due to the assumption that purchasing behavior of medium units are significantly different from very large units. Thus, for each city there are three observations per quarter, which differ across housing market variables but will share the same city and national macroeconomic values.  The data from CREIS are transaction records of new homes, which do not include information on resale properties. Since information on resale properties is available only for Beijing and Shanghai, this information was not considered in this analysis. Data categorized by property size allows for the comparison of different market segments. Since financing policies in China are different for the types of home purchased, the data allows for an analysis of the effects of the policies to differ between large and small properties. The panel data of nine Chinese cities offers the opportunity to examine the effectiveness of the policies in specific cities and critically to 	   52	  determine whether policies have differential policies across housing types and potentially re-direct rather than restricting demand for housing. 4.1.2 MACROECONOMIC DATA The macroeconomic data used here are mortgage rates for loans over 5 years, real estate investment, GDP per capita, and fixed investment. Real estate investment is the RMB amount of investments made by real estate development companies in structure and exclude expenditures for land transactions. City GDP per capita is calculated using the number of residents rather than citizens as a better determining factor that delineate each city’s activities. Fixed investment is investment in construction projects totaling 5 million Yuan or more, which includes both real estate transactions but also infrastructure and industrial facilities. Other macroeconomic data, which include unemployment rates, disposable income, CPI, and domestic loan amounts, are collected from the Statistical Bureau of China. Together with the real user cost of housing, which is calculated by subtracting the home price appreciation over the last quarter from the 5-year mortgage interest rate, these variables are used in the regression model.  4.1.3 REGULATORY VARIABLES To measure government efforts to restrict housing demand I use three policy variables that were introduced at different times in different cities.  The three are minimum down payment requirements, limits on the number of homes that may be purchased, and the minimum lending rate.  Depending on the policy instrument, it may apply to funds borrowed from the Housing 	   53	  Provident Fund or may only apply to bank loans These bundle of policies were examined for its effect on house price appreciation, unit sales of new housing, and total floor area sold of new housing. The POLICY dummy indicates whether restrictive regulations are prevalent in the selected nine cities. The regulatory measure used in this empirical analysis is a policy dummy variable that marks the time period that the restrictive policies are implemented as 1 and the time period during which the policies were absent as 0. The first restrictive financing policy from 2007 to 2013 was announced by the national government in January 2010 (No.4 [2010] of the State Council) and later imposed by the municipal governments in various time periods. Table 4.2 summarizes the time periods during which POLICY=1 for each city.  Table	  4.2: Summary of the POLICY Variable City POLICY=1 Beijing 2010q2  -  2013q4 Shanghai 2010q4  -  2013q3 Shenzhen 2010q2  -  2013q2 Dalian 2010q3  -  2013q1 Wuhan 2011q1  -  2013q0 Shenyang 2010q2  -  2013q1 Xi'an 2010q4  -  2013q2 Nanchang 2010q3  -  2013q3 Dongguan 2012q3  -  2013q4  	   54	  The other individual regulation variables, ℎ𝑝𝑓_1™? ,ℎ𝑝𝑓_1™? ,𝑏𝑎𝑛𝑘_2™? ,: the level of minimum down payment when using the Housing Provident Fund to purchase the first property, the level of minimum down payment when using the Housing Provident Fund to purchase the second property, and the level of minimum down payment when using a bank loan to borrow for a second property, respectively.  Dummy variable of the second home purchase restriction, 𝑝𝑢𝑟𝑐ℎ𝑎𝑠𝑒™? , and gap of mortgage rate, ΔR, which is the difference between discount or premium relative to the benchmark loan rate of the first home and that of the second home, are used in the regression work to examine different implications of each policy.   4.2 Descriptive Statistics  Table 4.3 gives descriptive statistics for the variables used in the empirical analysis. The geographically aggregated panel data pools data on different cities of diffident tiers across eastern China, which provides an overview of the real estate market across the entire region. Among the nine selected cities, nominal housing prices increase by an average of 12.28% per quarter over the period. As shown in Figure 4.2, a majority of housing price change data collected lie above the baseline and the distribution of data becomes more volatile after 2010Q1, even though the State Council announced the No. 4 document with the intention to establish an intensified microeconomic control and stabilize the market.  The data is consistent with the public’s doubts about the overwhelming real estate market and sparks the initial research question of whether macroeconomic control policies during that period effectively controlled the 	   55	  market.  Quarterly sales volume also varies significantly during the period. The change of sales volume ranges from -13,336 units to 13,777 units, and the average increase of approximately 81 units of properties sold account for 1.28% of the mean sales volume. Figure 4.3 shows the distribution of changes of quarterly sales volume.  The pattern indicates a certain seasonality trend. In general, the change of sales volume is less significant in the first quarter and it gradually grows until year-end. This is due to the Chinese Spring Festival, which is a seven-day national holiday. This festival is based on the Chinese lunar calendar and falls either in January or February. Business activities are practically stagnant during this period.  The average change in total floor space of properties sold is 1.08% of the mean level, while the average change of floor space of new constructions is 2.46% of the mean level. The data is consistent with the fact that construction rates are faster than its selling counterpart.  At the national level floor space sold and new construction completed have tracked closely, but new construction started has risen dramatically at the national level from 83 million m2 in 2008 to 131 million m2 in 2012., pointing to an explanation for the current high and rising vacancy rate in the real estate market.    	   56	  Table 4.3: Descriptive Statistics of the Panel Data  Variable Mean Standard Deviation Minimum Maximum Citizens (0,000)      736.10 393.13 171.26 1426.93 Residences (0,000) 1071.81 552.90 458.06 2415.15 Unemployment Rate (%) 3.07 0.97 1.21 4.50 Regional GDP (00,000,000 ¥) 1937.15 1414.54 280.13 6127.99 Regional GDP per capita  (0,000 ¥) 1.69 0.66 0.41 4.16 Fixed Investment  (00,000,000 ¥)     Level 1666.55 2375.82 73.91 13561.17 Changes 160.98 2165.55 -12438.41 5230.73 Real Estate Investment (00,000,000 ¥)     Level 188.71 140.60 10.60 663.62 Changes 9.74 119.22 -380.49 481.62 Domestic loan (00,000,000 ¥)     Level 106.21 127.11 0.22 832.31 Changes 4.30 93.77 -621.46 736.27 Long Term Benchmark Loan  Interest Rate (%) 0.07 0.01 0.06 0.08 Disposable Income (¥) 6648.87 2366.00 980.00 16455.07 New Construction (0,000 m2)     Level 345.30 233.47 8.46 1352.03 Changes 8.48 241.06 -647.74 1033.09 Average Price (¥/m2)     Level 11778.26 7939.95 3135.50 46944.00 Changes 256.57 2063.67 -14873.00 24244.70 % Price change -0.5345 1.4959 0.1228 0.2124 Number of Homes Sold (unit)     Level 6289.93 5692.90 63.00 32545.00 Changes 80.79 2859.49 -13336.00 13777 Floor Space of Residential Sold  (0,000 m2)     Level 62.07 53.82 1.40 349.27 Changes 0.67 28.74 -145.93 155.58 	   57	  Figure 4.2: Changes in Nominal Housing Prices  Figure 4.3: Changes in Quarterly Sales Volume  	   58	  The percentage price change during the period has been presented in Table 4.4. The mean price change dropped in year 2008, mainly due to the negative impact of financial crisis on housing market. A slowdown in price growth rate can be observed after 2010. In 2012, mean price change experienced a negative change of 3.7%. Table 4.4: Descriptive Statistics of Year on Year % Price Change from 2006 to 2013 Year Minimum Maximum Mean Standard Deviation 2007 -0.0574 0.2728 0.1102 0.0906 2008 -0.4193 0.7379 0.0949 0.2599 2009 -0.5345 1.4959 0.1114 0.2726 2010 -0.3198 1.2747 0.2941 0.2138 2011 -0.2723 0.5452 0.1284 0.1459 2012 -0.2910 0.3192 -0.0037 0.1080 2013 -0.1536 0.8160 0.0992 0.1502 Total -0.5345 1.4959 0.1228 0.2124  Geographically aggregated data leads to biased estimates and shrouds the underlying relationships and behavior across cities (Goodman, 1998). For example, as shown in Figure 4.2, regional GDP per capita ranges from ¥4115.52 to ¥ 41555.48, and fixed investment ranges from ¥7,391 million to ¥1,356.117 billion. The interesting variation across cities, as derived from Table 4.5, cannot be observed from Table 4.3. Therefore, city level data is preferable to geographically aggregated data in this research.  Table 4.5 gives descriptive statistics for the key variables used in the empirical analysis by city, showing different variation across the local real estate markets. Part II indicates that on 	   59	  average all cities experienced price inflation during the period, and the average year on year percentage increase of housing price ranges from 12.31% in Shenyang to 16.06% in Beijing The average price increases for Tier 1 cities (Beijing, Shanghai, and Shenzhen) are above the mean price change. Tier 2 and Tier 3 cities have mean price increase below the mean price change, with Dalian as an exception, reflecting their lower price levels. The nominal average price follows the same pattern. The average quarterly housing price in Beijing, Shanghai, Shenzhen, and Dalian are above the mean price level of the panel. Similarly, standard deviations of average price reflect large variation in Tier 1 cities. With a standard deviation of 0.3608, the average price change in Shenzhen is the most volatile, with year on year price changes declines as large as -41.63% and quarterly price increase as high as 149.59%. Although price inflation in Tier 1 cities was severe, increase in the number of homes sold for Tier 1 cities were all below the average. On the other hand, cities with price inflation less than average (Part II of Table 9) experienced sales volume changes that are higher than average (Part III of Table 9). More specifically, mean changes of sales volume in Beijing and Dalian were negative, with 140 units and 1 unit below average, respectively. It gives a sign that purchasing activity in Tier 1 cities may slow down and Tier 2 and Tier 3 cities are catching up, which leads to the second empirical question of whether the housing control policies have differential impact across city tiers.  Part I and Part IV of Table 4.5 present a summary statistics of floor space constructed and sold in each city. Newly constructed residential housing in Beijing increased by 16,250 m2 per quarter; however, average floor space of residential sold decreases by 6344 m2, indicating that 	   60	  the residential housing may have been overdeveloped in Beijing. When the financing environment for homebuyers worsens, reduction in demand may result in an overall decrease of real estate development in local markets. This is inconsistent with the data, mainly because of the developers’ positive expectation of the real estate market. According to interviews conducted by Wei et al., developers believe that the policies are implemented in isolation. When comparing the mean value of residential sold and built, all selected cities have more floor space built than sold. Therefore, it is worthwhile to examine whether policies have effectively reduced the real estate development.  	   	  	   61	  Table	  4.5: Descriptive Statistics of Key Variables by City Part 1. Variable: New Construction (0,000 m2) Part I City Minimum Maximum Mean Standard Deviation Beijing     Level 118.79 944.52 431.04 216.94 Changes -615.89 515.94 16.25 329.19 Shanghai     Level 321.45 745.90 469.38 110.22 Changes -320.18 265.80 1.30 128.84 Shenzhen     Level 33.07 292.56 130.94 58.76 Changes -136.08 172.32 5.23 61.54 Dalian     Level 93.86 627.51 291.75 122.12 Changes -418.30 372.11 4.15 193.66 Wuhan     Level 87.35 735.09 382.18 175.49 Changes -647.74 437.63 12.08 258.91 Xian     Level 85.98 716.02 385.12 183.61 Changes -315.77 385.20 9.62 188.42 Shenyang     Level 162.90 1352.03 609.41 324.35 Changes -568.40 1033.09 11.21 465.99 Nanchang     Level 24.59 336.63 135.51 75.40 Changes -250.85 260.63 7.50 105.33 Dongguan     Level 8.46 195.69 114.60 63.61 Changes -38.40 117.17 12.81 45.78 Total     Level 8.46 1352.03 345.30 233.47 Changes -647.74 1033.09 8.48 241.06  	   	  	   62	  Table4.5: Descriptive Statistics of Key Variables by City (Continue) Part 2. Variable: Percentage Year on Year Changes of Average Housing Price (¥/m2) Part II City Minimum Maximum Mean Standard Deviation Beijing -0.4193 1.2064 0.1606 0.2853 Shanghai -0.1528 0.6429 0.1488 0.1640 Shenzhen -0.4163 1.4959 0.1292 0.3608 Dalian -0.2199 0.4162 0.1232 0.1551 Wuhan -0.2955 0.4369 0.0621 0.1438 Xian -0.1203 0.7024 0.1453 0.1975 Shenyang -0.1176 0.5800 0.1231 0.1368 Nanchang -0.5345 0.6601 0.1468 0.2035 Dongguan -0.4151 0.3722 0.0614 0.1474 Total -0.5345 1.4959 0.1228 0.2124 	   	  	   63	  Table4.5: Descriptive Statistics of Key Variables by City (Continue) Part 3. Variable: Number of Homes Sold (unit) Part III City Minimum Maximum Mean Standard Deviation Beijing     Level 673 22909 6853.5 4907.9 Changes -13336 11992 -59.4 3404.4 Shanghai     Level 2036 28610 8798.1 5312.5 Changes -12643 13777 79.9 4035.0 Shenzhen     Level 63 13839 2852.1 3323.8 Changes -6223 4857 74.2 1538.5 Dalian     Level 224 9349 2713.7 2184.3 Changes -4771 4566 -0.5 1290.4 Wuhan     Level 1158 23463 9322.4 6383.1 Changes -6656 6775 188.9 2376.2 Xian     Level 127 20745 6079.0 5206.5 Changes -6852 7062 93.7 2381.5 Shenyang     Level 819 32545 11103.3 7978.9 Changes -11947 11919 19.6 4841.8 Nanchang     Level 368 5941 2623.6 1552.9 Changes -1928 1998 79.2 984.3 Dongguan     Level 509 13867 4848.4 3078.7 Changes -5508 5199 262.3 1702.0 Total     Level 63 32545 6289.9 5692.9 Changes -13336.00 13777 80.79 2859.49       	   	  	   64	  Table4.5: Descriptive Statistics of Key Variables by City (Continue) Part 4. Variable: Floor Space of Residential Sold (0,000 m2) Part IV City Minimum Maximum Mean Standard Deviation Beijing     Level 17.21 124.86 45.30 24.66 Changes -58.89 44.91 -0.63 23.70 Shanghai     Level 31.22 349.27 106.31 64.64 Changes -145.93 155.58 0.98 51.19 Shenzhen     Level 1.40 111.97 25.59 24.92 Changes -53.98 44.95 0.47 14.45 Dalian     Level 4.66 72.00 25.47 15.16 Changes -46.19 31.01 0.01 11.68 Wuhan     Level 18.56 251.30 95.09 61.11 Changes -79.45 108.32 1.14 27.83 Xian     Level 2.85 231.28 63.58 53.80 Changes -76.50 82.33 0.57 25.41 Shenyang     Level 14.89 236.94 106.53 56.73 Changes -90.29 91.17 0.48 41.20 Nanchang     Level 9.28 69.44 28.47 12.82 Changes -24.32 24.38 0.73 10.41 Dongguan     Level 8.99 69.69 45.93 17.27 Changes -19.46 31.70 2.31 13.65 Total     Level 1.40 349.27 62.07 53.82 Changes -145.930 155.580 0.669 28.739  The fourth and key empirical question for this thesis is whether the restrictive policies impact different size groups differently. Is demand diverted between groups because of the 	   65	  policies varying by housing unit size will be discussed. The quarterly housing transaction data are grouped into three size categories: ≤ 90 m2 (Group 1), 90 – 140 m2 (Group 2), >140m2 (Group 3) to observe the market trends across the different size categories.  As observed from Table 4.6, average price increase for homes smaller than 90 m2 is below the market average, while bigger units, homes greater than 140 m2 has an average price increase of 12.87%. The standard deviation of the percentage price change for three groups are similar. The mid-size units, group 2, have a higher standard deviation of 0.2153.   Part II and Part III of Table 4.6, show that the patterns in the number of homes and residential floor space sold are different from price changes across the groups. Group 1 has the highest number of units and floor space sold, whereas Group 3 has the least amount of sales volume and floor space sold. Percentage changes in sales volume across groups fluctuate substantially, with group 3 having the greatest in percentage increase in sales, with a standard deviation of 1.09. Mean percentage change of sales volume and total floor space sold for Group 2 and Group 3 are below average percentage change of the market. Thus, whether the government’s implementation of restrictive policies has succeeded in cooling down the high-end market and redirecting housing demand into smaller homes will be explored in later section.  Government policies and objectives may have been at cross-purposes, something that is tested below.  For instance, although it was the government’s intention to promote the market for smaller sized homes by introducing the second home purchase restriction, this action may have given individuals an incentive to buy bigger homes if they were only going to be able to 	   66	  buy one home.. Thus this begs the question of whether the implementation of second home purchase restriction redirects the housing demand to bigger homes, a result that is the opposite of the government’s stated policy intention. Table	  4.6: Descriptive Statistics of Key Variables by Size Category Part I. Average Price (¥/m2) Size Category Minimum Maximum Mean Standard Deviation <=90 m2     Level 3135.50 20844.90 9289.85 4808.85 Changes -4051.20 4413.00 204.73 1164.33 % Changes -0.4193 1.2064 0.1170 0.2125 90-140 m2     Level 3235.83 23930.50 9728.35 5312.00 Changes -6602.13 8071.50 205.14 1284.49 % Changes -0.5345 1.4959 0.1227 0.2153 >140 m2     Level 4183.67 46944.00 16297.44 10348.91 Changes -14873.00 24244.70 359.38 3125.36 % Changes -0.4151 0.8160 0.1287 0.2103 Total     Level 3135.50 46944.00 11778.26 7939.95 Changes -14873.00 24244.70 256.57 2063.67 % Changes -0.5345 1.4959 0.1228 0.2124                	   67	  Table4.6: Descriptive Statistics of Key Variables by Size Category (Continue) Part II. Number of Homes Sold (unit) Size Category Minimum Maximum Mean Standard Deviation <=90 m2     Level 924 32545 8689.41 5416.89 Changes -13336 11992 159.76 3461.31 % Changes -0.7744 4.4145 0.2634 0.7248 90-140 m2     Level 367 28610 8232.15 5967.96 Changes -12643 13777 68.29 3311.18 % Changes -0.8974 3.1892 0.2343 0.7371 >140 m2     Level 63 12692 1966.56 2059.84 Changes -5508 6628 14.60 1294.51 % Changes -0.9618 9.7635 0.3183 1.0948 Total     Level 63 32545 6289.93 5692.90 Changes -13336 13777 80.79 2859.49 % Changes -0.9618 9.7635 0.2721 0.8690                	   68	   Table4.6: Descriptive Statistics of Key Variables by Size Category (Continue) Part III. Floor Space of Residential Sold (0,000 m2) Size Category Minimum Maximum Mean Standard Deviation <=90 m2     Level 8.99 236.94 62.45 40.56 Changes -90.29 86.66 1.16 24.82 % Changes -0.8260 3.5106 0.2718 0.7630 90-140 m2     Level 4.26 349.27 83.44 66.75 Changes -145.93 154.11 0.50 35.68 % Changes -0.9015 3.6243 0.2477 0.8114 >140 m2     Level 1.40 246.04 40.40 41.09 Changes -107.80 155.58 0.35 24.42 % Changes -0.9004 4.2105 0.2870 0.8878 Total     Level 1.40 349.27 62.07 53.82 Changes -145.93 155.58 0.67 28.74 % Changes -0.9015 4.2105 0.2689 0.8212  	  	    	   69	  5. Empirical Analysis The empirical investigation begins by applying the empirical models of the housing market that have been built by previous literature (Arnott,1987; Schwartz, 1988; Smith, Rosen, and Fallis, 1988; Megblolugbe, Marks, and Schwartz(1991); Dua, Miller, and Smyth, 1995). Theories suggest that factors including home prices, the mortgage interest rates, and some other economic activities may explain movements in the unit home sales. Mayer and Somerville (2000) provide an empirical model of residential construction, by regressing new housing starts on the current and lagged changes in log housing prices, the change in real interest rate, population, quarterly dummies, city time trend, and policy variables. The empirical methodology provided by the literature above has laid the foundation for the work that can be seen in this paper.  This paper uses a city-level panel of data to investigate the policy impact on the housing market. More specifically, the panel consists of 9 cities, and for each city, there are three market segments categorized by size for observation. Although	  restrictive	  regulations	  are	  implemented	  at	  different	  times	  and	  intensities	  across	  cities	   and	  unit	   size	   submarkets,	   regulations	  may	  not	  be	  uniformly	   enforced	   across	   cities.	  Tier	   1	   cities	   usually	   strictly	   enforce	   all	   the	   regulations	   because	   they	   are	   either	   directly	  controlled	  municipalities	  or	  capital	  cities	   in	  respective	  provinces,	  where	  housing	  markets	  are	   heavily	   regulated	   and	   more	   mature	   than	   other	   cities.	   Nevertheless,	   regulations	   are	  typically	  uniformly	  enforced	  across	  different	  submarkets	  within	  a	  city.	  Since	  the	  research	  focuses	  on	  investigating	  the	  differential	  impact	  across	  unit	  size	  submarket	  instead	  of	  policy	  	   70	  effects	   across	   cities,	   the	   issue	   of	   inconsistent	   regulation	   across	   cities	   would	   not	  fundamentally	  affect	  the	  empirical	  analysis. To gain initial insight into how change of house prices, unit sales and total floor space sold vary with implementation of various policies over the sample period, I employ the following panel fixed effect regressions (5.1) below, which allows us to see whether macroeconomic control policies during the period have effectively impact the housing market movements: Y™? = 𝛼 + β ∙ ln  (𝑝𝑟𝑖𝑐𝑒  𝑐ℎ𝑎𝑛𝑔𝑒𝑠) ™ ,???   + 𝜔 ∙ Δreal  estate  𝑖𝑛𝑣𝑒𝑠𝑡𝑚𝑒𝑛𝑡??? + 𝛿∙ Δ𝐺𝐷𝑃  𝑝𝑒𝑟  𝑐𝑎𝑝𝑖𝑡𝑎??? + 𝜋 ∙ Δ𝑖𝑛𝑡𝑒𝑟𝑒𝑠𝑡  𝑟𝑎𝑡𝑒??? + 𝜃 ∙ 𝑃𝑜𝑙𝑖𝑐𝑦™? + 𝜑∙ 𝑃𝑜𝑙𝑖𝑐𝑦™ ,??? + 𝛾 ∙ 𝑃𝑜𝑙𝑖𝑐𝑦™ ,???×  Σ𝑔𝑟𝑜𝑢𝑝? + 𝜌 ∙ 𝑡𝑖𝑚𝑒  𝑡𝑟𝑒𝑛𝑑? + 𝐺𝑟𝑜𝑢𝑝  𝐹𝐸+ 𝐶𝑖𝑡𝑦  𝐹𝐸 + 𝑄𝑢𝑎𝑟𝑡𝑒𝑟  𝐹𝐸 + 𝜖 ™? ;        5.1  where 𝑃™?  refers to policy variables, 𝑡𝑟𝑒𝑛𝑑? is anational time trend, and 𝜀™?   is the error term that is assumed to vary non-stochastically over time or group.  𝑌™?  is a vector of current and lagged value of dependent variables. This research focus on testing policy effect on: i) Changes in log prices, ii) Log unit sales , iii) Log total floor space sold.  The policies in question are as follows:  • Down payment level of purchasing the first home using the Housing Provident Fund: ℎ𝑝𝑓_1™?  • Down payment level of purchasing the second home using the Housing Provident Fund:  ℎ𝑝𝑓_2™?  	   71	  • Down payment level of purchasing the second home using the bank loan: 𝑏𝑎𝑛𝑘_2™?  • Third home purchase restriction dummy: 𝑝𝑢𝑟𝑐ℎ𝑎𝑠𝑒™?  The differential financial policies may affect change of housing price differently and the purchase restriction may give individuals more incentive to buy bigger homes in one time. The interaction terms between lag policy and group dummies is to test whether each restrictive policy succeeded in cooling down the high-end housing market and in redirecting housing demand into homes that are more accessible to marginal homebuyers. While dependent variable (1) to (3) is different across groups, log new construction area measures the aggregated construction activity in each city. All explanatory variables are lagged one quarter with respect to the dependent variable to avoid endogenous issue. The econometric estimation controls for group fixed effects 𝜑?, city fixed effect 𝛿?, and quarter fixed effect  𝜏, to absorb other possible unobserved determinants of the housing market. Moreover, 𝑖 indexes the selected cities, 𝑡 represents each quarter from 2006 to 2013, 𝑐 refers to market groups, and 𝑚 and 𝑛 represents the lag of the variable. In addition, the subscript 𝑞 represents the number of quarter that each policy effect is statistically from zero. Assuming movements of sub-markets are correlated with each other, the standard error is clustered by city-time such that observations within the cluster are correlated while observations between clusters are uncorrelated. City-time specifies to which group each observation belongs. In this situation, housing transactions within a particular city in a particular time period are correlated while any two transactions from different cities in different period are independent. It 	   72	  indicates that three unit size sub-markets are not independent distributed within a city at a specific time period. The way that the econometric model clustered affects the standard errors and variance-covariance matrix but not the estimated coefficients.  Using quarterly data from a panel of 27 market groups between 2006 and 2013, the econometric model tests the following hypothesis. Based on theory and existing literature, the research expects that demand-side housing restrictions have a negative correlation with quantity demanded and housing price changes. Furthermore, due to the preferential policies imposed to group 1 units, the research presumes that quantity demanded should be diverted from bigger units to smaller units.  Hypothesis 1: H0: 𝜃 = 0, the policy does not have significant relationship with the dependent variable  H1: 𝜃 ≠ 0, the policy is significant associated with the dependent variable Hypothesis 2:  H0:  𝛾 = 0,  the policy effect across groups does not have significant difference H1: 𝛾 ≠ 0, the policy diverted demand from one group to another Table 5.1 – 5.3 present OLS estimation of regression (5.1), which test the aggregated policy impact on housing market and diversion of quantity demanded across sub-markets. The null hypotheses that the policy does not have significant impact on the dependent variable are rejected when testing each dependent variable. The estimates show that housing markets are 	   73	  significantly correlated with the lagged value of policies. More specifically, changes of housing prices are positively correlated with investment in real estate, regional GDP per capita, the benchmark interest rate. The finding of positive relationship between housing price changes and investment or GDP per capita is consistent with the common perception and previous literatures. However, positive correlation between interest rate and housing price growth are in contrast with the expected negative relationship and previous theory (Campbell and Shiller, 1988a, b). The positive relationship suggests the rapid growth in housing prices in China. Changes of housing price exhibit a short run serial correlation and start to show mean reversion in lag 2, which is consistent with the previous theory (Case and Shiller,1989; Capozza et al. 2004), showing high volatility of housing price changes.  Table 5.1 presents the relationship between the right hand side variables and the changes in average housing prices. The regression estimates suggest that the down payment requirements when purchasing the 2nd home have a significant negative impact on the changes of housing prices, which is consistent with government’s expectation. The coefficient for ℎ𝑝𝑓_2™ ,??? is significant at 99% confident level, which means a 10-percentage point increase in the required minimum down payment is negatively associated with a 38% change in price increase rate, so if price appreciation is 10% per annum, this would result in lowering the growth rate by 6.2%.  The minimum down payment when purchasing the second home using a bank loan has greater impact on housing price changes. The coefficient of -0.0669 indicates that a 10-percentage point increase in the minimum down payment is negatively associated with a price change of 66.9%.  	   74	  The HPF loan policy on the second home down payment has a similar but less pronounced impact in each period when compared to the 2st home.    	   75	  Table 5.1: Housing demand, Dependent variable: log (price changes) Regression Type Regr. (a1)  Regr. (a2)  Regr. (a3)  Regr. (a4)  Tested Policy 1st home HPF Loan 2nd home HPF loan 2nd Home Bank Loan 3rd Home Purchase Restriction       Log price change 0.385*** 0.343*** 0.314*** 0.322***  (𝑡 − 1) (0.0844) (0.0802) (0.0783) (0.0798) Log price change -0.124* -0.0830 -0.0542 -0.0687  (𝑡 − 2) (0.0659) (0.0650) (0.0606) (0.0605) Log real estate investment change 0.232 0.342 0.323 0.245   (𝑡 − 1) (0.462) (0.444) (0.459) (0.441) Change in the loan interest rate 0.419 1.144 3.485** 3.024**   (𝑡 − 1) (1.432) (1.566) (1.518) (1.460) Change in Regional GDP per capita 0.167 0.134 0.0924 0.151    (𝑡 − 1) (0.189) (0.195) (0.212) (0.179) Down payment level 0.0488 0.0105 0.0120   (0.0325) (0.0200) (0.0144)  Down payment level -0.0475 0.00173 -0.0669***        (𝑡 − 1)  (0.0359) (0.0197) (0.0204)   Down payment level  -0.0384***       (𝑡 − 2)  (0.0133)   Lag (1) policy interact with group 1 0 0.00508 -0.000925 -0.0172     (t-q)*group1  (0) (0.0120) (0.0182) (0.323) Lag (1) policy interact with group 2 0.00522 0.00270 0.00161 0.0177     (t-q)*group2 (0.0231) (0.00740) (0.0131) (0.227) Lag (1) policy interact with group 3 0 0 0 0     (t-q)*group3 (0) (0) (0) (0) Purchase restriction    0.852*** (Dummy variable)    (0.253) Constant -2.305*** -1.572*** 1.443 -1.987***  (0.681) (0.590) (1.119) (0.329) Number of groups 27 27 27 27 Observations 271 271 276 276 R-squared 0.293 0.326 0.355 0.343  • 𝑞 = 1 in Regr. (a1), 𝑞 = 2 in Regr. (a2), 𝑞 = 1 in Regr. (a3), 𝑞 = 1 in Regr. (a4) • Robust standard errors in parentheses (*** p<0.01, ** p<0.05, * p<0.1). Data are quarterly from 2006 to 2013. All regressions include quarterly dummies, city dummies, group dummies, and a national time trend. Standard error for variables clustered by city_time.  	   76	  Contrary to the prediction, coefficients on the 3rd home purchase restriction in Table 5.1 are positive and statistically different from zero. The coefficient on purchase restriction shows that whenever a purchase restriction is announced in the city, the average housing price increases by 85.2% in the current period. This result seems improbably large, indicating that the increase demand may come from marginal buyers, who own one home and still be able to purchase a second home. Marginal homebuyers may advance their second home purchase with the anticipation of future more restrictive rules. Moreover, since the purchase dummy variable that only has the value of 1 for a small period in one city would have very little variation to generate identification, so that the coefficient would just reflect the difference for the city over that short period, which makes it susceptible to spurious correlation with some unobserved time varying factor. Other than insufficient variation in purchase restriction, the endogeneity of restraining policies with housing market movements can be another reason that contributes to the large coefficient on the purchase restriction dummy.  The policy interaction variables suggest some diversion of quantity demanded in that the coefficients are consistently positive, but none are statistically different from zero. The positive sign implies that increased restrictions in one area lead to faster price appreciation in other sub-markets, consistent with a diversion in demand from one sub-market to another.  Due to the time gap between the announcement of national policy and the implementation of local policy, rumors regarding the implementation of housing control policies in the city could impact the households’ purchasing behavior. In general, a tightened financing or purchasing 	   77	  policy rumored to be implemented in the future may result in an increase in transactions for the purpose of avoiding a higher financial burden. Above coefficients in the quarter of implementation of the policy variable are all positive, but not statistically different from zero. Coefficients in lagged quarter are statistically negative, indicating that the public is treading cautiously and taking a conservative stance before the implementation of the policy.  Table 5.2 tests individual the policy effect on log unit sales. The 1st home HPF down payment requirement are significant in lag 2 at 99% confidence level, and the 2nd home HPF down payment requirement are significant in lag 1 at 99% confidence level. A 10-percengae point increases in down payment requirements is negatively associated with sales volume by between 27% and 45%. The effect is transitory, affecting sales volume in the short run. As with prices, there seems to be some advanced acceleration of purchases as sales rises in the quarter of implementation, though as above the increase is not statistically different from zero.   The interaction variables also indicate some diversion of quantity demanded from bigger units to smaller units. For smaller units, about 40% of the overall reduction in quantity demanded from the increase in down payment requirements is offset by demand diverted from the large unit size sub-markets.    	   78	  Table 5.2: Housing demand, Dependent variable: log (unit sales) Regression Type Regr. (a5)  Regr. (a6)  Regr. (a7)  Regr. (a8)  Tested Policy 1st home HPF Loan 2nd home HPF loan 2nd home Bank Loan 3rd Home Purchase Restriction       Log price change -0.162 -0.219 -0.200 -0.167   (𝑡 − 1) (0.291) (0.255) (0.242) (0.243) Log real estate investment change 0.102 0.161 0.216 0.200   (𝑡 − 1) (0.224) (0.206) (0.206) (0.212) Change in the loan interest rate -2.116*** -0.987* -1.779*** -2.290***   (𝑡 − 1) (0.506) (0.548) (0.547) (0.516) Change in Regional GDP per capita 0.0786 0.0167 0.0492 0.0692    (𝑡 − 1) (0.0832) (0.0834) (0.0774) (0.0836) National time trend -0.00701* -0.0478*** -0.0408*** -0.0261***  (0.00391) (0.00839) (0.00896) (0.00780) Down payment level 0.0169* 0.000123 -0.00574   (0.00981) (0.00606) (0.00861)  Down payment level -0.0457*** -0.0278*** -0.0326***        (𝑡 − 1)  (0.0119) (0.00649) (0.00944)  Down payment level -0.0311**        (𝑡 − 2) (0.0156)    Lag (1) policy interact with group 1 0 0.0102** 0.0159** 0.0164     (t-q)*group1  (0) (0.00396) (0.00723) (0.226) Lag (1) policy interact with group 2 -0.00824 0.00203 0.00715 0.133     (t-q)*group2 (0.0151) (0.00343) (0.00567) (0.228) Lag (1) policy interact with group 3 0 0 0 0.229     (t-q)*group3 (0) (0) (0) (0.250) Purchase restriction    0.241** (Dummy variable)    (0.104) Purchase restriction    0.0154    (𝑡 − 1)      (0.120) Constant 9.572*** 9.511*** 10.58*** 9.121***  (0.303) (0.171) (0.481) (0.102) Number of groups 27 27 27 27 Observations 581 587 610 610 R-squared 0.771 0.782 0.783 0.778 • 𝑞 = 2 in Regr. (a5), 𝑞 = 2 in Regr. (a6), 𝑞 = 1 in Regr. (a7), 𝑞 = 1 in Regr. (a8)  • Robust standard errors in parentheses (*** p<0.01, ** p<0.05, * p<0.1). Data are quarterly from 2006 to 2013. All regressions include quarterly dummies, city dummies, group dummies, and a national time trend. Standard error for variables clustered by city_time.   	   79	  Table 5.3 presents the results when the dependent variable is total floor space sold, showing a similar pattern as unit sales. Total floor space sold allows seeing whether there is an effect on unit size. The coefficients suggest that there is no effect on first home size, but total floor area sold for the second homes falls by 31 to 37% for a 10-percentage point increase in the relevant minimum required down payment. More specifically, the estimates show that the restrictive policies is negatively associated with the total number of units demanded and the size of units demand.  As with units sold, the interaction terms suggest diversion of quantity demanded from larger to smaller units, as the interaction with group 1 is significantly positive. Similarly, the interaction variables for total floor space sold also provide evidence of diversion of demand, as the interaction with group 2 and 3 are significantly negative.  Therefore, the coefficients on interaction variables are consistent with the initial expectation that smaller homes are now in demand and the high end housing market is slowing down due to these housing control policies.      	   80	  Table 5.3: Housing demand, Dependent variable: log (total floor space sold) Regression Type Regr. (a9)  Regr. (a10)  Regr. (a11)  Regr. (a12)  Tested Policy 1st home HPF Loan 2nd home HPF loan 2nd home Bank Loan 3rd Home Purchase Restriction       Log price change -0.0839*** -0.0697** -0.102*** -0.117***   (𝑡 − 1) (0.0310) (0.0320) (0.0319) (0.0331) Log real estate investment change 0.151 0.174 0.227 0.239   (𝑡 − 1) (0.247) (0.235) (0.242) (0.241) Change in the loan interest rate -1.568** -0.0295 -0.593 -1.079   (𝑡 − 1) (0.689) (0.818) (0.846) (0.759) Change in Regional GDP per capita 0.121 0.0186 0.0736 0.110    (𝑡 − 1) (0.0974) (0.103) (0.103) (0.0987) Down payment level 0.0122 0.00132 -0.000393   (0.0129) (0.00788) (0.00794)  Down payment level 5.24e-05 -0.0309*** -0.0366***        (𝑡 − 1)  (0.0198) (0.00806) (0.00858)  Lag (1) policy interact with group 1   0.00918** 0.00897 -0.0733     (t-q)*group1    (0.00432) (0.00735) (0.138) Lag (1) policy interact with group 2 -0.0477** 0.00315 0.00562 -0.0389     (t-q)*group2 (0.0201) (0.00465) (0.00812) (0.153) Lag (1) policy interact with group 3 -0.0427** 0 0 0     (t-q)*group3 (0.0189) (0) (0) (0) Purchase restriction    0.286** (Dummy variable)    (0.117) Purchase restriction    0.402***    (𝑡 − 1)      (0.140) Constant 3.384*** 4.336*** 5.487*** 3.729***  (0.404) (0.218) (0.502) (0.126) Number of groups 27 27 27 27 Observations 387 391 401 401 R-squared 0.676 0.698 0.696 0.701     • Robust standard errors in parentheses (*** p<0.01, ** p<0.05, * p<0.1). Data are quarterly from 2006 to 2013. All regressions include quarterly dummies, city dummies, group dummies, and a national time trend. Standard error for variables clustered by city_time allows for intragroup correlation. • 𝑞 = 2 in Regr. (a9), 𝑞 = 2 in Regr. (a10), 𝑞 = 1 in Regr. (a11), 𝑞 = 1 in Regr. (a12)  	   81	  6. Conclusion This paper provides a history of housing policy reforms in China, focusing on investigating policy adjustments that took place from 2006 to 2013. Using quarterly data of 9 cities from the past 7 years, the empirical research finds that while minimum down payment required by the Housing Provident Fund tends to have a significant negative correlation with the housing price change and quantity demanded of the previous period, quantity demanded tends to shift back in the period when the policy actually takes effect.  The minimum down payment required for buying the 2nd home using a bank loan has a stronger negative association with housing price changes and quantity demanded than the HPF loan. A 10-percentage point increase in the minimum down payment of a bank loan correlated with a 68.1% drop in price change rate, a 33.1% drop in unit sales, and a 36.9% decrease in total floor space sold over two quarters.  Surprisingly, the estimates show that the purchase restriction has a strong positive correlation with housing price changes and quantity demanded. Cities that implemented the 3rd home purchase restriction is positively associated with 85.2% increase in price changes, 24.1% increase of sales volume and 68.8% increase of total floor space sold over two quarters.  In addition, the regression analysis provides evidence that the impacts caused by policies differ based on the size of units sold. Restraining financing policies significantly increases total floor space sold for properties smaller than 90m2 by as much as 9.8% more than bigger units, which means housing demand are redirecting from bigger units to smaller homes.  	   82	  These results suggest a number of questions. One issue is that the positive effect caused by the purchase restrictions on housing demand contradicts the very purpose of establishing such policies. Owning multiple properties is prevalence in many developed countries. Chinese government can be better off by removing the purchase restriction to let the market adjust itself. Restrictions shifting the housing demand from larger units towards smaller homes suggest that the affordability of consumers is lowered. It may not be fair to families who need larger units to accommodate their family members in small units due to financial burdens and purchase restrictions.  Another outstanding question is the extent to which the housing control policies themselves are endogenous. It is possible that municipal governments impose more onerous policies on a market segment that has intrinsically lower demand for housing. Also, municipal governments can implement restrictions with the anticipation that the local housing market will heat up in the future. Moreover, since cities have flexibility as to when to implement national guidelines, a city with higher housing prices and unit sales is more likely to impose more stringent policies sooner. However, an instrument variable for possible simultaneity of housing demand and regulation is difficult to find, therefore the regression does not include any instrument variables.   Data analysis in this research mainly focuses on the impact to the housing demand. Decrease in housing unit supply and demand will result in ambiguous price changes. The analysis could be more accurate if it is possible to obtain more data from the supply side of real estate market in China. The development activity across different market segment answers the question whether 	   83	  the policy redirect development and investment into other less restricted sub-markets. Therefore, future research on this topic should look at the policy impact from both sides.  	    	   84	  Bibliography Arnott, R. (1987), “Economic Theory and Housing,” in E.S. Mills, ed., Handbook of Regional and Urban Economics, Volume II, Urban Economics, 959-88, Amsterdam: North-Holland Publishing. Bhattacharya, R., & Kim, S. (2011). Economic Fundamentals, Subprime Lending and Housing Prices: Evidence from MSA-level Panel Data. Housing Studies, 26(6), 897-910. Cao, J. A., & Keivani, R. (2014). The Limits and Potentials of the Housing Market Enabling Paradigm: An Evaluation of China's Housing Policies from 1998 to 2011. Housing Studies, 29(1), 44-68. Cao, Cherry. "SUNAC Agrees to Buy 24.3% Stake in Greentown." Shanghai Daily. N.p., 18 Apr. 2013. Web. 30 June 2014.  Campbell, J. Y. & Shiller, R. J., 1988a. The dividend-price ratio and expectations of future dividends and discount factors. Review of Financial Studies 1, 195-227.  Campbell, J. Y. & Shiller, R. J., 1988b. Stock prices, earnings, and expected dividends. Journal of Finance 43, 661-676.  Yunaxin, Cui. “ Who Become Millionaires in 2009 ?” Sina Real Estate. 31. Dec. 2009. Web. 30 June 2014. Case Karl E., Shiller Robert J.1989. “The Efficiency of the Market for Single-Family Homes”. American Economic Review 79 (1): 125–37.v 79, Capozza D R, Hendershott P H, Mack C, Mayer C J (2002). Determinants of real house price dynamics. NBER Working Paper, No. W9262 Capozza, Dennis R., Patric H. Hendershott, and Charlotte Mack. "An Anatomy of Price Dynamics in Illiquid Markets: Analysis and Evidence from Local Housing Markets." Real Estate Economics 32.1 (2004): 1-32. Web. China Statistical Yearbook. (1988). Beijing: China Statistical Publication House.  China Statistical Yearbook 2013. (2013). Beijing: China Statistics Press.   Chiu, R. L. (1996). Housing Affordability In Shenzhen Special Economic Zone: A Forerunner Of China's Housing Reform. Housing Studies, 11(4), 561-580.   	   85	  Ding, C. (2001). The Benchmark Land Price System And Urban Land Use Efficiency In China. Chinese Geographical Science, 11(4), 306-314.  Collyns C, Senhadji A (2002). Lending booms, real estate bubbles and the Asian Crisis. IMF Working Paper No. 02/20 Ding, C., & Knaap, G. (2005). Urban Land Policy Reform in China's Transitional Economy. Emerging land and housing markets in China (p. 10). Cambridge, Mass.: Lincoln Institute of Land Policy. Dua, P. and D.J. Smyth (1995), “Forecasting U.S. Home Sales Using BVAR Models and Survey Data on Households’ Buying Attitudes for Homes,” Journal of Forecasting, 14, 167-180. Duca, John V; Muellbauer, John; Murphy, Anthony (2011). House prices and credit constraints:making sense of the US experience. The Economic Journal, ISSN 0013-0133, 2011, Volume 121,Issue 552, pp. 533 – 551. Engle, R. F., & Granger, C. W. (1987). Co-integration and error correction: representation, estimation, and testing. Econometrica: journal of the Econometric Society, 251-276. Gallin, J. (2006). The Long-Run Relationship between House Prices and Income: Evidence from Local Housing Markets. Real Estate Economics, 34(3), 417-438. Gerlach S, Peng W (2005). Bank lending and property prices in Hong Kong. Journal of Banking & Finance, 29: 461–481 Goodman Jr, J. L. (1998). Aggregation of local housing markets. The Journal of Real Estate Finance and Economics, 16(1), 43-53. Han, Sunsheng (1998). Real Estate Development in China: A Regional Perspective. Journal of Real Estate Literature, 6: 121-133. Hook, B. (2011). The Development of the Urban Housing Market: Social Implications. China today: economic reforms, social cohesion and collective identities (pp. 105-120). London: Routledge.   Hu, Chunzhi. (2000). On the Development, Methodology and management of Land Value Appraisal System. Internal Report of the State Land Administration, Beijing. Jim, Clare. "China Vanke Says Property Sector's 'golden Era' Over, Demand Outlook Solid." Reuters. Thomson Reuters, 28 May 2014. Web. 30 June 2014. 	   86	   Johansen, S. (1991). Estimation and hypothesis testing of cointegration vectors in Gaussian vector autoregressive models. Econometrica: Journal of the Econometric Society, 1551-1580.  Lau, K., & Lee, J. (2001). Housing Policy Reform. The market in Chinese social policy (pp. 112-137). Houndmills, Basingstoke, Hampshire: Palgrave.  Lee, J. (2000). From Welfare Housing To Home Ownership: The Dilemma Of China's Housing Reform. Housing Studies, 15(1), 61-76.  Malpezzi, S. (1999). A simple error correction model of house prices. Journal of Housing Economics, 8(1), 27-62. Mayer, C. J., & Somerville, C. (2000). Land use regulation and new construction. Regional Science and Urban Economics, 30(6), 639-662. Mayer, C., & Somerville, C. (2000). Residential Construction: Using The Urban Growth Model To Estimate Housing Supply. Journal of Urban Economics, 48(1), 85-109. Meen, G. (2002). The time-series behavior of house prices: a transatlantic divide? Journal of Housing Economics, 11(1), 1-23. Megbolugbe, I.F., Marks, A.P., and Schwartz, M.B. (1991), “The Economic Theory of Housing Demand: A Critical Review,” Journal of Real Estate Research, 6, 381-93. Ortalo-Magne, F., & Rady, S. (2006). Housing market dynamics: On the contribution of income shocks and credit constraints. The Review of Economic Studies, 73(2), 459-485. Pesaran, M. H., Shin, Y., & Smith, R. J. (2001). Bounds testing approaches to the analysis of level relationships. Journal of applied econometrics, 16(3), 289-326. Qi Liang and Hua Cao, ‘Property prices and bank lending in China’, Journal of Asian Economics 18(1), (2007), pp. 63–75. Real estate price in Tianjin increase substantially over the last year . (2004, April 16). National Bureau of Statistics of China. Retrieved November 1, 2013, from http://www.stats.gov.cn/was40/gjtjj_detail.jsp?channelid=57792&record=3642 Schwartz, M.B. (1988), “The Estimation of Housing Demand: A Review of the Research and Its Implications,” Working Paper, National Association of Realtors, Forecasting and Policy Analysis Division.  	   87	  Seven Looming Financial Bubbles - Forbes. (2009, December 15). Information for the World's Business Leaders - Forbes.com. Retrieved July 7, 2013, from http://www.forbes.com/2009/12/15/looming-financial-bubbles-personal-finance-bubbles.html   Smith, L.B., Rosen, K.T., and Fallis, G. (1988), “Recent Developments in Economic Models of Housing Markets,” Journal of Economic Literature, 26, 29-64.  Wang, R. (2009). The Structure Of Chinese Urban Land Prices: Estimates From Benchmark Land Price Data. The Journal of Real Estate Finance and Economics, 39(1), 24-38. Vargassilva, C. (2008). Monetary Policy And The US Housing Market: A VAR Analysis Imposing Sign Restrictions. Journal of Macroeconomics, 30(3), 977-990. Wang, Y., & Murie, A. (1999). Housing policy and practice in China. New York: St. Martin's Press.   Westlake Intime Project Phase Two Completed (2013, May 6). Investment Invitation. Uptown District of Hangzhou. Retrieved November 1, 2013, from http://122.224.136.14/foreground/newsContent.jsp?newsId=445. Wei, Y., Lam, P. T., Chiang, Y., Leung, B. Y., & Seabrooke, W. (2014). An Exploratory Analysis of Impediments to China's Credit Control on the Real Estate Industry: an institutional perspective. Journal of Contemporary China, 23(85), 44-67. Xue, M. (1985). Almanac of China's economy, 1984. Hong Kong: Modern Cultural Co. Yang, Z., & Shen, Y. (2008). The Affordability Of Owner Occupied Housing In Beijing. Journal of Housing and the Built Environment, 23(4), 317-335.   Yeh, A. (2005). The Dual Land Market and Urban Development in China. Emerging land and housing markets in China (p. 40). Cambridge, Mass.: Lincoln Institute of Land Policy. Ye, J. P., & Wu, Z. H. (2008). Urban housing policy in China in the macro-regulation period 2004–2007. Urban Policy and Research, 26(3), 283-295. Yu, H. (2010). China’s house price: Affected by economic fundamentals or real estate policy? Frontiers of Economics in China, 5(1), 25-51.    	   88	  Appendices Appendix 1: Summary of policies at city and national level from 2007 to 2013 	  	  	   89	  	  	  	   90	  	  	  	   91	  	  	  	   	  	   92	  Appendix 2: Summary of national policies from 1978 to 2013 	  	  	   93	  	  

Cite

Citation Scheme:

        

Citations by CSL (citeproc-js)

Usage Statistics

Share

Embed

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"
                            src="{[{embed.src}]}"
                            data-item="{[{embed.item}]}"
                            data-collection="{[{embed.collection}]}"
                            data-metadata="{[{embed.showMetadata}]}"
                            data-width="{[{embed.width}]}"
                            async >
                            </script>
                            </div>
                        
                    
IIIF logo Our image viewer uses the IIIF 2.0 standard. To load this item in other compatible viewers, use this url:
http://iiif.library.ubc.ca/presentation/dsp.24.1-0102479/manifest

Comment

Related Items