Relationship between Proportion of Private Housing Completions, Amount of Private Housing Completions, and Property Prices in Hong Kong Bauhinia Foundation Research Centre May 2014
Background Tackling Hong Kong s housing problem is the top priority of the current Government, and the shortage of housing supply is viewed as the crux of the prevailing housing problem. 1 In order to resolve the housing shortage problem, the Government aims to provide a total of 470,000 units in the coming ten years, of which 60% will be public housing. 2 For the first five-year period (2014-2018), the Government plans to provide about 80,000 public housing units, and the private sector is expected to produce about 68,000 flats. 3 For the second fiveyear period (2019-2023), the Government aims to build about 200,000 public housing units, and the private sector is expected to complete about 122,000 flats. 4 2. The Government is trying to tackle the housing shortage by increasing the overall housing supply. 5 It is worth noting that the Government intends to increase the proportion of public housing in the future housing completions. While some grassroots groups have argued that the proportion of public housing completions should reach at least 60%, or even 70%, others are concerned about the decline in the proportion of private housing completions and its impact on the property market. 6 People are worried that once the amount of private housing completions declines, property prices may rise further. 3. In fact, even though the proportion of public and private housing completions varied over the past ten years in Hong Kong, property prices still kept rising. Is there any relationship between the proportion of private housing completions and property prices? And it is worth examining whether there is any relationship between the amount of private housing completions and property prices. 1 The 2013 Policy Address, 16 January, 2013. 2 The 2014 Policy Address, 15 January, 2014. 3 Private housing completions are defined as those units granted an Occupation Permit which is a statutory document issued by the building department certifying the satisfactory completion of a new building, including those private flats converted from subsidised sale flats. 4 Speech from the Secretary for Transport and Housing at a special meeting of the Legislative Council Panel on the 2014 Policy Address, press release by Information Services Department, 28 January, 2014. 5 Same as 2. 6 Cheung Bing-leung: maintain the recommended ratio public and private housing as 6:4, the RTHK website, 21 December, 2013. 1
Study Objectives and Methodology 4. This study aims to examine two areas as follows: (1) the relationship between the proportion of private housing completions and property prices in Hong Kong; and (2) the relationship between the amount of private housing completions and property prices in Hong Kong. By using the housing statistics over the past ten years (2003-2012), we first examine the situation of the proportion of private housing completions and changes in property prices, and further explore whether the two variables are related in terms of their correlation coefficient. Besides, we employ the Granger causality test (details on Page 6) to examine if the lead-lag relationship exists between the changes in both private housing completions and property prices. Study Limitations 5. This study is not focusing on property market analysis, so other important factors that may affect property prices, including the quantitative easing policy in the United States, the hot money inflow from the Mainland, the linked exchange rate system, etc., are not covered in this study. In addition, previous property journals generally focused on the relationship between two variables, such as the examination of the causal relationship between property prices and rents in Hong Kong (Cheung et al., 1995) 7, and the analysis of the association between property prices and the performance of real estate stocks in Hong Kong (Newell and Chau, 1996) 8. Given the premise that the land supply is mainly managed by the Government, our analysis will only focus on the relationship between the proportion of private housing completions and property prices, and the relationship between the amount of private housing completions and property prices. 7 Cheung, Y.L., Tsang, S.K., and Mak, S.C. (1995) The causal relationships between residential property prices and rentals in Hong Kong: 1982-1991, Journal of Real Estate Finance and Economics, 10: 23-35. 8 Newell, G., and Chau, K.W. (1996) Linkages between direct and indirect property performance in Hong Kong, Journal of Property Finance, 7: 9-29. 2
Study Findings The proportion of private housing completions and changes in property prices 6. According to the data of housing completions in the past ten years (2003-2012), the average proportions of the public and private housing completions were 42:58 for the period of 2003-2007 and 62:38 for the period of 2008-2012 respectively (see Table 1). In 2003-2007, the average proportion of private housing completions was higher than that of public housing completions, and property prices rose nearly 70% over the same period. Although property prices fell by 12% in 2003, due to the outbreak of Severe Acute Respiratory Syndrome (commonly known as the SARS ) in Hong Kong, property prices rebounded rapidly by 27% in 2004. The cumulative increase of 68% was recorded in the following five years between 2003 and 2007. In 2008-2012, the average proportion of public housing completions was higher than that of private housing completions, and property prices still increased by 70%. 7. Thus, even though the average proportions of public and private housing completions in two periods (2003-2007 and 2008-2012) were different (i.e. a higher proportion of private housing in 2003-2007 and a higher proportion of public housing in 2008-2012), property prices still kept rising as a result of the continuously low supply of private housing completions. 3
Table 1: Number of Public and Private Housing Completions, Proportion of Private Housing Completions and Changes in Property Price Index during 2003-2012 Year Public Housing Units (including public rental housing and subsidised sale flats Private Housing Units (excluding village houses) Total Number of Housing Completions Public and Private Housing Proportion of Completions 9 2003 14,268 26,397 40,665 35 : 65 Average Proportion of Completions Year-on- Year Changes in Property Price Index -11.9% 2004 20,947 26,036 46,983 45 : 55 26.6% 2005 24,691 17,321 42,012 59 : 41 42 : 58 17.9% 2006 4,430 16,579 21,009 21 : 79 0.8% 2007 6,805 10,471 17,276 39 : 61 11.7% 2008 25,831 8,776 34,607 75 : 25 16.4% 2009 19,391 7,157 26,548 73 : 27 0.7% 2010 7,495 13,405 20,900 36 : 64 62 : 38 24.4% 2011 17,787 9,449 27,236 65 : 35 20.7% 2012 9,778 10,149 19,927 49 : 51 13.2% Source: Information note on Long Term Housing Strategy by Legislative Council Secretariat, Hong Kong Housing Authority, and Rating and Valuation Department. 68% 71% 9 Proportion of public housing completions = Public housing completions / Total housing completions; Proportion of private housing completions = Private housing completions / Total housing completions. 4
Number of Units The relationship between the proportion of private housing completions and property prices 8. We try to examine the relationship between the proportion of private housing completions and property prices in terms of their correlation coefficient 10. Data from 2003 to 2012 show that the correlation coefficient between the proportion of private housing completions and property prices is -0.24 11, which indicates a negative correlation between the two variables. However, with the p-value of 0.50 12, it shows that their relationship is not statistically significant (see Table 2). 9. It is noted that private housing completions in Hong Kong over the past ten years (2003-2012) was only 45% of that in 1993-2002 (see Table 1 and Figure 1). Under the circumstances of continuous shortage of private housing completions, property prices kept rising, regardless of the proportion of private housing completions. This may account for no significant relationship between the proportion of private housing completions and property prices. Figure 1: Number of Private Housing Completions for Every Ten-year during 1983-2012 350,000 300,000 250,000 200,000 150,000 100,000 50,000 0 302,200 266,230 145,740 1983-1992 1993-2002 2003-2012 10 The correlation coefficient is used to reflect the relationship between two variables. When the value is greater than 0, it means that the two variables are positively correlated; when the value is less than 0, it means that the two variables are negatively correlated. 11 The degree of correlation can be generally divided into three levels: low correlation when the absolute value is less than 0.4; significant correlation when the absolute value is between 0.4 and 0.7; high correlation when the absolute value is between 0.7 and 1.0. 12 p-value is the probability that reflects the possibility of an event occurring; generally p- value<=0.01 indicates very significant, p-value<=0.05 indicates moderate significant, and p- value<=0.1 indicates slight significant. 5
Source: Rating and Valuation Department. Table 2: Correlation between Proportion of Private Housing Completions and Changes in Property Price Index during 2003-2012 Proportion of Private Housing (including all classes) Year Private Housing Property Price Year-on-Year Changes in Completions Index 13 Property Price Index 2003 65% 61.6-11.9% 2004 55% 78.0 26.6% 2005 41% 92.0 17.9% 2006 79% 92.7 0.8% 2007 61% 103.5 11.7% 2008 25% 120.5 16.4% 2009 27% 121.3 0.7% 2010 64% 150.9 24.4% 2011 35% 182.1 20.7% 2012 51% 206.2 13.2% Correlation -0.24 = (p-value = 0.50) Source: Information note on Long Term Housing Strategy by Legislative Council Secretariat, and Rating and Valuation Department. The relationship between the amount of private housing completions and property prices 10. We try to employ the Granger causality test 14 to examine whether the lead-lag relationship exists between the changes in both private housing completions and property prices. 11. By using the Granger causality test, we first examine two sets of data, including year-on-year changes in private housing completions (in percentage) and year-on-year changes in property prices (in percentage) over the past ten years (2003-2012). The test results over the past ten years (2003-2012) show that there is no lead-lag relationship between the changes in private housing completions in previous one year, two years or three years and the changes in property prices in the current year. Similarly, there is no lead-lag relationship between the changes in property prices in previous one year, two years or three 13 Based on 1999 = 100. 14 The Granger causality test is a statistical hypothesis test for determining whether the leadlag relationship between the two variables exists at the same period of time. 6
years and the changes in private housing completions in the current year (see Table 4(A)). 12. We further examine the same set of data covering the period of 1991-2012 (see Table 3) by using the Granger causality test. Taking five years as the general completion period for a residential housing project 15 into account, we include the changes in both private housing completions and property prices in previous five years in our analysis. The test results show that there is no lead-lag relationship between the changes in private housing completions in previous one year, three years or five years and the changes in property prices in the current year for the period of 1991-2012 (see Table 4(B)). However, there is lead-lag relationship between the changes in property prices in previous five years and the changes in private housing completions in the current year. It may be related to several major financial crises in Hong Kong in the past, which led to a slump in the economy and also in the property market. Developers might therefore have lost confidence in the property market, which was reflected in the reduction of private housing completions after five years. 15 Lot owners are required to obtain an occupation permit from the Building Authority within the Building Covenant period generally ranges from 48 to 72 months from the date of the land lease document. We try to use the mid-point, 60 months (five years) as the general completion period for a residential housing project in our analysis. 7
Table 3: Number of Private Housing Completions, Percentage Changes in Private Housing Completions, Property Price Index, and Percentage Changes in Property Price Index during 1990-2012 Year Private Housing Completions (excluding village houses) Year-on-Year Changes in Private Housing Completions Property Price Index 1990 29,400 44.8 Year-on-Year Changes on Property Price Index 1991 33,380 0% 61.1 36% 1992 26,222-5% 85.2 39% 1993 27,673-4% 93.0 9% 1994 34,173-2% 114.9 24% 1995 22,621-4% 107.3-7% 1996 19,875-9% 116.9 9% 1997 18,202-6% 163.1 40% 1998 22,278-4% 117.1-28% 1999 35,322 1% 100.0-15% 2000 25,790 3% 89.6-10% 2001 26,262 5% 78.7-12% 2002 34,035 12% 69.9-11% 2003 26,397 3% 61.6-12% 2004 26,036-6% 78.0 27% 2005 17,321-6% 92.0 18% 2006 16,579-7% 92.7 1% 2007 10,471-20% 103.5 12% 2008 8,776-18% 120.5 16% 2009 7,157-24% 121.3 1% 2010 13,405-6% 150.9 24% 2011 9,449-13% 182.1 21% 2012 10,149-1% 206.2 13% Source: Rating and Valuation Department. 8
Table 4: Results of Granger Causality Test for Changes in Private Housing Completions and Changes in Property Price Index Null Hypothesis (1 ) (A) Assuming changes in private housing completions in year (t-1) do not have lead-lag effect on changes in property prices in year (t) (2) Assuming changes in property prices in year (t-1) do not have lead-lag effect on changes in private housing completions in year (t) Assuming changes in private housing completions in year (t-2) do not have lead-lag effect on changes in property prices in year (t) Assuming changes in property prices in year (t-2) do not have lead-lag effect on changes in private housing completions in year (t) Assuming changes in private housing completions in year (t-3) do not have lead-lag effect on changes in property prices in year (t) Assuming changes in property prices in year (t-3) do not have lead-lag effect on changes in private housing completions in year (t) Null Hypothesis (B) Assuming changes in private housing completions in year (t-1) do not have lead-lag effect on changes in property prices in year (t) Assuming changes in property prices in year (t-1) do not have lead-lag effect on changes in private housing completions in year (t) Assuming changes in private housing completions in year (t-3) do not have lead-lag effect on changes in property prices in year (t) Assuming changes in property prices in year (t-3) do not have lead-lag effect on changes in private housing completions in year (t) 2003-2012 F statistics p-value (3) 0.02 0.90 0.32 0.59 0.07 0.94 0.51 0.63 0.30 0.82 0.05 0.98 1991-2012 F statistics p-value 0.19 0.66 1.77 0.20 0.80 0.52 1.91 0.18 Results Results 9
Null Hypothesis (B) Assuming changes in private housing completions in year (t-5) do not have lead-lag effect on changes in property prices in year (t) Assuming changes in property prices in year (t-5) do not have lead-lag effect on changes in private housing completions in year (t) 1991-2012 F statistics p-value Results 0.68 0.66 4.72 0.04** Notes (1): The null hypothesis is generally assumed true until evidence indicates otherwise. (2): In brief, assuming year t is 2012, where we test if property prices in 2012 are affected by private housing completions in previous year(s) (one year, three years or five years), we adopt the Granger causality test by using the changes in private housing completions in previous one year (i.e. 2011, t- 1), three years (i.e. 2009, t-3) or five years (i.e. 2007, t-5) to examine whether the lead-lag relationship exists between private housing completions and property prices. (3): When the p-value < 0.1, the null hypothesis can be rejected. Conclusion 13. In recent years, Hong Kong s property prices have continued to rise and exceeded what most people can afford, leading to a strong demand of affordable housing supply in the society. According to the statistics from the Rating and Valuation Department, the property price index was 246.3 in August 2013, recording a new historical high. At the same time, the sufficient global liquidity and the sustained low interest rate environment were still pushing up the property prices. Based on the above data analysis, we can conclude the following two key points: (i) There is no relationship between the proportion of private housing completions and property prices in terms of their correlation coefficient. In recent years, the amount of Hong Kong s private housing completions has decreased (see Table 3), but under such circumstances property prices have continued to rise. Based on the projections by the Long Term Housing Strategy Steering Committee, the quantity demand of Hong Kong s private housing 10
would be 18,800 units per year from 2014 to 2023. 16 However, the average private housing completions only met 14,500 units per year from 2003 to 2012. Hence, in order to assist home ownership and maintain a healthy growth in the property market, the adequate private housing completion is essential, while the proportion of private housing completions is relatively less important; and (ii) In general, previous changes in private housing completions should have some effects on property prices, but our study finds that their relationship does not exist in Hong Kong. There is no lead-lag relationship between the changes in private housing completions in previous one year, two years, or three years and the changes in property prices in the current year. It might be due to the continuous low level of private housing completions over the past ten years (2003-2012), which makes the changes in private housing completions irrelevant. Nevertheless, when examining the data from a longer period of time (1991-2012), we find that the changes in private housing completions can be affected by the changes in five years ago. It reflects that developers become prudent and conservative due to the prevailing economic downturn, which affects private housing completions after five years. 14. In conclusion, in order to tackle Hong Kong s housing problem, it is necessary for the Government to raise the land supply for housing completions. Since the Government has introduced the demand side management measures 17, property prices eventually cooled down in the third quarter of 2013. However, the property prices are still at a high level and most people who intend to purchase their own home still cannot afford home ownership. Hence, we are of the view that the Government should expedite the land supply for housing completions on the one hand, while implementing the demand management measures on the other hand, to curb speculative activities leading to the continuous rise in property prices. 16 Long Term Housing Strategy Consultation Document, Long Term Housing Strategy Steering Committee, September 2013. 17 The demand side management measures include Special Stamp Duty, Buyer Stamp Duty as well as Double Stamp Duty and Residential Properties (First-hand Sales) Ordinance introduced in November 2010, October 2012 and February 2013 respectively. 11
~END~ 12