Abstract
This Selected Issues paper for the People’s Republic of China—Hong Kong Special Administrative Region (SAR) reviews the residential property market and implications of an aging population. The fiscal sector impact of price fluctuations is important in Hong Kong SAR with land sales and stamp duties providing an important source of government revenue. The simulations are based on a dynamic general equilibrium system with forward-looking behavior and rational expectations. Hong Kong SAR’s revenue flows are characterized by relatively volatile nontax items.
I. The Residential Property Market in Hong Kong SAR1
A. Introduction
1. After declining for a number of years following the Asian crisis, property prices rebounded sharply from mid-2003 to early 2005 and then stabilized. As elsewhere, developments in property prices have an important influence on building construction, household consumption, business costs and consumer price inflation. The fiscal sector impact of price fluctuations is also very important in Hong Kong SAR with land sales and stamp duties providing an important source of government revenue (see Chapter III). Over the past two years, real residential prices have increased by 52 percent; office, retail and flatted factory property prices have increased by even more.
2. This recovery in property prices has been broadly in line with developments in fundamentals. While demand side factors, such as high housing affordability, low interest rates, and solid GDP growth have been important, supply-side variables, such as below-average public housing completions and land sales may have also had an influence on property prices. In the period ahead, solid housing affordability and a continuation of the economic expansion should help to attenuate the impact of further interest rate rises, although much depends on the uncertain external environment.
B. Recent Developments in the Property Market
3. Real residential property prices have rebounded sharply over the past two years after having experienced a sustained price decline (Figure I.1). Following the Asian crisis, real prices of dwellings halved over a five-year period, pushing one-in-five mortgagees into a negative equity position. Then, in September 2003, prices and transaction volumes began to rebound sharply, with prices rising, off a much lower base, by 52 percent in two years. While only about 2 percent of mortgagees still have negative equity, prices remain 44 percent below their peak level in 1997. In recent months, prices have fallen slightly and transaction volumes have stabilized, most likely reflecting the 2 percentage point rise in mortgage interest rates over the first three quarters of 2005.
Real Property Price Indices
(Log index: 1980Q4=100)
Citation: IMF Staff Country Reports 2006, 051; 10.5089/9781451816945.002.A001
Sources: CEIC and Staff calculation.Real Property Price Indices
(Log index: 1980Q4=100)
Citation: IMF Staff Country Reports 2006, 051; 10.5089/9781451816945.002.A001
Sources: CEIC and Staff calculation.Real Property Price Indices
(Log index: 1980Q4=100)
Citation: IMF Staff Country Reports 2006, 051; 10.5089/9781451816945.002.A001
Sources: CEIC and Staff calculation.4. Property prices across type of residential property and across sectors have evolved similarly although the magnitudes vary, with luxury housing and office prices leading the way. Within the residential sector, price increases have been most dramatic at the luxury-end of the market, reflecting significant ‘trading up’ and investor activity. This pattern is consistent with past cycles — during booms the price of luxury housing tends to rise by more than the average, however, when the property market is weak, the price of luxury housing falls by a commensurate amount. Prices of offices, retail and flatted factory properties have followed a similar pattern to residential prices although the extent of the price swings has varied – real office prices have risen by a phenomenal 126 percent over the past two years. While the sharp turnaround in prices is notable, it does not look out of place when viewed in the context of earlier property price changes (Table I.1).
Real Residential Property Price Cycles
(year-on-year growth)
Real Residential Property Price Cycles
(year-on-year growth)
| 81–85 | 85–93 | 94–95 | 96–97 | 98–03 | 03–05 | |
|---|---|---|---|---|---|---|
| Average | −16.3 | 13.7 | −10.4 | 27.1 | −13.8 | 20.1 |
| Max | −2.8 | 47.4 | −2.0 | 41.1 | 0.5 | 28.8 |
| Low | −25.0 | −2.4 | −16.1 | 7.3 | −39.9 | 0.5 |
Real Residential Property Price Cycles
(year-on-year growth)
| 81–85 | 85–93 | 94–95 | 96–97 | 98–03 | 03–05 | |
|---|---|---|---|---|---|---|
| Average | −16.3 | 13.7 | −10.4 | 27.1 | −13.8 | 20.1 |
| Max | −2.8 | 47.4 | −2.0 | 41.1 | 0.5 | 28.8 |
| Low | −25.0 | −2.4 | −16.1 | 7.3 | −39.9 | 0.5 |
C. How Do Property Market Cycles Compare With Those In The Region?
5. Property price cycles in Hong Kong SAR tend to be more pronounced and frequent than elsewhere (Figure I.2).2 A cross-country comparison of property price cycles indicates that the scale and, in particular, the frequency of property price swings in Hong Kong SAR is greater than that experienced elsewhere. For instance, over the period 1986–2003 changes in Hong Kong SAR’s property prices were around 3 times more volatile than in other industrial countries. Relative to the rest of Asia, Singapore is the only country whose property price volatility is comparable, which perhaps is not all that surprising given the similarities between the two economies: limited supply of land; small and extremely open economies vulnerable to similar types of external shocks; and large government involvement in the provision of public housing.
Real Property Price Indices in Asia
(Log index: 1991Q3=100)
Citation: IMF Staff Country Reports 2006, 051; 10.5089/9781451816945.002.A001
Note: Prices of all markets are indexed to the same starting point for comparing the growth in property prices relative to their own levels in 1991. The chart does not compare the dollar value of prices across economies.Sources: CEIC and Staff calculations.Real Property Price Indices in Asia
(Log index: 1991Q3=100)
Citation: IMF Staff Country Reports 2006, 051; 10.5089/9781451816945.002.A001
Note: Prices of all markets are indexed to the same starting point for comparing the growth in property prices relative to their own levels in 1991. The chart does not compare the dollar value of prices across economies.Sources: CEIC and Staff calculations.Real Property Price Indices in Asia
(Log index: 1991Q3=100)
Citation: IMF Staff Country Reports 2006, 051; 10.5089/9781451816945.002.A001
Note: Prices of all markets are indexed to the same starting point for comparing the growth in property prices relative to their own levels in 1991. The chart does not compare the dollar value of prices across economies.Sources: CEIC and Staff calculations.6. Over the past couple of years, property prices in Hong Kong SAR have risen by substantially more than in comparator economies. Many industrial countries are experiencing a well-documented prolonged property price boom. In stark contrast, real property prices are currently lower than their 1991 levels in Hong Kong SAR, Taiwan Province of China, Thailand, Korea and Tokyo. Outside of Tokyo, prices appear to have bottomed out, although the speed and scale of the pick up in Hong Kong SAR stands out. That said, there have been rapid price increases in parts of Seoul, which prompted the government to announce in August 2005 a drastic real estate reform package aimed at keeping housing affordable.3
D. Explaining the Current Property Price Cycle
7. Demand, supply and speculative factors are all likely contributors to the sharp turnaround in property prices and the high level of observed volatility. In considering the possible drivers behind the current (and earlier) property price cycle the following features of the property market are relevant. First, the supply of housing is very slow to adjust, as the number of completions is small relative to the existing stock, and there is a long delay from the decision to build to additional supply being available. This means that shocks to demand for housing — such as that caused by changes in household income — are borne by price changes in the short run. Second, the government has historically played an active role in the housing market with policy changes likely to have affected prices. While the government has taken significant steps to withdraw itself from the market, it still maintains a strong presence through the provision of public rental housing and land sales.
Demand factors
8. A greatly improved macroeconomic environment over the past few years has supported the turn around in prices. After suffering from the effects of the Asian crisis, the IT downturn in 2001 and SARS in 2003, economic growth has now been sustained for nine consecutive quarters. Accompanying the expansion has been a significant decline in the unemployment rate from a historical high, low real interest rates, and confidence effects from the Closer Economic Partnership Arrangement and rapid growth in Mainland China. Over the longer term, the trend increase in the demand for housing has been underpinned by demographic effects, including population growth (although the share of the population in the main household formation group (aged 25–44) has been falling) and a decline in average household size.
9. Relatedly, property prices are likely to have been boosted by historically high housing affordability and pent-up demand from potential buyers who stayed out of the market during the sustained price decline. The scale of the property price decline, together with low interest rates, lifted housing affordability to an all-time high by late 2003.4 It is also quite possible that as property prices kept on falling, young people delayed purchasing their first home (and existing home-owners delayed upgrading to a new home) until the market showed signs of turning around, and once it did the large pent-up demand pushed prices up quite quickly.
Affordability Index of Home Purchas ers
(1991Q4 =100)
Citation: IMF Staff Country Reports 2006, 051; 10.5089/9781451816945.002.A001
Source: HKMA.Affordability Index of Home Purchas ers
(1991Q4 =100)
Citation: IMF Staff Country Reports 2006, 051; 10.5089/9781451816945.002.A001
Source: HKMA.Affordability Index of Home Purchas ers
(1991Q4 =100)
Citation: IMF Staff Country Reports 2006, 051; 10.5089/9781451816945.002.A001
Source: HKMA.Supply factors
10. The government operates a still-extensive public housing program that was scaled back substantially in 2002. Hong Kong SAR has one of the largest public housing sectors in the capitalist world, with almost half of the population living in various forms of public housing. The main elements of the program for a long time were public rental units and subsidized home ownership schemes. However, in November 2002 the government announced that it would discontinue the construction and sale of flats under the Home Ownership Scheme (HOS) from 2003 onwards, citing a desire to rely more heavily on the market economy.5 As the government put up an average of almost 13,800 HOS units per year in the last 10 years of the program (which is over half of private unit completions), removing them from the market is likely to have had led to a rise in the demand for private housing to the extent that they are substitutes (Figure I.4)6 The provision of public rental housing continues with the number of completions guided by the goal of maintaining an average waiting time of 3 years for applicants.
Housing Completions
(number of units ‘000s)
Citation: IMF Staff Country Reports 2006, 051; 10.5089/9781451816945.002.A001
Source: CEIC.Housing Completions
(number of units ‘000s)
Citation: IMF Staff Country Reports 2006, 051; 10.5089/9781451816945.002.A001
Source: CEIC.Housing Completions
(number of units ‘000s)
Citation: IMF Staff Country Reports 2006, 051; 10.5089/9781451816945.002.A001
Source: CEIC.11. The government’s ownership of the land and the means of which it sells control of it to developers also plays a critical role in the property market.7 During 1985–1995, an annual limit of 50 hectares was imposed on land sales, as provided under the Sino-British Joint Declaration. The limit was relaxed from 1994 and lifted entirely following the transfer of sovereignty on July 1, 1997. At this time, the government announced its first five-year land disposal program, with the intention of reducing uncertainty over land supply in the medium term. However, the sharp decline in property prices following the Asian crisis prompted the government to suspend land sales in June 1998 for nine months. It is often argued that the restricted supply of land may have helped to reinforce the property boom in the mid 1990s, and that the additional supply that came online around the time of the Asian crisis contributed to the subsequent sustained price decline.
12. In recent years, both the area and value of land sold has been well below historical levels (Figure I.5). In November 2002, it was announced that starting in 2004 land sales would only be triggered through the Application List system first introduced in 1999. Under this system, the government provides developers with a list of lots that will be available in the coming financial year. An interested developer submits to the government an application together with a guaranteed bid price for a particular site. If the guaranteed bid price matches or exceeds the government’s (undisclosed) reserve price the site is then auctioned or tendered publicly to the highest bidder. After a period where no bids successfully triggered an auction, the government announced in June that a bid of at least 80 percent of its (still undisclosed) reserve price would be enough to trigger an auction.8 Still, the site will not be sold unless the highest bid meets or exceeds 100 percent of government’s reserve price. Since then bids successfully triggered auctions for three sites on September 27, 2005.
Land Sales
Citation: IMF Staff Country Reports 2006, 051; 10.5089/9781451816945.002.A001
Source: CEIC.Land Sales
Citation: IMF Staff Country Reports 2006, 051; 10.5089/9781451816945.002.A001
Source: CEIC.Land Sales
Citation: IMF Staff Country Reports 2006, 051; 10.5089/9781451816945.002.A001
Source: CEIC.13. The scaling back of public housing completions and limited land sales may have supported property prices. Peng and Wheaton (1994) find that over the period 1965 to 1990 there is a strong negative correlation between land sales and house price growth with a one-year lag. However, they only find a very weak relationship between land sales and housing completions. The interpretation is that reduced land sales are perceived to cause a smaller long-run supply of land, which eventually will raise rents for housing and land, and in a rational market, these anticipated higher rents are capitalized into higher current housing prices. In contrast, Tse (1998) uses Granger causality tests to show that there is no causal relationship between land supply and house prices. Using annual data from 1975 to 2004, we find that the level of land sales does in fact Granger cause changes in house prices (see Table I.A1 in ANNEX I.1).9
Speculative Activity
14. Excess movements in property prices that cannot be explained by fundamentals are sometimes attributed to speculative activity. The nature of any speculative activity that exacerbates property prices is likely to be different to that commonly seen in financial markets because of high transaction costs and the large share of owner-occupiers whose property decisions are not purely driven by profit motives. However, some authors such as Levin and Wright (1997) argue that these factors can exacerbate speculative behavior. Owner-occupiers are unlikely to be deterred by transaction costs, and as for most households whose home is typically their largest asset, there is a large implicit cost of not entering the market or trading up when property prices are rising. This, together with households’ tendency to base expected future capital gains on recent price movements, suggests a strong desire for “not missing out” builds additional momentum into prices, particularly during a period of rising prices.10
15. Indicators of speculative activity have remained at a consistently high level but are well below the levels reached in the mid-1990s when a bubble is widely thought to have developed. Econometric studies (Kalra et. al. (2000) and Peng (2002) find support for the notion that a bubble in property prices built up in the period before the Asian financial crisis and then subsequently burst. One direct measure of speculative activity is the number of confirmor transactions - investors who buy a property and resell it before transfer takes place on the initial transaction. This indicator has increased and remained at quite a high level, although it is still below the peak reached in 1997.
Number of Confirmor Transactions
Citation: IMF Staff Country Reports 2006, 051; 10.5089/9781451816945.002.A001
Souce : HKMA.Number of Confirmor Transactions
Citation: IMF Staff Country Reports 2006, 051; 10.5089/9781451816945.002.A001
Souce : HKMA.Number of Confirmor Transactions
Citation: IMF Staff Country Reports 2006, 051; 10.5089/9781451816945.002.A001
Souce : HKMA.E. Some Econometric Results
16. An empirical model was estimated to assess how the growth rate of real house prices are related to relevant demand and supply side factors.11 This model is similar to that used in the IMF (2003) and the variables considered were:
Past growth rates of real house prices. If house price growth is persistent, perhaps due to wealth effects or adaptive expectations on behalf of owner-occupiers and investors, then the current growth rate will be positively correlated with past growth.
Past housing affordability ratio. If house price growth shows long-run reversion to fundamentals, this implies that prices would tend to decrease when affordability is low.
Demand fundamentals. House price growth is expected to be positively affected by real per capita GDP growth, real rental rates,12 and population growth; and negatively affected by real interest rates.13 The real effective exchange rate (given the economy’s openness and its significance in an earlier study)14 and stock prices (has both an income and substitution effect) were also considered. Real credit growth is another variable that is often used; however, Gerlach and Peng (2005) show that in Hong Kong SAR causation runs from prices to credit and not the other way around.
Supply fundamentals. House price growth is expected to be negatively affected by public housing completions, residential land sales and construction costs. Causation may also run in the other direction, particularly in regards to private housing completions.
17. The estimated model performs reasonably well in explaining variations in property price changes, including the recent increases (Table I.A3 and Figure I.A1 in ANNEX I.2). Over the estimation period as a whole, the independent variables explain about 70 percent of the variation in property price movements, and the diagnostic and instability tests are broadly satisfactory. The model performs least well during very rapid price changes, such as that during the 1990s price boom, but picks up much of the recent sharp price increase.
18. The model suggests that real house prices display persistence and only a weak long-run reversion to fundamentals. Real property price growth is quite persistent, suggesting that households may use observed changes in house prices as an indicator of future price changes, which raises the risk that prices could deviate from fundamentals for a period of time. Consistent with this, when prices become out of line with income and interest rates, as measured by housing affordability, there is only a very gradual tendency for this misalignment to be corrected.
19. Property prices were found to depend on per capita GDP, real interest rates, rental prices, population, and the real effective exchange rate. All of the demand fundamentals have the expected sign and, with the exception of real stock prices, were statistically significant. The long-run impact of a change in the demand variables is quite significant. For instance, a permanent reduction in interest rates of 1 percentage point would imply over time an increase of 0.4 percent in real house price growth.15 While there are problems associated with comparing parameter estimates across studies, these results suggest that property prices in Hong Kong SAR display less persistence, are less sensitive to interest rates, but more sensitive to some of the other explanatory variables than what property prices in industrial countries are (IMF (2003)).
20. Supply variables were statistically insignificant determinants of property prices, although this may reflect the difficulty in modeling such relationships. In the long run, the level of property prices should be determined by both supply and demand factors, while short-term volatility of property prices tends to be demand driven. Given the model specification, only the short-term price movements could be explored. Thus, while the results confirm the importance of demand-side factors in explaining short-term property price growth, it does not address the conjecture that housing completions and residential land sales may be important determinants of the level of property prices. Another explanation for the statistical insignificance of the supply variables is that the relationship may be too complex to pick up econometrically. For example, prices may respond to the announcement of future land sales in addition to the housing supply when it eventually becomes available; and the method of selling land has changed a number of times through the sample, which may have affected the nature of the relationship between land sales and prices.16
Granger Causality Tests
Granger Causality Tests
| Sample: 1975 2004 | |||
|---|---|---|---|
| Lags: 2 | |||
| Null Hypothesis | Obs | F-Statistic | Probability |
| Real property price growth does not Granger Cause land supply | 27 | 0.201 | 0.819 |
| Land supply does not Granger Cause real property price growth | 5.973 | 0.008 |
Granger Causality Tests
| Sample: 1975 2004 | |||
|---|---|---|---|
| Lags: 2 | |||
| Null Hypothesis | Obs | F-Statistic | Probability |
| Real property price growth does not Granger Cause land supply | 27 | 0.201 | 0.819 |
| Land supply does not Granger Cause real property price growth | 5.973 | 0.008 |
Augmented Dickey-Fuller Unit Root Tests
Augmented Dickey-Fuller Unit Root Tests
| Levels | First differences | |
|---|---|---|
| Real house prices | −2.19 | −4.58* |
| Residential land sales | −8.77* | -- |
| Affordability ratio | −4.17* | -- |
| Real rental index | −2.19 | −4.58* |
| Real GDP per capita | −2.21 | −8.74* |
| Real interest rates | −3.24* | -- |
| Real effective exchange rate | −1.26 | −6.71* |
| Population | −1.22 | −4.54* |
| Hang Seng index | −1.81 | −9.98* |
| Residential land disposal | −8.80* | -- |
| Public housing completions | −6.83* | -- |
| Private housing completions | −2.36 | −2.58 |
Augmented Dickey-Fuller Unit Root Tests
| Levels | First differences | |
|---|---|---|
| Real house prices | −2.19 | −4.58* |
| Residential land sales | −8.77* | -- |
| Affordability ratio | −4.17* | -- |
| Real rental index | −2.19 | −4.58* |
| Real GDP per capita | −2.21 | −8.74* |
| Real interest rates | −3.24* | -- |
| Real effective exchange rate | −1.26 | −6.71* |
| Population | −1.22 | −4.54* |
| Hang Seng index | −1.81 | −9.98* |
| Residential land disposal | −8.80* | -- |
| Public housing completions | −6.83* | -- |
| Private housing completions | −2.36 | −2.58 |
Actual and Fitted Property Price Growth
Citation: IMF Staff Country Reports 2006, 051; 10.5089/9781451816945.002.A001
Actual and Fitted Property Price Growth
Citation: IMF Staff Country Reports 2006, 051; 10.5089/9781451816945.002.A001
Actual and Fitted Property Price Growth
Citation: IMF Staff Country Reports 2006, 051; 10.5089/9781451816945.002.A001
A Model of Real Residential Property Price Growth Sample Period: 1985:4 to 2005:2
A Model of Real Residential Property Price Growth Sample Period: 1985:4 to 2005:2
| Variable | Coefficient | Std. Error | t-Statistic | Prob. | |
|---|---|---|---|---|---|
| Constant | 0.127 | 0.047 | 2.691 | 0.009 | |
| Lagged dependent variable | |||||
| Lagged real house price growth | 0.420 | 0.123 | 3.413 | 0.001 | |
| Reversion | |||||
| Lagged housing affordability ratio | −0.031 | 0.011 | −2.768 | 0.007 | |
| Fundamentals | |||||
| Real per capita GDP (growth) | 0.821 | 0.293 | 2.801 | 0.007 | |
| Real interest rate (percent) | −0.220 | 0.001 | −2.326 | 0.023 | |
| Real rental index (growth) | 0.711 | 0.289 | 2.458 | 0.017 | |
| Real rental index (growth, lagged) | −0.552 | 0.206 | −2.676 | 0.009 | |
| Population (growth, lagged twice) | 2.824 | 0.958 | 2.949 | 0.004 | |
| Real effective exchange rate (growth, lagged) | 0.536 | 0.168 | 3.186 | 0.002 | |
| Real stock price (growth, lagged) | 0.066 | 0.044 | 1.500 | 0.138 | |
| Adjusted R-Squared | 0.69 | ||||
| Durbin-Watson | 205.00 | ||||
| Breusch-Godfrey Serial Correlation LM Test | 0.09 | ||||
| ARCH LM Test | 1.73 | ||||
| White Heteroscedasticity Test | 2.42* | ||||
| Jarque-Bera Normality Test | 0.02 | ||||
| Ramsey RESET Test | 1.56 | ||||
| Chow Test (Break 1994:1) | 1.83 | ||||
| Chow Test (Break 1998:1) | 0.78 | ||||
A Model of Real Residential Property Price Growth Sample Period: 1985:4 to 2005:2
| Variable | Coefficient | Std. Error | t-Statistic | Prob. | |
|---|---|---|---|---|---|
| Constant | 0.127 | 0.047 | 2.691 | 0.009 | |
| Lagged dependent variable | |||||
| Lagged real house price growth | 0.420 | 0.123 | 3.413 | 0.001 | |
| Reversion | |||||
| Lagged housing affordability ratio | −0.031 | 0.011 | −2.768 | 0.007 | |
| Fundamentals | |||||
| Real per capita GDP (growth) | 0.821 | 0.293 | 2.801 | 0.007 | |
| Real interest rate (percent) | −0.220 | 0.001 | −2.326 | 0.023 | |
| Real rental index (growth) | 0.711 | 0.289 | 2.458 | 0.017 | |
| Real rental index (growth, lagged) | −0.552 | 0.206 | −2.676 | 0.009 | |
| Population (growth, lagged twice) | 2.824 | 0.958 | 2.949 | 0.004 | |
| Real effective exchange rate (growth, lagged) | 0.536 | 0.168 | 3.186 | 0.002 | |
| Real stock price (growth, lagged) | 0.066 | 0.044 | 1.500 | 0.138 | |
| Adjusted R-Squared | 0.69 | ||||
| Durbin-Watson | 205.00 | ||||
| Breusch-Godfrey Serial Correlation LM Test | 0.09 | ||||
| ARCH LM Test | 1.73 | ||||
| White Heteroscedasticity Test | 2.42* | ||||
| Jarque-Bera Normality Test | 0.02 | ||||
| Ramsey RESET Test | 1.56 | ||||
| Chow Test (Break 1994:1) | 1.83 | ||||
| Chow Test (Break 1998:1) | 0.78 | ||||
References
Case, Karl E., John M Quigley, and Robert Shiller, 2003, “Home-buyers, Housing and the Macroeconomy”, prepared for the Reserve Bank of Australia conference on Asset Prices and Monetary Policy, Sydney, 18–19 August 2003.
Chan-Lau, Jorge A., 2003, “Determinants and Prospects for Property Prices in Hong Kong SAR”, People’s Republic of China — Hong Kong Special Administrative Region: Selected Issues, (Washington, D.C.: International Monetary Fund)
Gerlach, Stefan, and Peng, Wensheng 2005, “Bank Lending and Property Prices in Hong Kong”, Journal of Banking and Finance, Vol. 29, pp. 461 –481
“Three Current Policy Issues”, September 2004 World Economic Outlook (Washington, D.C.: International Monetary Fund)
Kalra, Sanjay, Mihaljek, Dubravko and Duenwald, Christoph 2000, “Property Prices and Speculative Bubbles: Evidence from Hong Kong SAR”, IMF Working Paper 00/2, (Washington, D.C.: International Monetary Fund)
Levin, E.J., and Wright, R.E. 1997, “The Impact of Speculation on House Prices in the United Kingdom”, Economic Modelling, Vol. 14, No. 4, pp. 567 –585
Peng, Ruijue, and Wheaton, William C. 1994, “Effects of Restrictive Land Supply on Housing in Hong Kong: An Econometric Analysis”, Journal of Housing Research, Vol. 5, No. 2, pp. 263 –291
Peng, Wensheng, 2002, “What Drives Property Prices in Hong Kong?”, Hong Kong Monetary Authority Quarterly Bulletin (August)
Tse, Raymond, 1998, “Housing Price, Land Supply and Revenue from Land Sales”, Urban Studies, Vol. 35, No. 8, pp. 1377 –1392
Prepared by Brenton Goldsworthy (X-36943).
Cross-country comparisons of property price cycles should be interpreted with caution as the data are not always comparable due to definitional and measurement differences.
The real estate reform package included measures such as: increasing financial support to those without home ownership; expanding public rental housing construction; and increasing the capital gains and ownership tax to discourage speculative purchases.
The index is derived by dividing a typical monthly mortgage repayment (assuming a 20-year mortgage on 70 percent of the purchase price for a 50m2 flat) by the median household income.
Prior to the November 2002 announcement, the government had a stated goal of achieving a 70 percent home ownership rate by end 2007 (the rate is currently around 57 percent).
There are still 16,595 units waiting to be disposed of under the HOS. To mitigate the impact these sales may have on the private market, the Housing Authority will announce the timing, number and pricing of the sales one year in advance (the intention is to sell about 6,000 units a year from the beginning of 2007).
Strictly speaking, the government does not sell the land but instead leases it with terms of 50 years or more. The buyer does, however, take full control over the land and can generally transfer the lease to another buyer.
Earlier in the year, the government also took steps to reduce the length of time it takes from the initial bid to the auction from 10 weeks to 7 weeks, simplify the deposit payment requirement, and publish on a monthly basis the number of unsuccessful applications.
Tse (1998) analyses the relationship between land sales and property prices in (a) levels and (b) in first differences to induce stationarity. However, unit root tests suggest that the level of land sales may in fact be stationary.
Case and Shiller (2003) present survey evidence that US homebuyers’ expectations of future capital gains are substantially affected by recent experience.
Quarterly data for the period 1983–2005 were used in the analysis. To obtain real values the nominal values were deflated by the Composite Consumer Price Index excluding the rental component. Most data are readily available from CEIC. The only exceptions are housing affordability, which was provided by the HKMA, and the construction cost index, which is available in the EMED database. The best lending rate was used as a proxy for the mortgage interest rate, although the spread of mortgage interest rate over the best lending rate has changed in recent years. Seasonality and unit root properties of the data series were examined (Table A2 in Annex I.1). For the series that exhibited stable seasonality, the seasonally adjusted series were used in the analysis.
Rental rates in Hong Kong SAR are typically fixed for a two-year period which means that causation cannot run in the other direction from property prices to rental rates (at least not on a contemporaneous basis).
Most mortgages in Hong Kong SAR are based on flexible rates.
A real exchange rate appreciation that reflects an increase in productivity may lead to an increase in expected future household income. Real exchange rate changes may also affect property prices through their impact on business competitiveness and a firm’s decision on where to locate.
Since the beginning of 2005 mortgage interest rates have risen by 2 percentage points. The model suggests that such a rise would reduce property price growth by 0.8 percent. However, over this period, property prices have fallen by much more, indicating that other factors are at play, or that prices this time around are more sensitive than they have historically been.
See Peng (2002) for an example of a study which finds housing supply to be a significant variable in modeling price movements.
