East Asia, which houses 22% of the total world population, has been experiencing an explosive and sustained output growth in the post World War II era. Besides the rapid economic growth, the region is also characterized by a high population density, with a density of 350 persons per square mile, about three times higher than the world average. In the process of the growth, and the industrialization and urbanization that came with it, the primate cities and other major urban areas in the region often faced with chronic shortage of land and real estate space, causing a rampant appreciation of property value. That, in turn, contributed to a rise of speculative demand for housing and unaffordable shelter for low and moderate income households. In the supply side, various natural as well as man-made constraints made the supply response inelastic and exacerbated the price volatility and housing affordability. Furthermore, thanks to the liberalization of financial markets, the residential mortgage lending sector has been rapidly evolving, which raised the affordability of home purchase for lowincome and less-creditworthy borrowers on the one hand but also made it more likely to have an inter-sectoral contagion between housing and macro-economy in those countries, both in upturns and in downturns.
This study aims to offer comparative analyses on real estate sectors of six East Asian (EA) countries – China, Hong Kong, Japan, Korea, Singapore and Taiwan – by: (1) documenting episodes of housing price boom-bust as identified in the literature; (2) performing data analyses, both exploratory and explanatory, on determinants of housing price dynamics in the study area; and, (3) investigating institutional factors of relevancy observed from those countries. In so doing, we put an extra effort to cover both market fundamentals and key institutional aspects of real estate system by focusing on land and housing supply, mortgage lending, and real estate taxation.
A special focus of this comparative study is on those housing cycles whose magnitude, i.e., its duration and amplification, is large enough to inflict a spillover to macro-economy and financial system in a given country, often referred to as a contagious real estate cycle. We attempt to document the EA countries’ experiences on such cycles, in terms of similarities and dissimilarities with the relevant evidences compiled to for the U.S. and European countries. Our analyses to date reveal the following as our main findings.
First, both mean and volatility of housing price changes observed from the EA countries are shown to be non-stationary (i.e., timevarying): that is, while Singapore and Hong Kong exhibit a high-growth high-volatility pattern in housing price dynamics in 1990s, the Chinese cities – Shanghai and Beijing – replace them in 2000s with the top level mean and volatility; and, the Korean cities – Seoul and Busan, are underperformers in 1990s with low return and high dispersion, but become an average performer in 2000s with the on-trend mean and volatility during the decade. The conditional volatility, i.e., the dispersion after controlling a persistence factor in housing price movement, confirms the above results as well.
Second, given that the mean and volatility examined in the above do not tell much on how risky a price cycle is in terms of its spillover to macroeconomy and financial system, we check the duration and depth of the price cycles observed from the EA countries, and compare them to those of other advanced countries as a comparison group that experience a pronounced boom-bust in the 2000s. The results show that: the cycles of the EA countries exhibit a similar average annual price appreciation during the boom (12.8% vs. 11.5% by the comparison group, both in real term), a slightly milder average annual price decline during the bust (-8.1% vs. -11.2%), but a much shorter average duration of the boom (11.4 quarters vs. 29.8 quarters by the comparison group). Similar results are documented by the recent studies in that: the most EA cities examined exhibit a smaler autocorrelation parameter but a larger mean reversion coefficient than the U.S. cities. Although this outcome may indicate that the housing price processes in the region are more informationally efficient with a lower serial correlation and a speedy mean-reversion, the result is possibly caused by data and institutional factors, which warrant a further investigation.
Third, in terms of the role of market fundamentals, we estimate the housing price equation as an inverse demand function and document both similar, as well as dissimilar, patterns between the EA countries and the Western states. That is, both household income and the user cost show expected signs with statistical significance for the EA countries, as is the case for the Western counterparts. However, the quantity variable shows a positive and significant effect on housing price for the most primate cities of the EA countries, whereas it is negative and significant for Los Angeles as expected. The result implies that, while the typical price-quantity interaction applies for Los Angeles, housing price and housing supply tend to co-move in the EA cities, possibly due to common economic factors that influence both.
Fourth, we estimate the housing price-to-rent ratio (PRR) equation to examine drivers of price dynamics from the viewpoint of housing asset market. Due to data limitation, the equation is fitted only for two geographical areas in Korea – the capital region (including Seoul and metropolitan areas in its vicinity) and non-capital region (including six major cities not in the capital region). As main findings, while the user cost is a negative and significant driver of PRR in Korea (but with a smaller elasticity compared to that of the U.S.), the mortgage lending variables also affect PRR positively and significantly, implying that PRR is an important channel that links the real and financial sectors of the housing markets in the country.
Fifth, in terms of other policy shocks, shifts in monetary policy as well as financial liberalization often triggered pronounced housing price boom-busts in the EA countries. To name a few, examples of such policy shifts include the Plaza Accord in Japan in the early 1980s, the financial liberalization in Taiwan in the mid- to late-1980s, the liberalization of real estate lending in Korea after the Asian Financial Crisis (AFC) in the late 1990s, and the stimulus package in China after the Global Fnancial Crisis (GFC) in 2007~09. Even a change in public housing policy combined with a strong economic growth is shown to have caused a cyclical price movement, as in Singapore in mid 1990s (i.e., the expansion of the eligibility for owning “the HDB flats”). Nevertheless, the region-wide economic shocks – AFC and GFC – exerted downward pressures in housing price paths for most EA countries, although Taiwan was virtually unaffected by both.
Sixth, as a small open economy, four EA countries – Hong Kong, Korea, Singapore, and Taiwan – are shown to be vulnerable to international capital flows, which often work as a volatility enhancer in real estate markets. Related to this point, there appears to be a “China effect” on housing price dynamics of several EA countries in mid- late-2000s. That is, possibly due to expectations of positive spillover from the China’s strong economic performance, housing prices in Taiwan, Singapore, and Hong Kong exhibit fairly strong upward trends from mid 2000s and again after GFC, while those in Korea and Japan do not show similar post-GFC co-movement patterns. A more formal investigation on this seeming effect, along with the role of external capital flow, are subject to future research.
Finally, we attempt to describe the key features of the institutions related to the functioning of the real estate sector in the six East Asian countries studied and to explain how the differences in institutions translate into the price volatility in their housing markets. To that end, examine key features of mortgage markets and government intervention thereof, followed by description of important aspects of arrangements and protection of property rights, land use regulations, regulations governing the allocation of new housing, and taxes on and subsidies to the housing sector. Using a composite index that represents property rights protection and strictness of supply-side regulations, the extent of supply elasticity and that of the price volatility are assessed. It appears that the first linkage is reasonably well-grounded while the second linkage is harder to explain the data. In light of the complexity of the factors that influence the second linkage, more and better data as well as better modelling methodology would be required.