Housing prices have begun falling this year in several advanced economies, a sharp contrast from the increase in prices seen during 2007 in almost all countries save the United States where a housing correction has been underway since late 2005. In real terms, and on a seasonally-adjusted basis, house prices fell in the first half of 2008 at an annual rate of 5 percent to 12 percent in Canada, Denmark, Spain, New Zealand, and the United Kingdom (first figure). How much further are house prices likely to come down? And what are consequences of the declines in house prices for the
Corrections in house prices. As a basis for assessing the potential for house price declines, a first step is to try to account for the increase in house prices that has taken place over the past decade in terms of important driving forces. To this end, real house price growth is modeled as a function of the following variables: growth in per capita disposable income, working age population, credit and equity prices; and the level of short-term and long-term interest rates. Dynamic effects of these variables are captured through the inclusion of lagged real house price growth and an affordability ratio (the lagged ratio of house prices to disposable incomes). This model is estimated for each country using quarterly data for the time period 1970 to 2007.
The increase in house prices not explained by these fundamental factors—referred to as the house price gap—is taken an estimate of the potential for correction in house prices. Of course, the gap estimates could partly reflect omitted fundamental factors, such as changes in supply-side factors in the housing market. Nevertheless, the estimates provide an indication of how large those omitted factors would have to be for the rise in house prices over the past years to be considered an equilibrium outcome.
The second figure shows the house price gaps—the percent increase in house prices during the period 1997 to end-2007 that is not accounted for by fundamentals. Also shown, as an indicator of the robustness of these results, is the range of gap estimates generated by small perturbations of the estimated models. These changes include: using the average value of housing prices over 1994 to 1997, instead of the 1997 value, as the starting point for computing the gap estimates; estimating a parsimonious version of the model with only incomes and interest rates as the driving forces; and changing the dynamic specification by estimating a vector autoregressive model for house prices instead of a single-equation model.
The countries that have experienced the largest unexplained increases in house prices over the past decade are Ireland, the United Kingdom and Australia; house prices in these countries were 20 percent to 30 percent higher in 2007 than could be attributed to fundamentals. A group of other countries—including France, Spain, the Netherlands and Italy—have house price gaps of between 10 percent and 20 percent. The gap estimate for the United States—about 7 percent—is smaller than for most other countries and has been narrowing compared to earlier estimates, partly reflecting the decline in U.S. house prices over the past eighteen months. The range of estimates for each country is about 3½ percent on average, though for Norway, Sweden and the Netherlands the range is considerably higher.
To put these gap estimates in perspective, it is useful to compare them to house price cycles in the advanced economies over the past several decades (OECD, 2006). Between 1970 and 2005, the average house price cycle lasted about ten years, with an expansion phase of six years during which real house prices increased by about 45 percent. During the subsequent four-year contraction phase, real house prices declined about 25 percent, with the range of declines across countries varying from about 10 percent in the United States to over 30 percent in Japan and several European countries.
Thus, if house price corrections were to occur in line with the gaps shown in the second figure, they would be well within the range of previous experience. Moreover, the past evidence indicates that corrections typically occur over several years. Evidence from countries with regional (i.e. sub-national) data suggests that for some regions, price level corrections could be much more pronounced and last longer than the national cycle (Calomiris, Longhofer and Miles, 2008; Estevao and Loungani, 2008).
Macroeconomic consequences. Experience during past housing market cycles can also be a guide to the macroeconomic consequences of these price corrections (Claessens, Kose, and Terrones, 2008; IMF, 2008; IMF, 2004). The evidence suggests, not surprisingly, that the consequences are more adverse if they occur against the context of a weakening economy and tight credit conditions, which is likely to be the situation facing many countries at present. Over the period 1960 to the present, recessions in advanced countries that are associated with house price busts and credit crunches are slightly longer and deeper than other recessions. The duration of a recession is more than one quarter longer in case of a housing bust, total output loss during the recession is somewhat higher, and the unemployment rate increases notably more and for longer in recessions with housing busts (third figure, top panel). Over the 12 quarters following the onset of a recession, the unemployment rate has increased on average by 1.5 percentage points. But in recessions associated with house price busts, the increase in unemployment is 3 percentage points.
There is some evidence that this pattern holds up at both the national and regional levels. As shown in the lower panel of the third figure, during regional recessions in the United States that are associated with a housing bust the peak impact on unemployment is an increase of 4 percentage points, compared with an increase of 2 percentage points for all regional recessions (Estevao and Loungani, 2008).
What about the impact of house price declines on the components of output? There is a growing literature on the possible impact of changes in housing wealth on consumption. Buiter (2008) demonstrates that changes in house prices are redistributions of wealth and hence do not have much impact on net wealth in the aggregate; however, they can affect individual consumption by relaxing collateral constraints. Consistent with this point, Muellbauer (2008) finds that with a careful modeling of the effect of credit market development and deregulation, which raises access to housing collateral, changes in house prices have a medium-run liquidity effect on U.S. and U.K consumption.
The impact on investment is more readily apparent. Claessens et al. find that investment—residential investment in particular—tends to fall more sharply in recessions associated with housing busts and with credit crunches than in other recessions. There are also significant cross-country differences in the extent of the residential investment declines, which in principle can depend on a wide range of characteristics of national financial and legal systems. One important dimension is the ease with which households can access mortgage credit. This can be measured either by the depth of mortgage markets or by an index that summarizes the institutional features of mortgage markets. The mortgage market index incorporates features such as the typical ratio of mortgage loans to property values, the standard length of mortgage loans, the capacity to borrow against accumulated home equity, and the degree of development of secondary markets for mortgage loans. As shown in the top two panels of the fourth figure, declines in residential investment have tended to be higher in countries where households have had more access to mortgage credit. 
Other factors can play a role in explaining the amplitude of the economic cycle following house price corrections. In addition to the characteristics of mortgage markets already discussed, a feature that is important at the present conjuncture is the prevalence of mortgages with variable (as opposed to fixed) interest rates. There are differences within Europe in this respect, with Finland, Ireland and Spain having mostly variable rate mortgages. Higher interest payments (relative to household disposable income) have also been historically associated with bigger declines in residential investment during housing busts—see the bottom panel in the fourth figure. Countries also differ in legal provisions, such as those that govern the recourse that residential mortgage lenders have in the case of defaulted residential mortgages, which can influence foreclosure rates. For many of the countries that have been the focus of study in this box—United Kingdom, Ireland, Germany, Netherlands, France and Spain—debtors are personally liable for the full amount of mortgaged debt, thus reducing incentives for foreclosure. In the United States, mortgage foreclosure is regulated at the state level. In six states, lenders only have recourse to the mortgaged property, which they can repossess and sell. In the other states, debtors are also personally liable for the full amount of the debt, but there are differences on the extent to which lenders can recover the difference between the mortgage debt and the foreclosure sale price; in practice, however, lenders may choose not to seek deficiency judgments mainly because of the time and cost involved.
Another factor is banking sector exposure to the housing sector, which varies across countries as well as across lending institutions within countries. The value of mortgage loans held by banks, expressed as a multiple of their overall market capitalization, gives an indication of their ability to withstand the deterioration of their real estate loan portfolios. This indicator varies from about 4 in Denmark and Germany, under 3 in Spain, about 1.5 in Canada, Japan and the U.K, and under 1 in the United States. Cross-country declines in residential investment during housing cycles have been higher in countries with greater banking sector exposure to mortgage lending, but the effect has not been as strong as that shown earlier with the mortgage debt-to-GDP ratio. Nevertheless, at the present conjuncture, with bank balance sheets under renewed stress and bank equity prices low, the potential for an adverse impact from banking system exposure to mortgage lending on the real economy is perhaps greater than in the past.
Conclusions. Many advanced economies experienced a house price run-up in recent years that is difficult to account for fully in terms of fundamental driving forces such as income growth and interest rates. The correction in house prices appears to have now begun in most of these economies. If past is prologue, these corrections could average around 25 percent and be spread out over a period of two to four years. Past evidence also suggests that cross-country differences in the impact of these corrections on the macroeconomy are likely to depend on the characteristics of their housing finance systems, particularly the ease with which household have been able to access mortgage credits in recent years. This feature is likely to be correlated with the extent of investment declines that occur during the house price corrections and could also have a dampening impact on consumption.
Ahearne, Alan G., John Ammer, Brian M. Doyle, Linda S. Kole and Robert F. Martin, 2005, “House Prices and Monetary Policy: A Cross-Country Study,” Federal Reserve Board, International Finance Discussion Papers, No. 841.
Buiter, Willem H., 2008, “Lessons from the North Atlantic Financial Crisis,” London School of Economics working paper.
Calomiris, Charles, Stanley D. Longhofer and William Miles, 2008, “The Foreclosure-House Price Nexus: Lessons from the 2007-2008 Housing Turmoil,” Columbia University working paper.
Claessens, Stijn, M. Ayhan Kose and Marco Terrones, 2008, “Mechanics of Recessions: Contractions, Crunches and Busts,” forthcoming IMF Working Paper.
Deutsche Bank, 2008, “Housing Correction: U.S. Ahead of Europe,” Global Economic Perspectives.
Estevao, Marcello and Prakash Loungani, 2008, “Housing Markets and Regional Labor Market Dynamics,” forthcoming IMF Working Paper..
Hilbers, Paul, Alexander W. Hoffmaister, Angana Banerji and Haiyan Shi, 2008, “House Price Developments in Europe: A Comparison,” forthcoming IMF Working Paper.
International Monetary Fund, 2008, “The Changing Housing Cycle and the Implications for Monetary Policy,” Chapter 3 in World Economic Outlook (April).
International Monetary Fund, 2007, “What Risks Do Housing Prices Pose for Global Growth, “ Box 2.1 in World Economic Outlook (October).
International Monetary Fund IMF 2004, “The Global House Price Boom,” Chapter II in World Economic Outlook (Spring).
Klyuev, Vladimir, 2008, “What Goes Up Must Come Down? House Price Dynamics in the United States,” IMF Working Paper 08/187.
Muellbauer, John, 2008, “Housing, Credit and Consumer Expenditure,” Centre for Economic Policy Research (CEPR) Discussion Paper No. 6782.
Organization for Economic Cooperation and Development (OECD), 2005, “Recent House Price Developments: The Role of Fundamentals,” OECD Economic Outlook (Paris).
Warnock, Francis E. and Veronica Cacdac Warnock, 2007, “Markets and Housing Finance,” NBER Working Paper No. 13081.
This material appeared in the IMF’s World Economic Outlook released on October 8, 2008. Ercument Tulun and Jair Rodriguez provided research assistance. This box updates analysis presented in the October 2007 and April 2008 World Economic Outlook reports.
These data are provided by the OECD and are based on commonly used national sources, as shown here: http://www.olis.oecd.org/olis/2006doc.nsf/linkto/ECO-WKP(2006)3 (page 34). The data are seasonally adjusted by the OECD in cases where the national authority does not provide a seasonally-adjusted series. The use of seasonally-adjusted data creates some difficulty in comparability with headline figures on house prices but may be a better indication of developments in house prices over the coming months.
The models estimated here focus on explaining short- to medium-run changes in house prices rather than the long-run level of house prices, which could differ considerably across countries reflecting national supply constraints and long-term institutional factors, such as the extent of taxation of housing (Poterba, 1984). A study of European housing markets by Hilbers, Hoffmaister, Banerji and Shi (2008) provides a good exposition of the role such factors can play in house price movements.
As noted in the 2008 IMF staff report for Australia, if some country-specific factors, particularly the impact of long-term migration on housing demand are taken into account, the results do not produce evidence of a significant overvaluation of house prices.
The 2008 Article IV staff report for the Netherlands notes that the estimated house price gap—estimated here as ranging from 9 to 15 percent—is likely to be much smaller if the rise in single-person households, which is very important in the Netherlands as a factor that boosts housing demand, is taken into account, together with institutional factors (e.g., strict zoning regulations and generous mortgage interest deductibility).
Hilbers et al. (2008) group European countries into “fast,” “average” or “slow movers” depending on the extent to which their house prices in recent years have risen above long-term averages. The gap estimates presented here turn out to be consonant with this classification: the average estimated gap for the three groups is 19 percent, 11 percent and -3 percent respectively. Recent Article IV staff reports that point to either a cooling of housing markets or the onset of a correction include Canada, Korea, New Zealand, Norway, Spain and the United Kingdom. For Germany, some studies have found higher undervaluation than the estimate of 5 percent reported here, perhaps reflecting supply side impacts from social housing in Germany post-reunification.
Klyuev (2008) estimates that single-family homes in the United States “remained 8 to 20 percent overvalued as of the first quarter of 2008.” The U.S. house price gap was estimated at about 12 percent in 2007 (IMF 2008, Box 3.1) and about 20 percent in 2006 (IMF 2007, Box 2.1).
Benito (2007) finds, using household-level data for the United Kingdom, that it is much more common for withdrawal from home equity to flow into residential investment than consumer spending, which suggests that the collateral channel stressed by Buiter and Muellbauer could be more stronger for investment than consumption.
Data on the depth of mortgage markets—the ratio of outstanding mortgage debt to income are reported in Warnock and Warnock (2008) and OECD 2006. The mortgage market index is described in IMF (2008). The debt measure used here is the ratio of mortgage debt to household disposable income for the 1990s (from OECD 2006) but the use of other measures of debt—for other years or expressed as a ratio to GDP—gives similar results. Controlling for the magnitude of the house price corrections makes the correlation between residential investment declines and the mortgage debt-to-GDP ratio stronger. Cardarelli, Monacelli, Rebucci and Sala (2008) take this analysis a step further by using sign restrictions to identify housing demand shocks and tracing through the impact of these shocks on house prices, residential investment and output. They conclude that housing finance innovation has amplified the spillovers from housing to the rest of the economy by strengthening the role of housing as collateral.
See Tsatsaronis and Zhu (2004). Warnock and Warnock (2008) add Greece, Portugal, Sweden, and the United Kingdom to the list of European countries with mostly variable rate mortgages; outside of Europe, the United States, Canada, and Japan are classified as countries with mostly fixed rate mortgages.
Estimates for countries other than the United States are from Ahearne et al. (2005) and are based on bank-level data on mortgage loans and market capitalization from Bloomberg and Worldscope; the U.S. estimate is based on total real estate loans by the banking sector and total banking sector market capitalization.