The Housing Market in Israel: Long-Run Equilibrium and Short-Run Dynamics
A study by Yossi Yakhin and Inon Gamrasni of the Bank of Israel’s Research Department examines the characteristics of the Israeli housing market and the factors that have contributed in recent decades to the development of rents, home prices and housing construction.
Since 1960, home prices have been on an increasing trend (Fig. 1). The real price of homes (net of the CPI excluding housing) is rising at a rate of about 2.9 percent per year (the slope of the trend line in Figure 1). The study finds that the long-term rate of home price increase in Israel is not anomalous compared to that in the OECD countries. The study also finds that price development is characterized by long cycles and therefore, despite the long-term upward trend, there may be long periods of price decline within these cycles. For example, the recent cycle in the housing market started in 1997 and has lasted for twenty years (assuming that at the end of the sample the market has actually reached its peak), and in the decade between 1997, the previous peak, and 2007, the most recent trough, home prices have decreased by 25 percent in real terms.
Figure 2 presents the development of home prices relative to GDP per capita. The most striking finding is that this ratio is trendless throughout the sample period. This indicates that over time the rate of increase in home prices is similar to the rate of increase in income. The long cycles are also noticeable in this figure, and it is clear that the prices at the end of the sample are indeed high relative to the income, but at a magnitude that is similar to past peaks.
Prolonged cycles indicate that deviations from the long-run equilibrium can last for many years. These deviations can affect market dynamics in the short run, and therefore, along with analyzing short-run developments, it is important to estimate the long-run relationships. To that end, the study relies on four decades of data, 1980–2019, and is based on structural relationships, as emerges from a basic theoretical model for the housing market, DiPasquale and Wheaton (1992).
The study estimates three basic long-run relationships: the demand for housing services; an asset-pricing equation—where home prices are determined by rents and financial returns; and the supply of housing construction. The shortage or surplus in housing is reflected in the gap between the existing stock of homes and the demand for homes; over- or under-valuation of homes is reflected by the deviation of home prices from their value under the asset pricing equation; and the gap between the actual volume of housing construction and the supply equation reflects the surplus or shortfall of construction activity.
Deviations from the long-run relationships have a decisive effect on short-run dynamics. A shortage of apartments supports an increase in rents, in home prices and in construction activity; overvaluation pushes home prices down and accelerates the pace of construction; and a shortfall of construction, relative to the supply curve, also accelerates the pace of housing construction.
Figure 3 presents the estimation for the surplus or shortage of homes, as derived from the demand equation for housing services. The figure shows the shortage created in the early 1990s as a result of the mass immigration wave from the former Soviet Union countries. Accelerated construction in those years led to some surplus in the market starting in 1998. However, this surplus has eroded over the years, and since 2007, the stock of homes has returned to being lower than required to satisfy demand. The excess demand for housing has continued since then, and at the end of the sample (2019), there is still a small excess demand for housing.
The residual of the asset pricing equation (Fig. 4) indicates that the increase in prices during the mass immigration wave led in its end—in the second half of the 1990s—to overpricing of homes (relative to what was anticipated with respect to the development of rent and financial returns). However, the gradual decline in real home prices, in the aftermath of the immigration wave (Fig. 1), against the backdrop of the surplus in homes (Fig. 3), led to underpricing in the early 2000s. The increase in prices during the last decade has led to some overpricing, but relative to the magnitude of their rate of increase, it is not very large. The study estimates that at the end of the sample, in 2019, home prices were about 5.5 percent higher than their value derived from the asset pricing equation.
The analysis in the study shows that the increase in home prices in 2008–11 was mainly due the underpricing of housing at the beginning of the period. An estimation of the asset pricing equation shows that on the eve of the rise in prices, in the years 2006–07, prices were lower than their long-run equilibrium level by an average of 13.7 percent. The estimation of the short-run dynamics does not track well the sharp increase in prices in 2008–11; however, this is in light of the preceding three years, 2005–07, where the model is unable to track well the decline in prices in those years (Fig. 5). Therefore, this result also supports the conclusion that home prices on the eve of their increase were too low relative to market conditions. We estimate overall that about half of the rise in prices in 2008–11 was due to the underpricing that prevailed on the eve of their rise.
The short-term real interest rate (the Bank of Israel interest rate less the expected inflation rate) explains about a quarter of the increase in home prices in 2008–11, but in subsequent years, its contribution was found to be nil (Figure 5). The shortage of housing and the increase in households' income are moderate but persistent factors contributing to the increase in prices, and since 2012, these have been the two main drivers that have fueled their rise (Figure 5).
The sample in the study ends in 2019, but as for 2020, it was characterized by a shock that was transitory but remarkable in its magnitude, in the COVID-19 crisis, which also affected the housing market: the pace of construction slowed, labor income fell along with an atypical increase in government support to households, and capital market yields declined sharply. Due to the temporary nature of the shock, it is difficult to deduce from the data an assessment regarding the long-run relationships in the market. For example, it appears that basic demand drivers in 2020 supported a significant slowdown in demand, thereby implying a reduction in the shortage of dwellings. Similarly, home prices have increased in the past year at a higher rate than rents, but a sharp drop in financial yields suggests that apartment overpricing has continued to decline. However, due to the temporary nature of the shock and because the model is designed to analyze the drivers and forces of the housing market in the long-run, these estimates should be treated with caution. With the recovery of the economy from the recent crisis, the picture of the long-run drivers of the housing market will be clarified, and the above estimates will be revised accordingly.
DiPasquale, D. and W. C. Wheaton (1992). "The Markets for Real Estate Assets and Space: A Conceptual Framework", Journal of the American Real Estate and Urban Economics Association 20 (1), 181–197.