Developments in the Housing Market

1. Against the backdrop of low interest rates and supply shortages, house price increases have been ranmpant. Nominal house prices have risen by 80 since 2007. In response to the global financial and Euro Area crises, the Bank of Israel (BoI) engaged in two rounds of monetary easing. The monetary stimulus supported economic growth, but it also boosted demand in the mortgage and housing markets. At the same time, owing to the peculiarities of the housing market in Israel, household formation has tended to outstrip the supply of homes². This phenomenon is largely explained by fast population growth (owing to mass immigration waves during the 20th century), space constraints related to the way the land market operates³, and delays in the land planning approval and building permit issuance processes⁴. As a result, positive demand shocks tend to have large price effects. By contrast, in countries where supply is not constrained, new construction quickly catches up with increased demand, thereby dampening house price movements.

2. Price-to-income and price-to-rent ratios are also well above their equilibrium value. These ratios are, respectively, 26 and 22 percent above their long-run average, suggesting a house price deviation from fundamentals of the same magnitude, as rents and incomes serve as long-run anchors for property prices. In the short run, however, house prices can rise above their equilibrium value, because supply cannot respond quickly to changes in demand. Similarly, increases in the availability of mortgage credit can lead to house price misalignments, owing to feedback loops between these variables, operating through the collateral channel.

Housing Valuation Metrics

(percent deviation from long-run average)

Citation: IMF Staff Country Reports 2014, 048; 10.5089/9781475590142.002.A001

Sources: OECD, IMF staff calculations

• Download Figure
• Download figure as PowerPoint slide

View Full Size

Housing Valuation Metrics

(percent deviation from long-run average)

Citation: IMF Staff Country Reports 2014, 048; 10.5089/9781475590142.002.A001

Sources: OECD, IMF staff calculations

• Download Figure
• Download figure as PowerPoint slide

Housing Valuation Metrics

(percent deviation from long-run average)

Citation: IMF Staff Country Reports 2014, 048; 10.5089/9781475590142.002.A001

Sources: OECD, IMF staff calculations

• Download Figure
• Download figure as PowerPoint slide

Box 1.

Cointegration Analysis

House price misalignments in Israel are estimated using an error correction model, which borrows elements from Andrew and Meen (2003), Glaeser, Gyuorko and Saiz (2008), and Caldera-Sánchez, and Johansson (2011). The specification of the model is as follows:

${\mathrm{\Delta }\mathrm{}\mathit{rph}}_t={\beta }_0+{\beta }_1{\mathit{\Delta rr}}_t+{\beta }_2{\mathit{\Delta rw}}_t+{\beta }_3{\mathit{\Delta my}}_{\mathit{t}}+{\mathit{\alpha ECT}}_{t-1}+{\epsilon }_t$

Δ is the first difference operator, rph_t denotes real house prices, rr_t the real rental price index, rw_t real wages per employee, and my_t the level of mortgage debt-to-GDP. Further, ECT is the error correction term, which corresponds to the residual of the long-run equation below:

${\mathit{rph}}_t={\gamma }_0+{\gamma }_1{\mathit{ps}}_t+y_2{\mathit{rr}}_t+{\gamma }_3{\mathit{rw}}_t+{\gamma }_4{\mathit{my}}_t+{\mathit{ECT}}_t$

ps_t denotes the population-to-housing stock variable, which is an indicator of supply-side constraints in the housing market. This variable was not included in the short-run dynamics equation, because it was not statistically significant.

The estimation results are reported in Table 1. All coefficients are statistically significant and have the expected signs. In addition, the error correction terms does not have a unit root, suggesting the existence of a cointegrating relationship between the variables in the long-run equation. This hypothesis was also supported by Johansen’s rank cointegration test.

Table 1.

Cointegration Analysis

*** p<0.01, ** p<0.05, * p<0.1Notes:

^1/The null hypothesis of a unit root process is rejected at the 5% significance level.

^2/The hypothesis of no cointegrating vectors is rejected at the 5% significance level, while the hypothesis of one cointegrating vector cannot be rejected.

Sources: Haver Analytics, Bank of Israel, and IMF staff estimates

Table 1.

Cointegration Analysis

Short-Run Dynamics
R-squared	0.29
Adj R-squared	0.23
Δ Real house prices	Coef.
Δ Real rent price	0.52 ***
Δ Real wage per employee	0.23 *
Δ Debt to GDP ratio	0.04 **
ECT (t-1)	−0.08 *
ECT Unit Root Tests
ADF	−3.28 ** 1/
Trace Statistic (max. rank 1)	39.72 ** 2/
Long-Run Equation
R-squared	0.79
Adj R-squared	0.77
Real house prices	Coef.
Population-stock ratio	1.47 **
Real rent price	1.11 ***
Real wage per employee	1.07 ***
Debt to GDP	0.08 **
Constant	−17.4 ***

*** p<0.01, ** p<0.05, * p<0.1Notes:

^1/The null hypothesis of a unit root process is rejected at the 5% significance level.

^2/The hypothesis of no cointegrating vectors is rejected at the 5% significance level, while the hypothesis of one cointegrating vector cannot be rejected.

Sources: Haver Analytics, Bank of Israel, and IMF staff estimates

View Table

Table 1.

Cointegration Analysis

Short-Run Dynamics
R-squared	0.29
Adj R-squared	0.23
Δ Real house prices	Coef.
Δ Real rent price	0.52 ***
Δ Real wage per employee	0.23 *
Δ Debt to GDP ratio	0.04 **
ECT (t-1)	−0.08 *
ECT Unit Root Tests
ADF	−3.28 ** 1/
Trace Statistic (max. rank 1)	39.72 ** 2/
Long-Run Equation
R-squared	0.79
Adj R-squared	0.77
Real house prices	Coef.
Population-stock ratio	1.47 **
Real rent price	1.11 ***
Real wage per employee	1.07 ***
Debt to GDP	0.08 **
Constant	−17.4 ***

*** p<0.01, ** p<0.05, * p<0.1Notes:

^1/The null hypothesis of a unit root process is rejected at the 5% significance level.

^2/The hypothesis of no cointegrating vectors is rejected at the 5% significance level, while the hypothesis of one cointegrating vector cannot be rejected.

Sources: Haver Analytics, Bank of Israel, and IMF staff estimates

The housing equilibrium value (ph^f) is derived from the long run equation, as follows:

${\mathit{rpg}}^f=-17.4+1.47{\mathit{ps}}^{\mathrm{avg}}+1.11{\mathit{rr}}^{\mathrm{avg}}+1.07{\mathit{rw}}^{\mathrm{avg}}+0.08{\mathit{my}}^{\mathrm{avg}}$

where the superscript ‘avg’ denotes long-run average (a proxy for the steady state value). Consequently, the misalignment in house prices is given by:

$m_t={\mathit{rph}}_t-{\mathit{rph}}^f$

If m_t>0, then house prices are above their equilibrium value. By this metric, house prices in Israel are currently 26 percent above their fundamental value. In addition, the estimates suggest that the speed of adjustment is somewhat slow, with 8 percent of the disequilibrium corrected every quarter. This implies that, other things being equal, the house price misalignment would be corrected in about 4 years.

House Price Misalignment

(Percent)

Citation: IMF Staff Country Reports 2014, 048; 10.5089/9781475590142.002.A001

Source: Bank of Israel, IMF staff estimates

• Download Figure
• Download figure as PowerPoint slide

View Full Size

House Price Misalignment

(Percent)

Citation: IMF Staff Country Reports 2014, 048; 10.5089/9781475590142.002.A001

Source: Bank of Israel, IMF staff estimates

• Download Figure
• Download figure as PowerPoint slide

House Price Misalignment

(Percent)

Citation: IMF Staff Country Reports 2014, 048; 10.5089/9781475590142.002.A001

Source: Bank of Israel, IMF staff estimates

• Download Figure
• Download figure as PowerPoint slide

3. Increased mortgage availability and supply-side rigidities are equally important drivers of the house price misalignment in Israel. Akin to the standard valuation metrics, cointegration analysis estimates indicate that house prices are about 25 percent above the equilibrium value (see Box1). In addition, around 50 percent of the current house price misalignment is explained by supply-side constraints, while another 50 percent is accounted for by above-average growth in mortgage debt⁵.

House Price Misalignment by Contributing Factors

(Percent)

Citation: IMF Staff Country Reports 2014, 048; 10.5089/9781475590142.002.A001

Sources: Haver Analytics, Bank of Israel, IMF staff estimates

• Download Figure
• Download figure as PowerPoint slide

View Full Size

House Price Misalignment by Contributing Factors

(Percent)

Citation: IMF Staff Country Reports 2014, 048; 10.5089/9781475590142.002.A001

Sources: Haver Analytics, Bank of Israel, IMF staff estimates

• Download Figure
• Download figure as PowerPoint slide

House Price Misalignment by Contributing Factors

(Percent)

Citation: IMF Staff Country Reports 2014, 048; 10.5089/9781475590142.002.A001

Sources: Haver Analytics, Bank of Israel, IMF staff estimates

• Download Figure
• Download figure as PowerPoint slide

4. After taking into account supply-side considerations, price-to-income and price-to rent ratios suggest that, by international standards, Israeli property prices are considerably elevated relative to their fundamental value. When housing demand is strong, house prices tend to rise faster in the presence of supply constraints. Thus, cross-country comparisons of housing valuations should internalize the effect of supply rigidities. A simple and sufficient way to address this issue is by expressing deviations of the price-to-income and price-to-rent ratios from their long-run average, in percentiles. This metric suggests that house prices in Israel are significantly above their equilibrium value, when compared to other advanced economies and to the country’s own history.

Sources: OECD, Bank of Israel, IMF staff calculations

• Download Figure
• Download figure as PowerPoint slide

View Full Size

Sources: OECD, Bank of Israel, IMF staff calculations

• Download Figure
• Download figure as PowerPoint slide

Sources: OECD, Bank of Israel, IMF staff calculations

• Download Figure
• Download figure as PowerPoint slide

Risks of a Correction in House Prices and Macroeconomic Consequences

5. With housing in short supply, Israel may be less exposed to a sharp correction in house prices. When the supply elasticity of housing is high, property price upswings are often accompanied by an overshooting in residential construction. In turn, during the downswing, the units overbuilt throughout the boom put additional downward pressure on prices.⁶ By contrast, in supply-constrained economies, household formation tends to outstrip the supply of houses continuously, raising structural affordability concerns. For this reason, in these countries property prices tend not to revert to their historical mean, relative to incomes; and house price corrections tend to be shorter and faster (see Table 2)⁷.

Table 2.

House Price Cycles in Advanced Economies

Sources: OECD; and IMF staff calculations.Notes: Amplitude is the average change in real house prices from trough-to-peak in upturns and from peak-to-trough in downturns. Supply constrained countries are Belgium, Israel, Italy, Netherlands, Switzerland and the UK. Countries that are not supply constrained are Canada, Germany, Denmark, Finland, France, Ireland, Japan, Norway, Sweden and the US.

Table 2.

House Price Cycles in Advanced Economies

	Upturns		Downturns
	Duration (qu arters)	Amplitude (percent)	Duration (quarters)	Amplitude (percent)
Supply constrained
Real house prices	59	80.4	22	−27.2
Price-to-Income ratio	47	51.4	18	−29.6
Price-to-Rent ratio	60	67.8	23	−34.7
Not supply constrained
Real house prices	41	61.1	32	−36.0
Price-to-Income ratio	27	36.0	34	−34.6
Price-to-Rent ratio	33	49.5	35	−42.7

Sources: OECD; and IMF staff calculations.Notes: Amplitude is the average change in real house prices from trough-to-peak in upturns and from peak-to-trough in downturns. Supply constrained countries are Belgium, Israel, Italy, Netherlands, Switzerland and the UK. Countries that are not supply constrained are Canada, Germany, Denmark, Finland, France, Ireland, Japan, Norway, Sweden and the US.

View Table

Table 2.

House Price Cycles in Advanced Economies

	Upturns		Downturns
	Duration (qu arters)	Amplitude (percent)	Duration (quarters)	Amplitude (percent)
Supply constrained
Real house prices	59	80.4	22	−27.2
Price-to-Income ratio	47	51.4	18	−29.6
Price-to-Rent ratio	60	67.8	23	−34.7
Not supply constrained
Real house prices	41	61.1	32	−36.0
Price-to-Income ratio	27	36.0	34	−34.6
Price-to-Rent ratio	33	49.5	35	−42.7

Sources: OECD; and IMF staff calculations.Notes: Amplitude is the average change in real house prices from trough-to-peak in upturns and from peak-to-trough in downturns. Supply constrained countries are Belgium, Israel, Italy, Netherlands, Switzerland and the UK. Countries that are not supply constrained are Canada, Germany, Denmark, Finland, France, Ireland, Japan, Norway, Sweden and the US.

6. Nevertheless, cross-country empirical analysis suggests that the risk of a housing downturn in Israel cannot be ignored. The probability of a housing bust within the next 5 years was estimated with a panel probit model. The outcome variable is a dummy that takes the value of 1 if real house prices fall by at least 10 percent within a five year window preceding the bust. The controls include variables which, as documented in the literature, are good predictors for house price adjustments: the current account balance, inflation, and residential investment-to-GDP and mortgage debt-to-GDP ratios.

Probability of Real House Prices Falling at Least 10% over the Next 5 Years

(Percent)

Citation: IMF Staff Country Reports 2014, 048; 10.5089/9781475590142.002.A001

Sources: OECD, Haver Analytics, Bank of Israel, IMF staff estimates

• Download Figure
• Download figure as PowerPoint slide

View Full Size

Probability of Real House Prices Falling at Least 10% over the Next 5 Years

(Percent)

Citation: IMF Staff Country Reports 2014, 048; 10.5089/9781475590142.002.A001

Sources: OECD, Haver Analytics, Bank of Israel, IMF staff estimates

• Download Figure
• Download figure as PowerPoint slide

Probability of Real House Prices Falling at Least 10% over the Next 5 Years

(Percent)

Citation: IMF Staff Country Reports 2014, 048; 10.5089/9781475590142.002.A001

Sources: OECD, Haver Analytics, Bank of Israel, IMF staff estimates

• Download Figure
• Download figure as PowerPoint slide

• The probability of a housing bust in Israel is about 20 percent. The factor contributing most to the likelihood of a downturn is the mortgage-to-GDP ratio, which has risen by 10 percentage points since 2007. However, due to the severity of supply-side constraints, the ratio of residential investment-to-GDP is low, which reduces the probability of an adjustment.

7. A house price correction in Israel could have important consequences for the real economy. Estimates from a VAR model for Israel suggest that a housing bust could hurt economic growth by weakening private consumption⁸: a one-standard deviation shock to real house prices -a 6.5 percentage point decline in annual house price inflation- would reduce consumption growth by 1.5 percentage points in the short-run and by nearly 3 percentage points over the long-run. To assess the impact of plausible future house price developments on the real economy, three alternative scenarios were considered: 1) a stable path, where prices adjust slowly to their equilibrium value; 2) a sharp and slow correction in house prices, akin to the adjustment that recently took place in the Netherlands; and 3) a sharp and quick fall in residential real estate prices, similar to the Israeli experience of the late 80’s⁹.

8. The macroeconomic effects of a correction in house prices depend on the speed of adjustment. A slow correction would allow the economy to escape a recessionary episode, but economic prospects would be weak for a prolonged period of time. By contrast, a rapid adjustment would lead the economy into recession, with consumption and output recovering two years after the shock.

Sources: Haver Analytics, Bank of Israel, Israel’s Central Bureaur of Statistics, and IMF staff estimates.

• Download Figure
• Download figure as PowerPoint slide

View Full Size

Sources: Haver Analytics, Bank of Israel, Israel’s Central Bureaur of Statistics, and IMF staff estimates.

• Download Figure
• Download figure as PowerPoint slide

Sources: Haver Analytics, Bank of Israel, Israel’s Central Bureaur of Statistics, and IMF staff estimates.

• Download Figure
• Download figure as PowerPoint slide

Role of Policies

9. The BoI has used monetary policy to support the real economy and macroprudential policies to alleviate housing sector related risks.

• At the height of the global financial crisis, the Bank of Israel cut the policy rate assertively, allowing the economy to recover quickly. As the economy and the housing market picked up, and inflationary pressures emerged, the BOI started to raise interest rates gradually, while tighter macroprudential measures were introduced to counteract the buoyant housing market (see Box 2).

• By the second half of 2011, global economic conditions became volatile, domestic economic growth slowed, and inflationary pressures receded. Notwithstanding these macroeconomic developments, house prices continued to rise fast. Consequently, the BoI began to lower the policy rate to support economic activity, and tightened its macroprudential policy stance in the pursuit of financial stability (see Box 2).

10. Macroprudential policies have until recently focused on indirect measures to restrain the buoyant mortgage and housing markets. Since late 2009, the BOI introduced supplementary provisioning and capital surcharges for mortgages (indirect measures), as well as restrictions to the adjustable rate component. More recently, direct measures, such as limits on loan-to-value and debt service-to-income, have been put in place for new housing loans.

11. While macroprudential policies have contributed to buttress the resilience of the financial system, their impact on mortgage and housing market activity has been narrower (see Box 3).

• Macro-prudential measures appear to have had an effect only over the six-month period following the intervention. Within this time horizon, direct measures have been more effective than indirect ones¹⁰. That is, restrictions on the size and risk of mortgages have been more successful than measures aimed at weakening banks’ incentives to lend.

• In the housing market, macroprudential policies have reduced somewhat the level of transactions, but there is no evidence that they have contributed to curb house price inflation. These results suggest that macroprudential policies may be helping to restrain speculative incentives. Intuitively, positive co-movements between house price inflation and the level of transactions signal the existence of speculative behavior, by indicating that market participants bargain over house price growth rather than property price levels (see Box 4). Although this result should be taken with caution, as the turnover data corresponds only to new built homes (which represent about 20 percent of the market), the finding is consistent with the evidence that tighter macroprudential policies have contributed to reduce the share of transactions undertaken by investors.

• The evidence also suggests that the origination of new housing loans has decelerated after the introduction of direct measures, while the impact on mortgage debt levels has not been significant. This may be partly explained by the slow-moving nature of the latter variable, which is a stock. More importantly, direct measures have not been given sufficient time to play out to assess their effectiveness, but early evidence is somewhat promising.

Box 3.

Assessing the Effectiveness of Macroprudential Policies

Following the approach proposed by Igan and Kan (2011), the effects of macroprudential regulation on house price dynamics, real estate activity and household leverage were examined by estimating the following equation

$V_t=\alpha +{\beta }^{\text{'}}X{\left(L\right)}_t+{\gamma }_d{D}_t^d+{\gamma }_i{D}_t^i+{\epsilon }_t$

where V_t denotes the variable of interest: house price inflation, turnover in the real estate market–expressed in terms of deviations from the long run average, the percent of transactions undertaken for investment purposes, and the growth rates of new mortgage originations and household mortgage debt.

To focus solely on the role of financial policies, changes to the monetary policy stance were controlled for by introducing a vector of controls, X_t. This vector includes indicators of economic activity and monetary conditions, such as the short-term interest rate, the mortgage spread, and the exchange rate. D_t ^d and D_t ⁱ are dummy variables that take the value of 1 in the months following the introduction of direct and indirect macroprudential measures, respectively. To assess the horizon over which financial policies have had an effect, the dummies were constructed for three and six-month windows. The results of the estimations are reported in Table 3 below.

Table 3:

Effects of Macro-prudential Policies on Housing and Mortgage Market Activity

(y/y percent change unless otherwise stated)

*, **, and *** denote significance at the 10, 5 and 10 percent level, respectively.Sources: Haver Analytics, Bank of Israel, and IMF staff estimates.

Table 3:

Effects of Macro-prudential Policies on Housing and Mortgage Market Activity

(y/y percent change unless otherwise stated)

		House Price Inflation	Turnover (demeaned)	Change in Investor Share	New Housing Loans	Mortgage Debt
Direct Measures
	3 months after	−0.07	−0.06	−0.21 **	−4.36	−0.01
	6 months after	−0.13	−0.12 *	−0.12 *	−5.05 *	0.04
Indirect Measures
	3 months after	0.03	−0.06	−0.02	0.66	0.09
	6 months after	0.07	0.04	−0.09 *	5.03 *	0.13
2009-2012 average		11.1	2.1	0.0	16.0	12.3

*, **, and *** denote significance at the 10, 5 and 10 percent level, respectively.Sources: Haver Analytics, Bank of Israel, and IMF staff estimates.

View Table

Table 3:

Effects of Macro-prudential Policies on Housing and Mortgage Market Activity

(y/y percent change unless otherwise stated)

		House Price Inflation	Turnover (demeaned)	Change in Investor Share	New Housing Loans	Mortgage Debt
Direct Measures
	3 months after	−0.07	−0.06	−0.21 **	−4.36	−0.01
	6 months after	−0.13	−0.12 *	−0.12 *	−5.05 *	0.04
Indirect Measures
	3 months after	0.03	−0.06	−0.02	0.66	0.09
	6 months after	0.07	0.04	−0.09 *	5.03 *	0.13
2009-2012 average		11.1	2.1	0.0	16.0	12.3

*, **, and *** denote significance at the 10, 5 and 10 percent level, respectively.Sources: Haver Analytics, Bank of Israel, and IMF staff estimates.

Box 4.

Deriving the Irrational Exuberance Component of House Price Inflation

A well-known pattern in housing markets is that prices changes and trading volumes, measured as the ratio of housing sales to stock (turnover), correlate with each other: relative to a falling market, trading activity is more intense when prices are rising. In Israel, the contemporaneous correlation between these variables in the market for new built homes is quite high (90 percent). This phenomenon can be explained by equity constraints and nominal loss aversion, or search frictions (see Poterba, 1984).

Peterson (2012) builds a theoretical model with search frictions in the housing market. He also develops an econometric approach to assess whether, in the presence of these frictions, house prices are driven by irrational exuberance.

Housing Market: Inflation and Trade

(y/y percent, 12ma percent deviation from mean 12ma)

Citation: IMF Staff Country Reports 2014, 048; 10.5089/9781475590142.002.A001

Sources: Haver Analytics, IMF staff calculations

• Download Figure
• Download figure as PowerPoint slide

View Full Size

Housing Market: Inflation and Trade

(y/y percent, 12ma percent deviation from mean 12ma)

Citation: IMF Staff Country Reports 2014, 048; 10.5089/9781475590142.002.A001

Sources: Haver Analytics, IMF staff calculations

• Download Figure
• Download figure as PowerPoint slide

Housing Market: Inflation and Trade

(y/y percent, 12ma percent deviation from mean 12ma)

Citation: IMF Staff Country Reports 2014, 048; 10.5089/9781475590142.002.A001

Sources: Haver Analytics, IMF staff calculations

• Download Figure
• Download figure as PowerPoint slide

The intuition underlying Peterson’s theory is that search frictions (or equity constraints) can cause prices to deviate from fundamentals due to demand being temporarily high (low), even though there has been no change to the underlying fundamentals. Since agents believe that the housing market is efficient, they think the prices of recent transactions reflect the fundamental value of houses. Therefore, buyers and sellers bargain over recent prices, or house price growth, rather than the level of property prices. To test this theory, Peterson formulates two models: one assumes that agents behave rationally (equation 1), and the other corresponds to a framework of irrational behavior (equation 2).

${\mathit{\Delta ph}}_t=\alpha +\beta \left({\mathit{ø}}_{\mathit{t}}\mathit{-}{\mathit{ø}}_{\mathit{t}\mathit{-}\mathit{4}}\right)+{\epsilon }_t\left(1\right)$

${\mathit{\Delta ph}}_t=\gamma +{\mathit{\Psi ø}}_t+{\epsilon }_t\left(2\right)$

Δph_t represents the annual growth rate of house prices, and ϕ_t denotes deviations from the long-run average of housing turnover.

Akin to what Peterson (2012) finds for the U.S., the results for Israel indicate that the rational model cannot explain house price growth, whereas the irrational exuberance model has a rather good fit. The estimates of the irrational model can be used to estimate a counterfactual that extracts the “irrational exuberance” component from house price developments. The chart to the right shows that when house prices started to accelerate in 2008, the role played by agents’ expectations was important. More recently, however, the speculative component of house price has become negligible.

House Prices

(January 2007=100)

Citation: IMF Staff Country Reports 2014, 048; 10.5089/9781475590142.002.A001

Sources: Haver Analytics, IMF staff estimates

• Download Figure
• Download figure as PowerPoint slide

View Full Size

House Prices

(January 2007=100)

Citation: IMF Staff Country Reports 2014, 048; 10.5089/9781475590142.002.A001

Sources: Haver Analytics, IMF staff estimates

• Download Figure
• Download figure as PowerPoint slide

House Prices

(January 2007=100)

Citation: IMF Staff Country Reports 2014, 048; 10.5089/9781475590142.002.A001

Sources: Haver Analytics, IMF staff estimates

• Download Figure
• Download figure as PowerPoint slide

Concluding Remarks and Policy Considerations

12. The analysis showed that house prices in Israel are well above their equilibrium value. The results from a cointegration model, as well as standard valuation metrics, suggest that house prices in Israel are about 25 percent higher than medium term fundamentals would suggest. This house price misalignment is largely explained by an increase in the availability of mortgage credit and supply-side rigidities. Furthermore, even after taking account of these supply-side considerations, by international standards house prices in Israel appear to be significantly elevated relative to their equilibrium value.

13. The scope for a correction in house prices is significant and could have important macroeconomic implications. Cross-country analysis suggests that the risk of a moderate to sharp correction in house price in Israel is around 20 percent. The factor contributing most to the likelihood such outcome is the mortgage debt-to-GDP ratio, which has increased rapidly over the past 5 years. By contrast, the relatively low level of construction has contributed to contain the probability of a housing downturn. If house prices drop sharply, economic growth would be undermined through weaker consumption.

14. The BoI has sought to balance growth concerns against the risks of macroeconomic imbalances in the housing market. Since 2008, the Bank has maintained an easing bias to support growth, notably in the tradable sector, which has been adversely impacted by the appreciation of the exchange rate (see External Competitiveness Annex). At the same time, macroprudential measures have been used to cool the housing market. Direct macroprudential policies, such as loan-to-value and debt service-to-income restrictions, have been more effective than measures aimed at weakening banks incentives to lend. Moreover, these policies have contributed to reduce financial stability risks and the volume of transactions in the housing market, although their impact so far in alleviating house price inflation has been more limited. If given more time to play out, such measures are expected to have a greater bearing in reducing risks of boom-bust dynamics in the housing market.

15. Additional house price increases would require assertive policy responses. If house prices continue to rise, direct macroprudential policies should be further tightened to curb mortgage lending and contain financial stability risks. In addition, to further reduce demand for housing, notably for speculative purposes, the property purchase tax for non-primary residencies could be increased. Finally, concerted efforts across the relevant agencies are needed to alleviate supply-side constraints and ensure a durable moderation of house price inflation.