Housing markets are large, and both house price and construction cycles are subject to sharp swings. The functioning of housing markets strongly influences countries’ exposure to economic crises and their capacity to recover from shocks. This chapter analyses the role that housing-related policies play in (i) mitigating or amplifying shocks and (ii) facilitating or hampering a recovery. It discusses how macro-prudential measures, rental regulation and taxation can contribute to greater economic resilience
Brick by Brick
Building Better Housing Policies
3. Enhancing Resilience
Abstract
Main policy lessons
There is a strong link between house price and construction cycles, and macroeconomic volatility. Macro-prudential, housing and tax policies provide complementary tools, which taken together can reduce the build-up of housing-related macroeconomic risks:
Loan-to-value ratios can be effective at containing the build-up of credit risk, and tighter loan-to-value ratios are linked with a lower risk of severe downturns. However, they are also associated with slower recoveries as they constrain borrowing.
Caps on debt-to-income ratios hold promising potential but have been used seldom, which limits possibilities to assess their effectiveness on an empirical basis.
More demanding capital requirements for mortgage-lending institutions are associated with more moderate output fluctuations and stronger recoveries after downturns.
Structural settings in the housing market also influence economic resilience:
Because it creates distortions in the housing market, tight rental market regulation is linked with higher crisis risk and more severe downturns. However, tight rental market regulation is also associated with less negative values of GDP at risk, suggesting consumption smoothing effects.
Higher taxation of housing (through property taxes or the treatment of housing under income tax) is linked with smoother housing cycles.
Main effects of tightening housing-related policies on resilience
Note: The table summarises the results of OECD empirical investigations in Cournède, Sakha and Ziemann, (2019[1]). Only significant results are displayed. Tightening means increases in policy indicators, except for LTV caps where a decrease of the policy value signifies a tightening. Green arrows show favourable outcomes, red ones unfavourable outcomes.
Recognise the role of housing for economic stability
Housing markets provide a sizeable contribution to economic activity. Fluctuations in house prices and residential investment can also be large, affecting the business cycle and amplifying shocks through balance sheet effects on households and lenders (Figure 3.1).1 In the boom phase, strong labour markets, economic growth and abundant credit supply feed strong demand, which pushes up real house prices. House price increases raise households’ collateral values and their net worth, which can, in turn, boost their consumption. Higher real house prices may lead to second-round effects as they may also create expectations of further price increases, feeding back into higher demand. Relaxation of lending standards and innovations in mortgage markets may further fuel house prices, a feedback loop that was at the centre of the global financial crisis.
Figure 3.1. House prices and business cycles are tightly linked
Annual real percentage change, OECD average
Source: OECD Economic Outlook database and OECD Analytical House Price database.
Housing market busts are characterised by the opposite developments. First, house price drops lower collateral values, which in turn increase the losses that lenders face in the event of a default with implications for financial stability. Second, household wealth and the prospects of the construction sector are negatively affected, which tends to decrease spending. This reduces overall economic activity, leading to deteriorating macroeconomic conditions and a weakening of the economic outlook and fiscal balances. Housing downturns seem to have particularly damaging effects for inclusion and productivity because of the role of homes as collateral in loans to small and new firms. Housing downturns are, therefore, often associated with severe recessions (Figure 3.2).
Figure 3.2. House price cycles have often been associated with severe recessions
Note: The light-blue areas represent the number of countries identified as being in a severe recession (from peak to trough). The global real house price index is measured in deviation from a moving average. The global real house price index is a weighted average across OECD countries.
Source: Hermansen and Röhn (2017[2]).
Address housing threats to macroeconomic resilience
Many factors affect housing demand. These include demography, including migration, changes in disposable income, house prices, interest rates or credit conditions. Demand shocks can stem from domestic factors but also international ones such as shifts in global capital flows, which can have large effects on some housing markets (Barcelona, Converse and Wong, 2020[3]). When housing demand changes, housing supply rigidities lead to either an increase in vacant homes (negative demand shock) or scarcity (positive demand shock), which result in housing investment and house price adjustments to clear housing markets. The extent to which the housing demand shock affects prices depends on the financial cycle (e.g. initial over or under-valuation of house prices, credit conditions), policies (inelastic supply due to zoning regulations, rent control, etc.) and cyclical or structural variables (e.g. construction costs, infrastructure).
Changes in house prices influence housing demand directly, but they also have indirect effects through the financial system. Movements in house prices have a strong impact on household balance sheets. Changes in household balance sheets affect, for instance, the number of non-performing loans and loan-to-value ratios. Changes in house prices also affect consumption depending on the size and institutional set-up of mortgage markets, such as the ease with which households can borrow against the value of their home.
Housing and the broader economy interact through various channels and policy interactions that affect the build-up of vulnerabilities, the severity of crises and the economy’s capacity to recover from them (Figure 3.3). An important distinction is between ex-ante (vulnerability to shocks) and ex-post (recovery from shocks) resilience.
Figure 3.3. Housing markets influence resilience through many channels
A conceptual framework
Source: Cournède, Sakha and Ziemann (2019[1]).
Both types of resilience can be gauged using a variety of indicators. For example, ex-ante resilience can be assessed by crisis probability, defined as the frequency of large downward deviations from trend, and GDP-at-risk, which measures the performance of the economy in bad times (i.e. GDP changes in the worst 5% periods). Ex-post resilience, on the other hand, can be gauged through measures of the severity of downturns (peak-to-trough changes in activity), the duration of business cycle downturns and the time needed to recover, that is, regain the pre-crisis level of output. On the basis of these indicators, cross-country evidence indeed suggests that where crisis probabilities are high, business cycle fluctuations are also high, and so is the strength of post-crisis recoveries (Figure 3.4).
Figure 3.4. Resilience differs across countries
Note: Crisis probability denotes the probability of experiencing a cumulative two percentage point decline in real GDP below its 1990-2017 trend over two consecutive quarters. GDP-at-Risk denotes the 5th percentile of the distribution of de-trended growth rate of real GDP over 1990-2017. Peaks and troughs of the deviation of GDP from its 1990-2017 trend are obtained using the Harding-Pagan (quarterly Bry-Boschan) business cycle dating procedure (Harding and Pagan, 2002[4]). Strength of recovery denotes growth over n quarters, n being the duration of the preceding downturn (period from peak to trough). Additional resilience indicators, covering business cycles (real GDP), real house price and housing investment cycles, can be found in Cournède, Sakha and Ziemann (2019[1]).
Source: Cournède, Sakha and Zieman (2019[1]).
Deploy macroprudential tools
The main objective of macro-prudential policy is to prevent financial threats to economic stability, by restraining the build-up of systemic risks by moderating credit and asset price cycles, while ensuring the presence of sufficient buffers in the financial system. A key advantage of macro-prudential regulation is that it can be tailored to risks of specific sectors, such as housing, or loan portfolios, such as mortgages. In contrast to interest rate hikes, macro-prudential tightening need not entail a generalised reduction of economic activity, limiting the potential costs of policy intervention.
The most common macro-prudential tools include:
Loan-to-value (LTV) caps, which limit the amount of loans below a share of the dwelling price (Figure 3.5). The experience of OECD countries shows that countries that apply tighter LTV caps face lower crisis risks (Box 1.8). However, more restrictive LTVs imply less vigorous recoveries. Besides, tightening LTV caps could in the short term involve a trade-off between financial stability and social-inclusion objectives, by making it more difficult for young households with low savings to purchase a home. In the medium-to-long term, however, lower house prices preserve the housing purchasing power of all households, including the young ones.
Debt-service-to-income ratios (DSTIs), which require households to pay no more than a certain proportion of their income to service their housing loans. In some countries, DSTIs are based on total rather than only housing debt servicing costs.
Loan-to-income ratios (LTIs), which limit the amount of debt to a certain fixed multiple of income, are less commonly used. They are equivalent to DSTIs for a given interest rate and repayment period but have the advantage of not becoming looser in times of booms when interest rates are low and banks offer more accommodative credit conditions.
Risk-weighted capital requirements, which set the minimum ratio of capital that banks must hold for housing loans depending on their riskiness. The strength of this requirement is determined by the combination of minimum capital ratios and risk weights. Regulatory frameworks that require banks to hold more capital against mortgage loans are linked with a reduced crisis probability and stronger recoveries from crises.
Macro-prudential policies have been used more intensively since the global financial crisis. In the aftermath of the crisis, both capital requirements and LTV caps have been mostly tightened. Since 2012, the balance is more uneven for LTV caps, as many countries loosened regulation following the euro area sovereign debt crisis. In the face of the COVID‑19 crisis, countries took measures to support mortgage borrowers and lenders (Box 1.7 and OECD (2020[5])). Furthermore, policies that keep mortgage borrowing in check are unlikely to entail costs in terms of foregone housing supply: from the high levels observed in OECD countries, further housing loan expansion seems to boost prices rather than construction (Kohl, 2020[6]).
Figure 3.5. Macroprudential measures often target mortgage loans
Panel A: Borrower-targeted measures 2019
Note: LTV denotes maximum loan-to-value ratios. DSTI stands for maximum debt-service-to-income ratios. If no cap exists, the respective value is set to 100. Blank points or no bars signal that no information is available for the respective country. Average risk-weights are obtained as the unweighted average of risk weights for mortgage loans with LTVs ranging from 50 to 130 (since Basel II, risk weights can differ by LTV).
Source: IMF Integrated Macroprudential Policy (iMapp) Database; ESRB Macroprudential Database; OECD 2019 Questionnaire of Affordable and Social Housing; Cournède, Sakha and Ziemann (2019), “Empirical links between housing markets and economic resilience”.
Align structural housing-related policies with the goal of economic resilience
Rental market regulation influences housing cycles
Rent controls and landlord-tenant rules have been devised for a variety of social and economic reasons, such as to provide affordable accommodation by limiting rent increases, and to balance the landlord-tenant bargaining power. However, excessively tight regulations can discourage investment in new dwellings and maintenance of the existing rental housing stock, and hamper the development of the rental market. This can lead to housing shortages, exacerbate speculative housing price bubbles and increase household debt, which poses significant vulnerabilities for macroeconomic stability and economic growth (Caldera and Johansson, 2013[7]; Cavalleri, Cournède and Özsöğüt, 2019[8]; Hermansen and Röhn, 2017[2]).
The tightness of rental market regulations varies considerably within the OECD area (Figure 3.6). Evidence suggests that tighter rental market regulations are associated with higher crisis risk and deeper business cycle downturns (Box 3.1), because they distort the adjustment of housing supply to demand, which can exacerbate the accumulation of imbalances. GDP fluctuations (measured by GDP at risk) however tend to be milder in countries with strong tenant protection as a result of the protection that such regulations provide to vulnerable tenants against the consequences of income shocks.
Figure 3.6. Rental market regulations and housing taxation vary considerably across countries
Note: METR stands for “marginal effective tax rate” for owner-occupied, debt-financed housing investments. The indicator is a preliminary 2019 update of (OECD, 2018), “Taxation of Household Savings”. The final version will be released as Brys et al. (2021[9]). The rental regulation index comprises rent control and tenure security items of the 2019 OECD Questionnaire on Affordable and Social Housing and ranges from 0 (no restrictions) to 1 (all types of restrictions).
Source: Cournède, De Pace and Ziemann (2020[10]).
Box 3.1. Empirical evidence of the influence of housing policies on economic resilience
Links between policies and ex-ante resilience
Macroprudential policies lower severe-downturn probabilities, a key component of ex-ante resilience. Probit regressions show that severe-downturn probabilities are lower in countries that implement tighter macroprudential measures (loan-to-value caps and capital requirements). This result is in line with the earlier economic literature, which finds a moderating effect of macroprudential measures on the build-up of housing booms and credit risk. This reduced probability of crises is not matched by a reduction in GDP at risk, the lowest 5% of the distribution of growth outcomes, as house price booms and credit bubbles typically build up slowly and burst with a protracted downturn (rather than frequent occurrences of very weak GDP growth).
A significant relationship is also found between rent regulations and ex ante resilience indicators. Quantile regressions indicate that tighter rent regulations are associated with a lower dispersion of growth outcomes, suggesting that they smooth household consumption. However, rent regulations are also linked with a greater risk of severe downturns. This link is consistent with the view that overly tight rent regulations (by creating disincentives to rent out existing properties or build rentals) create a bias towards home ownership, which in turn can result in excessive mortgage borrowing. By reducing the expected rental value of dwellings, they also contribute to a lower responsiveness of supply, exacerbating the risk of housing market boom-bust cycles.
Links between policies and ex-post resilience
Ex-post resilience measures and indicators of the housing cycle have been related to macroprudential and housing policy indicators using pooled regressions. The results indicate that:
Tighter LTV caps are associated with shorter booms and milder downturns but more sluggish recoveries.
Tighter capital requirements are also linked with shallower cycles, but by contrast with LTV caps, they appear to be associated with stronger recoveries.
More stringent rental market regulations are associated with shorter and less pronounced booms but also shallower downturns, which would suggest that they might provide a certain degree of smoothing through the protection they offer.
Higher levels of property taxation are also linked with a more limited amplitude of business cycles, also pointing to a smoothing role.
Additional research have investigated the causality that runs from policy changes to real, financial and housing variables. The impact of a policy change on real, financial and housing variables has been assessed using propensity-score matching techniques, which enable comparing the countries where policy changes occurred to similar countries that left policies unchanged. The idea is to compare two episodes that are as similar as possible, one with a policy change (treatment) and one without (control). The policy change can then be considered as exogenous, so that observed differences in outcomes between the treatment and control groups can be attributed to the policy change.
The empirical analysis suggests that tightening LTV caps curbs credit to households and slows real house prices. Furthermore, there appears to be no sizeable impact on private consumption or aggregate output. These two sets of results point to stabilisation benefits of LTV caps with no significant macroeconomic costs.
Figure 3.7. Tightening LTV caps usually reduces credit and house price growth
Note: The treatment group consists of episodes where countries tightened their LTV caps at time=0. The control group comprises episodes where countries did not implement such a policy change although their conditions were otherwise similar. This treatment group has been determined by propensity matching techniques using a probit model with real and financial variables as covariates. The lines show averages for each group.
Source: Cournède, Sakha and Ziemann (2019[1]).
Property taxation can also have an effect on housing market dynamics
Housing markets are affected differently by different tax instruments. For example, stamp duties can slow down house price rises by reducing the expected returns on speculative house purchases. Higher stamp duties therefore reduce housing transaction volumes, but they also raise housing transaction costs and can lead to a lock-in effect, which poses an obstacle to reallocation in the labour market (see Chapter 6). By contrast, recurring taxes on property are broadly neutral with respect to the cyclical behaviour of housing markets and economic resilience. Their main effect is to reduce the size of the housing market, by making housing more expensive. As a result, it is important to gauge the combined effect of all tax instruments, rather than that of individual instruments, through marginal effective tax rates (METR) on owner-occupied and rental housing (Figure 3.6).
METRs are derived as the difference between the pre and post-tax rates of return of a marginal investment divided by the cost of capital of that investment where the post-tax real rate is the minimum rate of return necessary to make the investment worthwhile (OECD, 2018[11]). Overall property taxation (measured by the METRs) generally smooths business cycles: higher METRs are associated with a reduced severity of downturns (Box 3.1).
Supply responsiveness also has implications for economic resilience
The responsiveness of housing supply to changes in demand is influenced by policies, such as rental market regulations and land-use, which influence the dynamics of housing cycles. Indeed, countries where housing supply responds more strongly to demand have higher volatility of homebuilding (Cavalleri, Cournède and Özsöğüt, 2019[8]).
References
[3] Barcelona, W., N. Converse and A. Wong (2020), US Housing as a Global Safe Asset: Evidence from China Shocks, https://www.banque-france.fr/sites/default/files/session1_c_presentation_converse.pdf.
[9] Brys, B. et al. (2021), Effective Taxation of Residential Property, forthcoming.
[7] Caldera, A. and Å. Johansson (2013), “The price responsiveness of housing supply in OECD countries”, Journal of Housing Economics, Vol. 22/3, pp. 231-249, http://dx.doi.org/10.1016/J.JHE.2013.05.002.
[8] Cavalleri, M., B. Cournède and E. Özsöğüt (2019), “How responsive are housing markets in the OECD? National level estimates”, OECD Economics Department Working Papers, No. 1589, OECD Publishing, Paris, https://dx.doi.org/10.1787/4777e29a-en.
[10] Cournède, B., F. De Pace and V. Ziemann (2020), The Future of Housing: Policy Scenarios.
[1] Cournède, B., S. Sakha and V. Ziemann (2019), “Empirical links between housing markets and economic resilience”, OECD Economics Department Working Papers, No. 1562, OECD Publishing, Paris, https://dx.doi.org/10.1787/aa029083-en.
[4] Harding, D. and A. Pagan (2002), “Dissecting the cycle: a methodological investigation”, Journal of Monetary Economics, Vol. 49/2, pp. 365-381, http://dx.doi.org/10.1016/S0304-3932(01)00108-8.
[2] Hermansen, M. and O. Röhn (2017), “Economic resilience: The usefulness of early warning indicators in OECD countries”, OECD Journal: Economic Studies, Vol. 2016/1, https://www.oecd-ilibrary.org/docserver/eco_studies-2016-5jg2ppjrd6r3.pdf?expires=1547983022&id=id&accname=ocid84004878&checksum=3A3C6140FEC597236FBC3673C21D5C42.
[6] Kohl, S. (2020), “Too much mortgage debt? The effect of housing financialization on housing supply and residential capital formation”, Socio-Economic Review, http://dx.doi.org/10.1093/ser/mwaa030.
[5] OECD (2020), Housing Amid COVID-19: Policy Responses and Challenges, https://www.oecd.org/coronavirus/policy-responses/housing-amid-covid-19-policy-responses-and-challenges-cfdc08a8/.
[11] OECD (2018), Taxation of Household Savings, OECD Tax Policy Studies, No. 25, OECD Publishing, Paris, https://dx.doi.org/10.1787/9789264289536-en.
Note
← 1. This chapter provides policy insights on the effect of housing on economic stability based on two background papers, which also provide detailed bibliographic references (Cournède, Sakha and Ziemann, 2019[43]; Cavalleri, Cournède and Ziemann, 2019[2]).