Introduction

Few macroprudential policies intervene as directly in household choices as those regulating mortgage borrowing. Limits on loan-to-value (LTV) and loan-to-income ratios, debt-service burdens, and amortization requirements (“mortgage constraints”) are used in most advanced economies. Authorities justify them primarily by the need to restrain credit growth and household leverage, with financial stability as the ultimate objective. A natural evaluation of such policies therefore asks whether household credit fell after a rule was introduced. Yet, a lower credit volume is an incomplete measure of policy success for three reasons. First, it says nothing about which households gave up borrowing and what the lost borrowing capacity was worth to them. Second, it does not capture the market responses that redistribute gains and losses across buyers, owners, renters, and lenders. Third, it does not reveal whether financial vulnerability was reduced or merely relocated to other borrowers, lenders, or assets.

The evidence needed to address these shortcomings is spread across several strands of the literature. Household-level studies identify who is constrained and how those households adjust. Studies of housing markets and financial intermediaries examine changes in prices, lending conditions, and the allocation of credit. Cross-country studies estimate aggregate policy effects on credit. And a separate literature relates credit growth, house prices, and debt servicing to subsequent financial distress. These estimates are rarely interpreted together.

In this article, we provide a common framework that organizes the literature on mortgage constraints. We decompose the effect of a mortgage tightening into three components: the direct effect on households whose effective borrowing capacity falls, the incidence of the equilibrium changes that follow, and the corrective benefit from lower financial vulnerability. We use this framework to review and synthesize the literature. Along the way, we classify the objectives authorities state when introducing or adjusting mortgage constraints, compare household-level and aggregate credit estimates, and compare the timing of policy changes relative to crisis risk. We end with a discussion of where the gaps in the literature are most pressing.

On the direct effect, the household-level evidence gives a consistent picture of who is affected and how they respond. LTV limits fall mainly on households with little wealth or housing equity, while income-based limits fall on households whose desired borrowing is high relative to income (DeFusco et al. 2020; Peydró et al. 2026). Young and first-time buyers are especially exposed under both (Kinghan and McCann 2019; Aastveit et al. 2022). Some affected households still buy housing, but typically with a smaller mortgage, more equity from savings or family, or a cheaper property (Van Bekkum et al. 2024; Aastveit et al. 2022; Tzur-Ilan 2023). On the other hand, an often sizable share of potential buyers delays the purchase or remains a renter (Aastveit et al. 2022; Eerola et al. 2023). The same decline in borrowing can therefore reflect costly exclusion from homeownership or relatively easy substitution toward savings, family support, cheaper housing, or renting. The quantitative credit response alone is thus a poor guide to the private burden of the policy.

The direct credit effect is also small relative to aggregate estimates. Applying household-level treatment effects to the exposed share of new originations and scaling by the outstanding debt stock implies a reduction in annual household credit growth of about 0.16–0.27 percentage points for a typical tightening, mostly through fewer mortgage originations. Estimates from macro-level studies, however, are between 5–12 times larger (Cerutti et al. 2017; Akinci and Olmstead-Rumsey 2018). Equilibrium amplification is one possible explanation for the difference. Household-level designs compare more and less exposed households within a market and therefore difference out changes in prices and credit conditions common to all households. Macro estimates retain these market-wide responses, at the cost of potentially weaker identification. The gap is consistent with amplification, although differences in samples and policy instruments could also contribute. It is also worth highlighting that there is substantial variation in the reforms between studies. For instance, some papers study reforms that make small adjustments to mortgage constraints, whereas other studies investigate the effect of the initial activation of policies or blunter measures of policy change.

The incidence of the potentially large general equilibrium effects depends on which channel is the driving force. The evidence here is thinner than for the direct credit effects, but points to three margins. First, house-price effects of mortgage regulation are small and imprecise in cross-country studies (Araujo et al. 2020) but larger in within-country studies (Acharya et al. 2022; Peydró et al. 2024; Chi et al. 2024). Lower house prices redistribute gains from net sellers to net buyers (Fagereng et al. 2025), suggesting that mortgage constraints have distributional effects through house prices. Lower house prices also tighten LTV constraints further through lower collateral values, which affects credit-constrained households. Second, rents appear to rise where mortgage constraints keep households out of ownership, so renters may bear part of the cost (Gete and Reher 2018; Castellanos et al. 2024). Third, lenders reallocate credit toward less exposed borrowers, regions, and asset classes (Acharya et al. 2022; Peydró et al. 2024; Cizel et al. 2019), which lowers financial vulnerability only if the new lending is safer than the lending it replaces. Few studies, however, quantify overall winners and losers, and the evidence on the equilibrium incidence of mortgage constraints remains largely qualitative.

The final component is the corrective benefit with respect to financial stability, which is the main rationale for mortgage constraints. Direct evidence from aggregate shocks are scarce. One exception is Peydró et al. (2024), who document that UK areas more exposed to the loan-to-income limit, and therefore less leveraged, had more resilient house-price growth and fewer mortgage defaults after the Brexit referendum. Evidence on household consumption is more mixed, since lower leverage comes with smaller liquid buffers that can leave households less able to smooth consumption after job loss (Aastveit et al. 2026). A few papers associated tightening of mortgage constraints with subsequently fewer defaults (Van Bekkum et al. 2024; DeFusco et al. 2020). At the aggregate level, the “GDP-at-risk” literature finds that macroprudential tightening improves the lower tail of the future GDP distribution (Franta and Gambacorta 2020; Brandão-Marques et al. 2020), although the specific contribution of mortgage constraints is less firmly established (Fernández-Gallardo et al. 2023).

A much larger literature links credit to crises and can be informative for thinking about whether the corrective effects of mortgage constraints are high or low. Long historical panels show that credit growth, and in the postwar period mortgage credit in particular, predicts financial crises (Schularick and Taylor 2012; Jordà et al. 2016), and that the risk is especially elevated when rapid household credit growth coincides with rapid house-price growth (Greenwood et al. 2022). Financial crises, in turn, are often followed by persistent, and sometimes deep, recessions (Jordà et al. 2013; Cerra and Saxena 2008; Romer and Romer 2017). An implication of this literature is that the corrective benefit of mortgage constraints is likely to be largest when credit and house prices grow rapidly together. We place 154 mortgage-constraint tightening episodes on a “crisis map” that sorts country-years by prior household credit growth and house-price growth. Only ten episodes fall in the region where both are growing exceptionally fast. This could mean that authorities acted preemptively, before credit and house prices reached the high-risk region, or that the policies were used when historical risk was low and as such had small corrective benefits. Distinguishing between these would require a counterfactual path for credit, house prices, and lending standards.

We end by summarizing our views on directions for future research in the context of our framework. In terms of direct effects, the main gap concerns the households that are kept out of the market. Following them over time would show whether mortgage constraints delay homeownership or exclude some households permanently, as well as shed light on other adjustments that might be informative about the overall burden of the policies. In terms of equilibrium incidence, direct and market-level responses should be estimated for the same reform, so that amplification and its channels can be estimated and understood in the same setting, and studies should trace whether credit removed from the regulated segment reappears elsewhere. In terms of the corrective benefits of mortgage constraints, it would be fruitful to measure it more directly. Following exposed and non-exposed households through downturns, for instance, would show whether lower leverage dampens the fall in spending and whether this protects the incomes of other households. Tightening episodes should also be evaluated by their effect on the subsequent path of credit and house prices, not only by the economy’s position when the policy was introduced. Finally, many countries implement multiple borrower-based measures, in addition to bank-based constraints aimed at targeting the same external costs. A holistic approach to evaluating and calibrating macroprudential policies would be a useful and potentially impactful step for policy.

The article proceeds as follows. Section 2 documents the use, objectives, and design of mortgage constraints. Section 3 introduces our organizing framework. Section 4 reviews policy exposure, implementation, and household adjustments. Section 5 connects these responses to aggregate effects and equilibrium incidence. Section 6 evaluates the evidence on corrective benefits. Section 7 sets out directions for future research, and Section 8 concludes.