The unintended effects of relaxing targeted credit constraints in the housing market : empirical evidence from Hong Kong

Abstract

Housing assets and mortgage loans are critical to household financial well-being and macroeconomic stability. The central role of credit conditions in the mortgage debt-service channel makes government interventions in the credit market an appealing macroprudential instrument. Despite the potentially salient policy implications, micro-level empirical evidence on the role of credit conditions in explaining housing dynamics is scarce and mixed. Furthermore, “targeted” housing policies that intervene in a specific market segment have attracted increasing academic attention. This study fills the gap in related literature by investigating the causal effects of relaxing credit constraints targeted at a particular market segment on housing dynamics. Exploiting plausibly exogenous variation in down payment constraints caused by mortgage insurance program (MIP) amendments in Hong Kong in 2019 as a quasi-experiment, we evaluate the unintended market responses and strategic behaviors using difference-in-differences (DID) and bunching regression.

We present the main findings based on property-level transaction data in the Hong Kong residential market. First, relaxing down payment constraints in the targeted market segments increases trading volumes and prices in this specific segment. Second, we provide empirical evidence that targeted policies have unexpected and far-reaching spillover effects on neighboring segments triggered by derived demand for moving up housing ladders. Specifically, the derived demand from the housing ladder transition can be attributed to (i) a relaxation in ex-post financing constraints due to credit-driven housing appreciation and (ii) an increase in the ability to sell the current homes of mismatched homeowners attempting to move. Our findings suggest that housing stimulus programs may be justified by their welfare-enhancing role of general equilibrium effects, particularly in a cold market where the equilibrium is far from efficient due to illiquidity and frictions.

Third, we exploit the exogenous policy-induced sharp discontinuity in required down payments around the MIP eligibility threshold to analyze the local behavioral responses of households. We use the bunching regression to find a substantial excess “bunching” mass in the density of home prices around the threshold and a “missing” mass just above the threshold. This implies that some credit-constrained households with optimal mortgage demands slightly above the qualifying MIP threshold tend to reduce their loan amount to the threshold level. Additionally, we find that households strategically underreport sales prices to qualify for MIP through tacit collusion with sellers and agents. Our analysis suggests that policymakers should account for the strategic behaviors of market participants when designing macroprudential policies.

We make several contributions to the existing literature. First, we provide reduced-form evidence on the role of credit conditions in explaining housing dynamics at the micro-level. Second, we add to the literature on cross-segment spillover effects triggered by trading up activities of mismatched homeowners or investors, which inspires further research on general equilibrium effects in segmented markets. Third, we provide empirical evidence on the bunching and collusive underreporting behaviors around the MIP eligibility threshold, which potentially induce unintended consequences and distortions in price signals. Our analysis sheds light on the significance of the behavioral responses of market participants for policymakers in designing macroprudential policies.