House on Norwegian currency
VoxEU Column Labour Markets

The price of leverage: Learning from the effect of loan-to-value constraints on job search and wages

Macroprudential policies restricting household leverage may affect job search and subsequently wages via opposing channels through a debt-overhang and a liquidity-constraints channel. This column describes how the introduction of a loan-to-value ratio restriction in Norway can be used to assess which channel dominates empirically. It finds that job-seeking workers, when forced to have lower leverage, obtain higher wages in their new jobs. A reduction in leverage enables workers to have longer and broader job search and to find different occupations and jobs in firms with higher wage premia. These unintended effects on labour markets are important in an overall assessment of the welfare implications of macroprudential policies.

The Global Crisis made the potential negative consequences of household leverage on the economy very salient. Research motivated by the steep downturn of 2007-2008 has shown how household leverage can induce a housing boom (Mian and Sufi 2011, Adelino et al. 2016), weaken financial stability (Schularick and Taylor 2012), and lower economic growth (Mian et al. 2017). Such negative consequences led many countries to implement macroprudential policies that could effectively curb credit booms and strengthen financial stability (Cerutti et al. 2017, DeFusco et al. 2019, Sandri et al. 2020, Gelos et al. 2023). 

Policies that restrict household leverage can, however, also bring about costs for households, for instance by forcing them to reduce their housing consumption (Tzur-Ilan 2023) or deplete liquid savings (Aastveit et al. 2020). 

In addition to these well-documented channels, macroprudential borrowing restrictions can potentially affect the economy in other ways, such as via its influence on job search. On the one hand, these policies can reduce wages by mitigating the debt-overhang problem, i.e. when workers’ incentives to work are weakened due to higher leverage requiring a larger portion of wages for debt service. A macroprudential borrowing constraint then increases labour supply. On the other hand, a borrowing restriction can increase wages because it relaxes workers’ liquidity constraints and, in turn, can enable them to have a longer and broader job search. This could lead to better jobs with higher wages. The negative correlation between household leverage and wages, as seen in Figure 1, suggests that the latter channel may be more important. Yet, this negative correlation could be driven by another factor, instead of leverage itself, or be a consequence of reverse causality.

Figure 1 Correlation between household leverage and wages

Figure 1 Correlation between household leverage and wages

In a recent paper (Kabas and Roszbach 2023), we investigate the causal effect of debt on job search and wages. To do so, we exploit the introduction of a loan-to-value (LTV) ratio restriction in Norway. Norway experienced strong growth both in home prices and household debt levels, posing a threat to financial stability. Therefore, the Financial Supervisory Authority of Norway introduced an LTV constraint that imposes an 85% cap on the size of mortgages relative to home values.

We use the exogenous restriction of household leverage produced by the constraint in a difference-in-differences setting. This enables us to compare the job search and subsequent wages of workers with restricted leverage to that of unrestricted workers. Identifying the restricted workers requires a special effort since the LTV constraint applies to all new homebuyers. For this purpose, we use a random forest algorithm in which the homebuyers before the restriction function as correctly-labelled observations. The algorithm will classify workers as treated if homebuyers with similar characteristics had LTV ratios above the cap level before the constraint was introduced. Workers who resemble homebuyers who had LTV’s below 85% before the policy was implemented are classified as controls.

We deal with the possibility that workers can influence the start of their job search by limiting our analysis to workers who were displaced in a mass layoff. Such workers’ job search will not be triggered by their individual characteristics since they lost their jobs due to exogenous conditions. Furthermore, to avoid the accumulation of unobserved home equity prior to the layoff, we restrict our sample to those workers who bought a house no more than 12 months before their displacement. Our difference-in-differences setting thus compares those displaced workers who had recently bought a house and are likely affected by the LTV restriction (treatment group) to those unaffected by the restriction (control group).

Our results reveal that job-seeking workers who have lower leverage due to the restriction have higher wages in their new jobs. In particular, as shown in Figure 2, a 25% decline in workers' debt-to-income (DTI) ratio leads to a 3.3 percentage point increase in wages compared to other displaced workers, who, on average, experience a 7.4 percentage point fall in wages. We show that wage improvement is not driven by changes in entry into the housing market. 

Figure 2 Household leverage and wages of job-seeking workers

Figure 2 Household leverage and wages of job-seeking workers

When we look into the underlying mechanism, we find that workers’ wage improvement is driven by the mitigation of constraints on job search. Workers are able to extend their job search by 2.5 months, suggesting that lower leverage reduces the pressure on displaced workers to quickly find or accept a new job. At the same time, workers broaden the scope of their job search leading to a 20% rise in the frequency of occupational and industry switches. Displaced workers also find matches with firms that pay higher wage premiums, which explains 20% of the relative rise in wages.

We further document how heterogeneity across people influences the effect on wages. Workers who are below median age, have a shorter job tenure with their previous employer, or have higher education, drive the improvement in wages. This is consistent with the notion that it is easier for workers who are younger or have higher education to invest in the human capital required for a different occupation or industry. Longer job tenure with the same firm, on the other hand, makes human capital more firm-specific and limits the value of improved job search. Overall, the wage improvement is larger for low-income workers. 

Finally, we find that workers who are forced to restrict their leverage also enjoy lower wage volatility, meaning that the rise in wages is not attributable to taking jobs with high wages but high discontinuation risk.

Our findings mean that macroprudential policies aimed at limiting household leverage generate likely unintended positive effects on labour market outcomes and the real economy. This will matter for policymakers’ cost-benefit analyses of macroprudential policy. The effect of household leverage on the labour market is potentially also important as high household leverage has been a common characteristic of recent recessions and pervasive in many countries.

References

Aastveit, K A, R Juelsrud and E Getz Wold (2020), “Mortgage Regulation and Financial Vulnerability at the Household Level”, Norges Bank Working Paper 2020/6.

Adelino, M, A Schoar and F Severino (2016), “Loan Originations and Defaults in the Mortgage Crisis: The role of the Middle Class”, The Review of Financial Studies 29(7): 1635–1670.

Cerutti, E, S Claessens and L Laeven (2017), “The Use and Effectiveness of Macroprudential Policies: New Evidence”, Journal of Financial Stability 28: 203–224.

DeFusco, A A, S Johnson and J Mondragon (2019), “Regulating Household Leverage”, The Review of Economic Studies 87: 914-958.

Gelos, G, M S M Pería, E Nier and F Valencia (2023), “Macroprudential Policies Are Effective, With Limited Side Effects – But Open Questions Remain”, VoxEU.org, 14 April.

Kabas, G and K Roszbach (2023), “The Price of Leverage: Learning from the Effect of LTV Constraints on Job Search and Wages”, Norges Bank Working Paper 21/14.

Mian, A and A Sufi (2011), “House prices, home equity-based borrowing, and the U.S. household leverage crisis”, American Economic Review 101(5): 2132–56.

Mian, A, A Sufi and E Verner (2017), “Household debt and business cycles worldwide”, The Quarterly Journal of Economics 132(4): 1755–1817.

Sandri, D, F Grigoli, N-J Hansen and K Bergant (2020), “Macroprudential Regulation Can Effectively Dampen Global Financial Shocks”, VoxEU.org, 12 August.

Schularick, M and A M Taylor (2012), “Credit booms gone bust: Monetary policy, leverage cycles, and financial crises, 1870-2008”, American Economic Review 102(2): 1029–61.

Tzur-Ilan, N (2023), “The real consequences of LTV limits on housing choices”, Review of Financial Studies, forthcoming.