Debt Service and Default: Calibrating Macroprudential Policy Using Micro Data

IMF Working Papers, August 22, 2019

Source details

Canonical URL: Debt Service and Default: Calibrating Macroprudential Policy Using Micro Data

The relationship between debt-service-to-income (DSTI) and probability of default (PD) is non-linear, with PD responding to increases in DSTI only after a certain threshold.
Consumer loan defaults occur at lower levels of DSTI compared to mortgages.
A counterfactual analysis indicates that had a DSTI limit been in place for all loans in the sample, the probability of default (PD) would have been lower by 23 percent.
The empirical results support the National Bank of Romania’s regulation limiting household DSTI at origination to 40 percent for new mortgages and consumer loans.

Empirical microdata on household indebtedness and loan performance in Romania underpin the calibration of a cap on household indebtedness expressed as a DSTI ratio.
The analysis identifies a threshold effect in the DSTI–PD relationship rather than a constant marginal effect across the DSTI distribution.
Distinct default dynamics are documented across loan types, with consumer loans exhibiting default at lower DSTI levels relative to mortgages.

Implementing a cap on DSTI at origination can materially reduce borrower default risk.
Calibrate DSTI limits recognizing non-linear risk increases: thresholds matter more than uniform incremental DSTI changes.
Differentiate DSTI-based macroprudential measures by loan type, given that consumer loans show higher default sensitivity at lower DSTI levels than mortgages.
The National Bank of Romania’s 40 percent DSTI cap for new mortgages and consumer loans is empirically supported by the analysis.

DSTI cap adopted at origination: 40 percent (for new mortgages and consumer loans).
Estimated reduction in probability of default under counterfactual DSTI cap: 23 percent.
Working Paper length: 45 pages.
Working Paper series: Working Paper No. 2019/182, Issue 182, Volume 2019.

Source: IMF Working Paper "Debt Service and Default: Calibrating Macroprudential Policy Using Micro Data" (August 22, 2019).