Tackling escalating credit card debt – lessons from Singapore
Agarwal, Qian and Zou (2024) examine the impact of a 2015 credit suspension policy implemented by the Monetary Authority of Singapore (MAS). Their research delves into how this policy effectively mitigated debt accumulation among heavily indebted individuals while simultaneously fostering greater financial prudence across the broader population.
Wong Wei Chen
19 December 2024
In an economy without debt, great plans go awry.
Imagine an entrepreneur with a groundbreaking idea. But he has no money, and neither do his friends and relatives. If equity funding is not forthcoming, and debt funding is non-existent, an idea basically remains what it is: a figment of the imagination, albeit a febrile one.
Yet, debt has its dark side. Carried to reckless extremes, debt could lead to defaults, and the Global Financial Crisis of 2008 is a stark reminder of reckless borrowing – and reckless lending as well.
Spiralling household debt has often been a precursor to economic slowdowns and downturns. On the one hand, highly indebted households may crimp expenditure and reduce consumption, which in turn softens the economy. On the other hand, such households are more vulnerable to shocks, such as job losses or income reductions, and for those who have gone too far into debt – default awaits. A double whammy of negative sentiment on the demand side coupled with widespread inability to repay loans could very well drive an economy into a recession.
Keeping an eye on household debt is therefore an important task. In their working paper “Liquidity constraint threat and household indebtedness”, Agarwal, Qian and Zou (2024) investigate the impact of the Monetary Authority of Singapore (MAS)’s credit suspension policy on heavily indebted credit card users who had persistently failed to repay their debt over an extended period.
Background
Commencing from June 2015, financial institutions were prohibited by MAS from providing new unsecured credit lines to individuals with significant debt burdens. Affected persons were defined as borrowers who had accumulated unsecured debt that was 24 times their monthly income for at least three consecutive months i.e. “revolvers” who made minimum or partial payments, and carried outstanding balance on their credit cards over to the next period. Financial institutions were also barred from increasing credit limits for these individuals.
Revolvers who do not promptly repay their debts in full have to contend with a snowballing effect, as they continue to pay interest on a growing debt balance. When a significant portion of the population carries high levels of credit card debt even as their ability for full repayment diminishes, the financial system and economy may eventually take a hit.
Identification strategy
Since the MAS policy targeted unsecured debt accumulation, revolvers, especially those carrying heavy debt burdens from month to month, would be substantially more affected by the credit restrictions than non-revolvers.
Agarwal et al. hence deployed a difference-in-differences (DID) approach which identified revolvers as the treatment group, with non-revolvers providing the benchmark as the control group. The corollaries are that the revolvers would make a significant attempt to reduce their debt burden, whereas the control group was anticipated to remain relatively indifferent or be less likely to significantly alter their credit card usage patterns. In so far as such a phenomenon was observed during the post-policy period, there would be evidence to suggest that the MAS policy had succeeded in reducing credit card debt among those heavily indebted.
Data analysis was conducted on a proprietary panel dataset obtained from a major credit card company, and the sample consisted of 180,000 card users in Singapore split into approximately 90,000 revolvers (treatment group) and another 90,000 non-revolvers (control group).1 The dataset included monthly credit card statements, transaction-level credit card spending data, as well as demographic information, and covered the period from January 2013 to June 2016, hence allowing the researchers to do pre- and post-policy investigations.
Pre-policy observations
During the pre-policy period, the average monthly accumulated credit card debt for all individuals in the sample was S$955, accompanied by an average domestic credit card spending of S$82 (excluding expenditure outside Singapore). The average monthly total balance owed to the credit card company was S$1,113.
These figures were however skewed by the recurring debt obligations of revolvers in the sample, thus making it seem as though everyone in the sample – both revolvers and non-revolvers – were mired in debt. After drilling down into the data, revolvers turned out to have a monthly average of S$2,066 in outstanding balance, which was nearly 14 times more than the S$152 average for non-revolvers.
Additionally, efforts at reducing credit-card debt obligations by revolvers relative to non-revolvers during the pre-policy period were statistically insignificant – thereby suggesting that there was an existing parallel trend between the treatment and control group, which was expected to diverge after policy implementation.
Post-policy findings
After policy implementation, revolvers on average reduced their monthly debt by a statistically significant 9.9% (or S$188) relative to non-revolvers. The treatment group exhibited an additional reduction of S$19.90 in monthly domestic credit card spending compared with the control group, while total balance owed to the bank dipped by S$211. These figures respectively represent 24% and 10.2% of the pre-policy averages.
A longitudinal analysis revealed a clear and gradual reduction in monthly accumulated credit card debt following the announcement of the credit suspension policy.
Robustness tests
The objective of the MAS policy was to curtail households' unsecured debt accumulation while safeguarding essential needs. Credit extended for purposes such as education, medical needs, or business were therefore exempted from the credit suspension penalty. Accordingly, if the observed changes in credit card behaviours genuinely stem from the credit suspension policy, spending related to needs-based purposes would not be affected.
Utilising transaction-level credit card spending data, Agarwal et al. found that the treatment group did not significantly reduce their spending on essential needs, and furthermore, the observed reduction in credit card spending was primarily driven by spending in other categories that were not exempted from the policy.
To further mitigate the potential impact of confounding factors, propensity score matching was employed to match revolvers with non-revolvers of highly similar demographics (e.g., gender, age, income, etc.). The initial results derived from the unmatched sample, which showed a significant difference in debt reduction between revolvers and non-revolvers, remained robust – post implementation, revolvers continued to significantly reduce their monthly credit card debt.
Additional findings
Though the MAS policy was specifically targeted at individuals at high risk of eventual default, Agarwal et al.’s study revealed other intriguing insights.
Despite the credit suspension policy not being binding for the majority of revolvers, a significant number of individuals – including those who were relatively distant from the penalty threshold – still demonstrated a substantial reduction in their debt obligations.
This finding suggests a precautionary motive at work (including hedging against extreme negative events), especially among those with lower incomes and homeowners with mortgages. A voluntary behavioural adjustment suggests that the effects of the credit suspension policy percolated to a wider segment of the population, and the policy was therefore also an effective tool to address the overall risks of escalating unsecured debt.
1. In a panel dataset, the same group of individuals (in context, the sample of nearly 180,000 credit card users) are tracked over a period of time through multiple observations.
Agarwal, Sumit is the Low Tuck Kwong Distinguished Professor at the School of Business and Professor in the departments of Economics, Finance and Real Estate at the National University of Singapore.
Qian, Wenlan is the Ng Teng Fong Chair Professor in Real Estate and Professor of Finance and Real Estate at the NUS Business School.
Zou, Xin is an associate professor of finance in Department of Accountancy, Economics and Finance at School of Business, Hong Kong Baptist University.