Are Information Disclosures Effective in the Credit Card Market? Evidence from Mexico
In the financial sector, regulators often mandate the disclosure of information such as interest rates and time to pay down debt in order to protect consumers. However, there is limited evidence on the effectiveness of such disclosures. Researchers tested the effect of seven different information disclosure messages on credit card clients of a large Mexican bank and found that non-standard information disclosures reduced debt levels and lowered delinquency more effectively than legally mandated disclosures.
Many countries have mandatory information disclosure laws in the financial sector intended to protect consumers. In credit card markets, the aim of such disclosures is to make individuals informed users of credit. However, there is little evidence on what information should be disclosed, how effective existing disclosure requirements are, and why they may or may not be effective. Moreover, there is little evidence that these mandatory disclosures are worth the cost, which is often borne by financial institutions. In the context of the credit card market, do consumers make decisions related to debt, default, and account closure in response to the type of information they receive?
Context of the evaluation
Thirteen percent1 of adults in Mexico report having a credit card compared to only 7 percent in low- and middle-income countries, but still well below other Latin American countries like Brazil (29 percent) and Argentina (22 percent). Mexico, along with the United States and several other countries, requires credit card statements to include information about interest rates and the number of months to pay (MTP) existing debt if the consumer made the minimum monthly payment with no further purchases.
The Mexican Banking Commission (CNBV) wanted to study the effectiveness of issuing a requirement that banks send warning messages to clients with personalized information about their risk level.
The partner bank in this study also had an interest in introducing new measures to reduce delinquency in its credit card portfolio. The riskiest third of its clients had a 26 percent likelihood of defaulting, on average. These clients had very high debt levels relative to their income, were unaware of their interest rate, and overconfident in their ability to repay their debt. Their debt averaged 17,800 Mexican pesos (US$1,339), well above their average income of around 8,500 pesos (US$640) per month. The bank expected to lose around 2,700 pesos (US$203) per account per month.
Details of the intervention
Researchers, in partnership with the Mexican Banking Commission and a large Mexican bank, conducted a randomized evaluation to understand whether different types of information disclosures affect credit card customers' levels of default and indebtedness. Researchers randomly selected 167,190 of the bank's riskiest clients for the study and divided them into 150 groups based on gender, age, credit limit, and default probability. They randomly selected 77,175 clients to receive one of seven information disclosure messages, and the remaining 90,015 clients received no messages.
Researchers compared the effects of sending clients two standard information disclosures, which contained legally mandated information about interest rates and months-to-pay (MTP), with five alternative disclosures, which provided social comparisons or warnings. Recipients of the non-standard disclosures still received the legally mandated information contained in the standard disclosures through their regular credit card statements.
Legally mandated information disclosure messages:
The Interest rate message informed clients about their personalized credit card interest rate (12,900 clients)
The Months-to-pay (MTP) message informed clients about the number of months it would take them to pay off existing debt if they made the minimum payment and did not make additional purchases (12,825 clients)
The High debt message informed clients that their debt levels were higher than average (12,825 clients)
The High debt with advice message informed clients that their debt levels were higher than average and gave them specific advice to reduce it (12,825 clients)
The High default risk message informed clients that their default risk was higher than the median predicted default probability and recommended specific measures to reduce the debt burden (6,444 clients)
The Low default risk message congratulated clients for their lower than the median predicted default probability risk (6,456 clients)
The Warning message cautioned clients against overconfidence in their ability to pay down credit card debt (12,900 clients).
The messages were printed and mailed to clients in February 2011 using the bank's existing system for mailing monthly statements.
Results and policy lessons
Non-standard information disclosures were more effective than legally-mandated disclosures in reducing debt levels and lowering delinquency rates. However, these effects were very small and short-lived, lasting for at most two months. These effects also varied between higher-income clients with income in the top 20 percent and low-income clients with income in the second quintile of the income distribution. It is impossible to know how many clients read the messages, so these results are the effect of sending the messages, not the effect of reading the information.
Impact on debt levels: The two legally-mandated disclosures had no effect on credit card debt levels. The high debt with advice message also had no effect. However, the warning message did lower debt levels. Higher-income clients who received either of the high debt messages or the warning message reduced debt by around 1,200 pesos (US$\90). This was not the case for low-income clients. Finally, while high risk of default messages reduced debt, only higher-income clients who received the low risk message increased debt levels by around 2,300 pesos (US$173).
Impact on delinquency rates: Only the non-standard disclosure messages related to risk had any impact on delinquency, measured as the percentage of accounts with payments overdue by 30, 60, or 90 days. For clients who received the high risk message, delinquency rates decreased by 1.5 percentage points after one month from a base of 18.3 percent. This effect was concentrated among low-income clients as their delinquency rates decreased by approximately 3 percentage points in the following two months. There was no change in delinquency among higher-income clients. The low risk message had an unintended negative effect, as delinquency among these clients increased over the next two months. The other messages did not affect delinquency.
Impact on account closures: Only the high risk message had an impact on account closures, with clients being 0.7 percentage points more likely to voluntarily close their account, from a base of 4.3 percent.
The results suggest that policymakers and financial institutions may be more successful in reducing indebtedness and risk levels of clients if clients receive information that is actionable, easy to understand, and appropriately benchmarked in place of the information disclosures that are currently mandated. Sending such messages is relatively cheap, so this may be a cost-effective strategy, even if the effects are small and relatively short-lived.
Elizondo, Alan, Eduardo Laguna-Müggenburg, and Seira, Enrique. 2017. "Are Information Disclosures Effective? Evidence from the Credit Card Market." American Economic Journal: Economic Policy, 9(1): 277-307.