Implicit Corruption in the Chinese Consumer Debt Industry? A Close Look at Recent Evidence

While many country’s bribery laws require an express quid pro quo—an agreement to exchange a specific benefit for a specific exercise of government power—in practice many corrupt relationships involve implicit quid pro quos, in which the private party provides something of value to government officials, and the government officials use their power to help their private benefactors, but there is never any express agreement, or even any direct connection between any individual official act and a particular benefit conferred by the private party. The context in which such implicit quid pro quos are most widely suspected and discussed is perhaps campaign finance in democracies, but such implicit quid pro quos can occur in many other contexts as well. It is often very difficult—not only for law enforcement agencies, but also for empirical researchers—to find sufficiently clear evidence of an implicit corrupt deal. Yet quantitative empirical researchers have been making important strides in using available data to detect evidence of hidden or implicit wrongdoing—an approach sometimes dubbed “forensic economics.”

A fascinating recent paper by Sumit Agarwal, Wenlan Qian, Amit Seru, and Jian Zhang (forthcoming in the Journal of Financial Economics) illustrates both the potential and limitations of this approach. The paper, entitled “Disguised Corruption: Evidence from Consumer Credit in China,” presents quantitative evidence of an implicit quid pro quo between a large Chinese bank and government officials who wield regulatory authority over the bank. The paper finds that the bank offers unusually favorable lending terms to government employees (the “quid”) and that in those provinces where this practice is more widespread, the bank receives more favorable treatment from governments (the “quo”). While this evidence alone cannot establish that there was an implicit exchange (the “pro”), the authors suggest that this is the most plausible explanation of the data.

The data is certainly susceptible to that interpretation, but there are other, more benign possibilities. I’ll first say a bit more about the main evidence the paper offers for an implicit quid pro quo, and then suggest (though not necessarily urge) a possible alternative explanation.

  • First, the “quid”: The study finds that, even after controlling for a variety of factors that might affect a customer’s creditworthiness, bank customers who are government employees receive credit lines that are, on average, 16% higher than those of similarly-situated customers who are not government employees. In fact, government employees actually exhibit a higher probability of delinquency in repaying their loans, implying that they are, on average, inferior borrowers. Moreover, the probability that the bank restores a delinquent customer’s credit line is higher for delinquent government employees than for other delinquent borrowers, which the authors interpret as additional evidence that the bank offers more favorable treatment to customers who are government employees. Moreover, the authors find that after the Chinese central government initiated a major anticorruption crackdown in a given province, government officials no longer received significantly higher credit lines from the bank.
  • Next, the “quo”: The paper finds that in regions where government employees receive the benefit of higher credit lines—and only in those regions—the bank received broader regulatory approval to open new branches, and larger government deposits.

These findings do indeed suggest the possibility of an implicit corrupt exchange between the bank and government officials—or if not an actual exchange, then at least evidence that the bank improperly curries favor with government decision-makers by offering them terms not available to other customers. That said, the evidence of both the alleged “quid” and the alleged “quo” are susceptible to alternative interpretations.

  • First, with respect to the quid: The authors acknowledge the possibility that differences in credit lines might be due to differences in creditworthiness, and attempt to address this problem by controlling for variables like income, age, education, and a few others. But there could be additional differences between government employees and other customers with respect to creditworthiness that are not captured by these control variables. For example, government employees have higher job security. To this, the authors would likely point to their evidence that government employees are more likely than non-government employees to become delinquent on their debts, which the authors interpret as evidence that government employees are in fact less creditworthy than other borrowers. But the authors overlook an important fact: Delinquent borrowers are often the ones who generate the most profits to banks, because most delinquent borrowers frequently end up paying their debt with punitive interest rates that exceed the actual costs to the bank of extending the term of the debt. For this reason, the relevant measure of a customer’s creditworthiness should be the likelihood of default, not delinquency. And here, as noted above, the authors find that although government employees are more likely to become delinquent, they are also more likely (compared to other delinquent borrowers) to get their credit line reinstated. Although the authors interpret this as further evidence that the bank gives government employees favorable treatment, that interpretation seems problematic. While I don’t have specific information on the practices of this Chinese bank, in most cases delinquent credit lines get reinstated because they are paid, not because the bank agreed to forgive the debt. So it seems to me that the data purporting to show that banks offer favorable treatment to government employees could just be evidence that government employees (perhaps by virtue of their employment security) are more creditworthy, and that while they do end up delinquent on their payments more often, they are more likely to eventually pay off the debt (which, if true, might supply another reason why banks would be more likely to lend to extend them higher lines of credit). All that said, the fact that the differences in credit lines extended to government employees narrowed or vanished after the central government started cracking down on corruption is indeed intriguing. Yet this result is also open to an alternative interpretation along the lines sketched above. The anticorruption crackdown was associated with decreased employment security for government employees, dissipating an important source of creditworthiness.
  • Next, with respect to the quo, while the authors convincingly show that in regions where government employees received higher credit lines, the bank got more regulatory leeway for opening new branches, the conclusion that this is part of an implicit corrupt exchange may be premature. It’s possible, for example, that the extension of higher credit lines is associated with greater profits in the region, and that it is this higher profitability that drives the decision to open new branches. Or it could be that government employees’ positive experience as customers of the bank affects their judgment regarding the desirability of granting the bank licenses to open new branches. That might still be troubling—and might still be consistent with the concern that the bank intentionally favors government employees in the hope of shaping their professional judgment—but It wouldn’t be consistent with the conclusion of a corrupt quid pro quo, even an implicit one.

Notwithstanding these limitations, this study makes a significant contribution. It advances our understanding of the market mechanisms of implicit corrupt quid pro quo between governments and players in the private sector, emphasizing the need to look for corrupt practices more broadly than where we are used to be looking. Any reciprocal interaction between the government and private firms—especially one that is unscrutinized by the law—should give us pause. While there may be good reasons not to criminalize this type of behavior, regulators should be made more aware of these concerns and consider imposing harsher restrictions when a service is provided to government employees.

Leave a Reply

Fill in your details below or click an icon to log in: Logo

You are commenting using your account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )

Connecting to %s

This site uses Akismet to reduce spam. Learn how your comment data is processed.