Even when street vendors are freed from debt and educated about the benefits of saving, they go back to borrowing from moneylenders at exorbitant rates

Many street vendors in developing countries, and in some developed countries too, rely on daily or weekly loans from moneylenders to finance working capital i.e. the capital which is used in day-to-day operations. Banerjee and Duflo (2007) report borrowing rates from 11% (rural poor in East Timor) to 93% (in Pakistan) across 13 developing countries, oftentimes from moneylenders (see Banerjee (2004) for a survey of the literature documenting high interest rate borrowing across developing countries). 

Why not borrow less?

Given the high interest rates, why don’t vendors use a little bit of each day’s earnings to buy working capital, thus borrowing less? By doing this little by little, pretty soon no debt would be needed to buy their working capital. To try to answer that question, we conducted three experiments in which we gave cash grants and brief financial training to Indian and Philippine vendors labouring under high interest rate debt.  We then tested whether, and how long before, they returned to debt. Studying such behaviour is important for understanding the financial decision making of households in dire circumstances and for setting appropriate consumer-protection policies.

Our results

One striking pattern emerged in our samples. Most vendors fell back into debt within six weeks, and two years later the likelihood and volume of borrowing at high interest rates were nearly the same for the vendors who received cash grants and those who did not. The brief financial education we provided did little to change that behaviour in the long term. That said, there were still important differences across experimental sites. Most notably, in one of the Philippine experiments, while long-run borrowing rates were the same for treatment and control vendors, those who received our cash grants enjoyed substantially higher profits.  

These results illuminate how hard it was for these entrepreneurs to free themselves from debt. Not only were they unable to save their way out of it, but also even after receiving a windfall grant from us, they soon went back to borrowing.

Trying to explain this pattern

No doubt there is no single explanation for everyone’s reversion, and determining the relative importance of each would require more observations and data than our sample provided. Still, we explored several possibilities. For example, these entrepreneurs may have been susceptible to unavoidable and unexpected expenses and income shocks that forced them to return to borrowing. It’s also possible many suffered from a lack of self-control or did not have access to reliable saving technology. Finally, they may not have understood the long-term cost of repeated borrowing at high rates.

Policy implications 

Although there are many questions that still need to be answered, research of this kind can have an impact on public policy. If, for instance, financial literacy and poor planning are at the heart of the problem, public policies that require disclosures focusing on the cost of debt over various horizons may change behaviour (Bertrand and Morse 2011). Note that a short financial education module on the cost of debt did not help. Similarly, if return to debt is driven by unabsorbed shocks, this makes salient how debt is substituting for a missing insurance market for the poor. Thus, improvements to social protection policies and insurance markets may help alleviate a problem erroneously perceived as a credit market problem.

What drives the debt cycle?

The cause of the debt cycle has important regulatory implications regarding rules about rolling over loans from one to the next. This is a hot issue in the United States today.

Suppose people find themselves in perpetual and expensive debt, all because they took out one expensive loan and then had to keep borrowing to pay back the last loan. The reason they did this was because they were falsely optimistic about their likelihood of paying off the first loan without having to borrow again. In this world, it may be best to have a law that prevents immediate rollovers and makes sure borrowers know this. Knowing they cannot just borrow again to pay back their current loan may trigger a cutback in expenses to then successfully payoff their loan.

If on the other hand, individuals are in what seems like perpetual debt because expenses simply keep arising, and they live at the edge of liquidity at most times, then rollover rules on debt serve little to no purpose. A second loan or third loan is satisfying some short-run need just the same as the first loan. Instead a focus should be on helping individuals plan expenses, and save during the times when they are not busy paying off expensive debt.

Because we find that vendors return to borrowing even after their initial debt was entirely paid off, our study suggests that the latter may be the right story. Of course context may matter, and we are keen to see this tested more, in both rich and poor countries.


Banerjee, A (2004), “Inequality and investment”, World Bank Growth Commission Chapter. 

Banerjee, A and E Duflo (2007), “The economic lives of the poor”, Journal of Economic Perspectives 21(1).

Bertrand, M and A Morse (2011), “Information disclosure, cognitive biases, and payday borrowing”, Journal of Finance 66(6): 1865–93.

Skiba, P and J Tobacman (2011), “Do payday loans cause bankruptcy?”, Working paper. Vanderbilt University, Nashville TN.

India Philippines Credit markets Debt Loans Interest rates Street vendors Working capital