The Impact of Debt Relief in India and the Philippines

Fieldwork by:
Location:
Chennai, India
Sample:
India 2007 (1000 market vendors), Philippines 2007 (250 market vendors), Philippines 2010 (701 market vendors)
Timeline:
2004 - 2007
Target group:
  • Entrepreneurs
  • Urban population
Outcome of interest:
  • Savings/deposits
  • Take-up of program/social service/healthy behavior
Intervention type:
  • Cash transfers
  • Financial literacy
  • Training
  • Unconditional cash transfers
AEA RCT registration number:
AEARCTR-0002688
Partners:

Repeated borrowing from moneylenders at high interest rates is common across many low-income communities. Researchers evaluated whether offering market vendors cash grants to pay off existing debt and financial training influenced future borrowing behavior. While market vendors were less likely to borrow and borrowed in smaller amounts in the short-term, most returned to debt within six weeks. One to two years later, individuals who received the cash grants and financial training were borrowing at the same rate as households who did not receive the cash grants and training. 

Policy issue

When in need of cash to expand their business, micro-entrepreneurs throughout the world often repeatedly borrow from moneylenders, who charge high interest rates for loans. Therefore, they can find themselves in a debt trap, when they cannot pay back these high interest loans and must refinance so that they are perpetually in debt. It is unclear why people take on these high-interest loans, when instead they could save and borrow less. Studying debt traps is important for understanding how households make financial decisions when facing resource constraints and also for setting appropriate consumer protection policies. Can paying off existing high-interest debt relieve people from borrowing from high-interest moneylenders in the future and help them avoid falling back into debt? 

Context of the evaluation

Repeated borrowing from moneylenders at high interest rates is common across many low-income communities. For example, in the state of Tamil Nadu, India—where one of the three studies took place—a 2010 survey revealed that 13 percent of urban households and 39 percent of rural households borrowed at a monthly interest rate above five percent. At the time of the studies, borrowers in India paid an average daily interest rate of 4.7 percent, while borrowers in the Philippines paid an average monthly interest rate of 13 percent.   

Across three separate studies in India and the Philippines, this evaluation focused on market vendors in urban centers with existing high-interest debt. Researchers identified vendors in the cities of Chennai, India and Cagayan de Oro, Philippines based on the three following characteristics. First, they were the primary decision-maker of the business. Second, they borrowed consistently from a professional moneylender (defined as offering interest rate at, or above, 5 percent per month, although most vendors borrowed at higher rates) for the past five years. Finally, they had an outstanding loan balance below a certain amount, such that researchers could pay off the entire debt. On average, study participants had PHP 2,172 (US$44 at the time of the evaluation) in debt from moneylenders in the Philippines 2007 study, PHP 3,234 (US$70) in the Philippines 2010 study, and INR 751 (US$17) in the India 2007 study.

Details of the intervention

Researchers conducted a series of randomized evaluations in India and the Philippines to evaluate the effects of a cash grant to pay off market vendors’ existing debt and a brief financial training on their future borrowing patterns.

In 2007, the researchers randomly assigned 1,000 market vendors in India and 250 market vendors in the Philippines to one of four equally sized groups: 

  1. Debt payoff: market vendors received cash grants to pay off their debts. The grant was equal to the amount of debt owed, up to PHP 5,000 (US$100) in the Philippines or INR 3,000 (US$50) in India.
  2. Financial education: market vendors received a brief financial training in small groups, implemented by a professional survey team. The training occurred prior to the announcement of the debt payoff, and emphasized two messages: lending at high rates from moneylenders is costly, and saving could help avoid taking out loans. The training included a lecture, discussion, and interactive activities. 
  3. Debt payoff and financial education: market vendors received both the cash grants and financial training.
  4. Comparison: market vendors received neither the cash grant nor financial training. 

In 2010, the researchers conducted a similar study again in Cagayan de Oro, Philippines, but in different markets than the previous 2007 study. To test if vendors borrowed because they lacked a way to save, researchers introduced a new treatment arm that offered vendors a savings account. However, due to the difficulty of providing the necessary documentation and identification to open an account, very few participants opened a savings account. As a result, researchers analyzed results for two groups: vendors who received the debt payoff combined with financial training (regardless if offered a savings account or not) and vendors in the comparison group.

In all three studies, researchers conducted three or four follow-up surveys starting one month after paying off the debt and ending, at the latest, after two years. Follow-up surveys measured vendors’ loans from money lenders and other types of debt, business performance, household shocks, consumption, and savings.

Results and policy lessons

Overall, debt payoffs had a modest and short-lived impact on debt levels and borrowing behavior. Across all three studies, most borrowers returned to debt within six weeks after their debt was paid off. One to two years after intervention, individuals who received a cash grant and financial training were borrowing at the same rate as households in the comparison group. 

Paying off market vendor’s existing debt from moneylenders had modest short-term effects, but no effect on future borrowing from a moneylender.

In the 2007 India study, borrowers granted debt relief were 17 percentage points (25 percent) less likely to borrow from a moneylender after two to four months, and borrowed US$8 less than the comparison group average of US$25 (a 32 percent decrease). These effects dissipated entirely nine to ten months after the intervention. Results were similar in the 2010 Philippines study, with a small effect one month after the intervention that faded away after four months, and up to nineteen months later.  

Effects were larger and more persistent in the 2007 Philippines study. Borrowers offered the debt payoff were 37 percentage points (54 percent) less likely to borrow from a moneylender one month after the payoff and borrowed US$47 less than the comparison group average of US$82 (a 57 percent decrease). After three months, they remained 28 percentage points (41 percent) less likely to borrow. 

Financial training alone had no effect on borrowing behavior.

In the 2007 studies in India and the Philippines, financial training alone had no impact on borrowing. In India, results were similar for borrowers who received debt payoff alone or combined with financial training. In the Philippines, however, the combination of debt payoff and financial training seemed to have had a larger effect than the debt payoff alone. This suggests that the financial training in the Philippines may have slowed how soon market vendors returned to debt. Nevertheless, the combination of debt payoff and financial training had no effect after 18 months.

Debt relief had little to no impact on household’s ability to cope with shocks, consumption, savings, or business profits.

In the 2007 India and the 2010 Philippines study, debt relief had small and short-lived effects on household’s ability to cope with income shocks. For example, in India, vendors who received debt relief were 12 percentage points (75 percent) more likely to use savings, and 6 percentage points (86 percent) less likely to reduce consumption, to cope with low household income after two to four months. However, these effects faded away after four to eight months. Effects were somewhat larger, but still modest, in the 2007 Philippines study. After one month, households offered debt relief were 45 percentage points (85 percent) less likely to borrow when facing an income shock, and those offered debt relief and financial training were 33 percentage points (62 percent) less likely. While this effect faded away for the vendors who received only debt relief, for those who received debt relief combined with financial training, the effect persisted after eighteen months.

Across the board, debt relief had little to no impact on vendors’ expenditure and savings. Only vendors in the 2010 Philippines study seemed to have experienced a sustained increase in business profits. After eighteen months, vendors offered debt relief earned an extra US$1.15 in profits on a typical day, relative to the comparison group average of US$6.9 (a 17 percent increase).

Taken together, these results illuminate the difficulty that small-scale entrepreneurs face in escaping high interest debt from moneylenders. These borrowers seemed unable to save their way out of debt, and, even after receiving debt relief, they fell back into borrowing at some point in the following months. Market vendors may repeatedly borrow from high interest moneylenders because there are high returns to investments to justify borrowing. Vendors may also face frequent shocks that force them to return to borrowing. Alternatively, vendors may suffer from self-control problems that make saving difficult, may not have access to a reliable way to save, or may not understand the long-term cost of repeated borrowing at high interest rates. Addressing these specific challenges may be effective in helping urban entrepreneurs escape debt traps.

Ananth, Bindu, Dean Karlan, and Sendhil Mullainathan. "Microentrepreneurs and Their Money: Three Anomalies." Working Paper, Harvard University, May 2007.