In developing countries, economic behavior does not always follow patterns set in developed ones.

Two papers presented at a conference on India’s Financial System explore such examples. One asks why the poorest Indian farmers choose not to purchase inexpensive rainfall insurance that could help them avoid starvation in a drought. A second explores an unexpected phenomenon in Pakistan, when banks failed to increase their business loans despite a rich increase in capital after the September 11 terrorist attacks. The conference, held at Wharton in April, was organized by the school’s Financial Institutions Center with the Centre for Analytical Finance at the Indian School of Business in Hyderabad and the Stockholm-based Swedish Institute for Financial Research.

Rainfall Insurance for the Most Vulnerable

Farming has always been risky, but in developed nations, farmers use futures contracts, insurance and other financial products to make it through the lean years.

One might expect poor farmers in a developing country like India to jump at the chance to do the same. But when a company started offering inexpensive rainfall insurance several years ago, fewer than 5% of the eligible farmers bought it.

Defying expectations, the insurance was more likely to be purchased by the farmers who needed it least, while it was rejected by the farmers at greatest risk of financial catastrophe if the rains failed. “Participation rates were lower among the more vulnerable households,” said James Vickery, a research economist at the Federal Reserve Bank of New York, who outlined his paper entitled, “Patterns of Rainfall Insurance Participation in Rural India.”

Why? The answer should make any marketing or advertising executive smile: Many potential customers simply didn’t understand the product — not well enough, at least, to dip into their severely limited resources to pay for it. With better marketing and a few modifications to the policies, it may be possible to get more subsistence-level farmers to embrace rainfall insurance, helping to soften periodic famine in some of India’s poorest areas.

“This type of insurance is in its nascent stages,” said Vickery. “The goal of the product is essentially to insure against insufficient rainfall during the monsoon season…. It may not be very well understood.” His co-authors are Xavier Gine of the World Bank and Robert Townsend of the University of Chicago.

While harvests can fail for many reasons, weather is the most common. Because a drought hits an entire geographical area, it is hard for a stricken family to get help from relatives and neighbors who most likely are suffering as well. Subsistence farmers in India try to reduce drought risks by scattering plots over as wide an area as possible and emphasizing hardy varieties of crops. But there is a cost: They avoid varieties that might be more productive but are more sensitive to weather.

For a number of years, the Indian government has tried to ease the problem with the National Agriculture Insurance Scheme. Farmers who take out crop loans for seed and other purchases are required to carry this insurance, and it is available to others as well. Payouts are based on measured crop yields in designated test plots. But participation has been very low — just 9% of the 138 million rural households.

The system has a number of problems. It takes a long time for claims to be paid, for example. And the geographical areas covered are so large that farmers worry that test plots may not suffer the same low yields their own plots do, denying them needed insurance payouts. In recent years, a number of insurers have tried to resolve this problem by offering insurance based on rainfall measured at gauges long operated by the Indian Meteorological Department. This provides a simple, trusted, transparent and low-cost index on which to base insurance claims.

The study looked at “micro-insurance” policies developed by ICICI Lombard with help from the Commodity Risk Management Group of the World Bank. The product was marketed in two rural regions by BASIX, a micro-finance institution that had been offering farmers small loans for years. The policies were first available in 2003; the study focused on 2004.

The policies were designed for the two main cash crops in the area, castor and groundnut, and were aimed at farmers with two to 10 acres and annual incomes of $375 to $750. A $5 policy covered one acre, and offered a maximum claim of about $150. Payments were based on the degree to which rain fell short of given thresholds during three stages of the growing season from June through September, or when there was too much rain. In Narayanpet, a mandel, or county, in the Mahaboobagar district, insured farmers received $43.50 an acre in 2004, for example.

Who Bought and Who Didn’t

In 2004, policies were bought by 315 farmers in 43 villages and covered 570 acres, giving the average farmer $240 in maximum coverage. The researchers conducted a survey at the end of the season to find out why some farmers bought insurance and others did not. They questioned 1052 households in 37 villages. In the surveyed area, only 267 of 5,805 eligible households bought the insurance — 4.6%.

Surveyed households had median landholdings of four acres, liquid assets of about $200 and net worth of less than $2,000. Household heads had a median of 3.3 years of formal education.

The survey found a number of differences between farmers who bought insurance and those who did not. “Buyers are around one-third wealthier, report around 50% more land and nearly twice as much in liquid assets,” the authors write.

A test asking farmers to choose between hypothetical bets found that insurance buyers were less concerned about risk than non-buyers. About a third of buyers were members of groups that shared wells, compared to 4% of the overall population, and 46% of buyers had outstanding loans from BASIX, compared to 7% of the overall population. “Buyers are twice as likely to be members of the area Gran Panchayat (local council), and are also more likely to self-identify as ‘progressive’ households,” the study found.

Among the households that bought insurance, 65% cited reasons such as reducing risk and assuring a harvest income. In general, insurance was purchased by farmers who, compared to non-buyers, were in a better position to make it through a drought. Theory predicts the opposite.

Why did the poorer farmers avoid insurance even though they needed it more? “Strikingly, the most frequently cited reason among non-purchasers is that the consumer did not understand the insurance product,” the authors write. About 25% of non-purchasers gave that reason, while 21% said they could not afford the premium. Another 24% worried there was too big a chance they would not receive an insurance payout, citing reasons such as the rain gauge being too far away.

Higher participation was found among farmers belonging to well associations, those with loans from BASIX and those who said they trusted BASIX.

The authors concluded that “even though insurance is available in principle to all households in the village, the households’ familiarity and trust in the insurance provider constitute the most important determinant of households’ purchase decisions.” Basically, the study found that insurance was purchased by the more sophisticated farmers who could afford to take a chance on a new product.

The authors suggested that some modest changes would encourage more farmers to buy insurance. Since farmers are short of cash during the growing season after paying for things like seed, one improvement would be to make insurance payouts immediately after each of the three phases of the growing season rather than in November, as was done in 2004. The insurer also could let farmers buy insurance on credit, allowing them to pay off the loans after the harvest. “A household with more land and more wealth is more likely to purchase insurance,” Vickery said. “That seems generally consistent with the idea that credit constraints are important.”

Farmers also would be more likely to buy insurance if the product were better explained. In 2004, the product was offered in marketing meetings with much of the explanation provided by other farmers who may not have understood the product very well themselves. “Early adopters of insurance are likely to be households where the cost of experimenting with the insurance is relatively low…,” the authors write, adding that “over time, lessons learned by insurance ‘early adopters’ will filter through to other households, generating higher penetration rates among poor households.”

Lending in Pakistan and “Demand Shock”

Money flows to where the potential returns are greatest. Except when it doesn’t.

In recent years, economists have puzzled over why money is flowing out of rapidly developing countries — through enormous foreign investments in U.S. Treasury bonds, for instance. “Developing countries, especially the higher growing countries such as China and India, are actually exporting money back to the developed countries,” said Bilal Zia, economist at the World Bank’s Development Research Group.

Theory and history predict the opposite: Capital should flow from the developed countries to the developing ones, where greater room for growth offers bigger investment returns.

What causes this unexpected pattern? For an answer, Zia and two colleagues looked at bank lending in Pakistan after the September 11 terrorist attacks in the U.S. They found that conservative bank lending practices impede developing countries’ ability to make the best use of growing capital reserves. “Banks are just not capable; they don’t have the ability to absorb the capital that’s going in,” Zia said. Because companies cannot get bigger loans, economic growth is slowed.

The results are reported in a paper titled, “Dollars Dollars Everywhere, Not a Dime to Lend: Credit Limit Constraints on Financial Sector Absorptive Capacity.” Co-authors are Asim Ijaz Khwaja of the Kennedy School of Government at Harvard and Atif Mian of the Graduate School of Business at the University of Chicago.

The researchers chose Pakistan because it offered a good example of a “demand shock” — a rapid economic change that provides good before and after data. Prior to September 11, Pakistan was in economic trouble, largely because of international sanctions imposed after the country exploded a nuclear bomb in 1998. Annual economic growth had fallen to between 3% and 4%, compared to 6% in the first half of the 1990s.

This quickly changed after September 11 when Pakistan was seen as a key ally in the war on terrorism. “This was primarily due to the removal of international financial sanctions, a reversal of capital flight, and a significant increase in international economic assistance,” the authors write. “The net result was an unexpected surge in the supply of liquidity, a sharp drop in real interest rates, and a rise in aggregate demand.”

Between June 2001 and June 2003, remittances from Pakistanis living abroad increased by 300%, and interest rates fell from 11% to 2.5%. Foreign exchange reserves quintupled in less than two years. Yet the amount of bank lending to companies remained virtually unchanged. “Given the low cost of funds and positive demand, one would expect an increase in overall bank lending to firms, absent any lending constraints,” the authors write. “In reality, the macro evidence is extremely stark and shows little change in corporate lending, despite such a large and positive net demand shock.”

Conservative Lending Practices

The problem, Zia and his colleagues conclude, was conservative lending practices. Rather than looking at a borrower’s potential for growth, for example, banks tend to base lending limits on the borrower’s existing collateral. “Credit limits, once set, are actually very sticky,” Zia said, explaining that banks are reluctant to raise them.

“The central bank’s ‘prudential regulations’ provide strict guidelines to banks in terms of how credit limits should be determined for applicants,” the authors write. “These guidelines are very conservative in terms of collateral requirement, and bind a firm’s credit limit to its past cash flows. For example, total unsecured lending of a given firm cannot exceed 500,000 Rs (about $8,500).” In addition, total debt is kept very low relative to a company’s total equity, they found.

Many private banks are even more conservative, though certain types of borrowers have better access to loans. Exporters, for example, often have higher credit limits, since they are doing business with foreign firms with good reputations, the authors find. Similarly, large firms have better access to credit than small ones, since large ones have better reputations and more collateral.

Excessively tight lending practices have a number of negative effects. Among the most important is the income lost because companies cannot get the capital they need to grow. The authors estimate this cost at the equivalent of 2.3% of Pakistan’s gross domestic product in 2000. “This is a huge loss,” Zia said.

In addition, according to the authors, reserves that are not disbursed through bank loans become easy money that fuels risky investment. In the two years following September 11, stock prices in Pakistan increased five-fold, while housing prices grew at well over 100% a year.

“Evidence that this was a speculative bubble is becoming increasingly apparent with the recent collapse of the real estate market and a noticeable cooling off in the equity markets,” the authors say, adding that to get the fullest benefit from economic growth, developing countries like Pakistan, China and India need to find better ways to put capital to work at home, rather than investing it overseas.