In the K R Puram slum in Bangalore, India, a group of 15 women gather in a small, muggy living room. The electricity comes and goes, turning the fan and the single bare lightbulb on and off. Flies buzz around the room, and children run in and out.

The women have borrowed $330 and meet weekly to make repayments. The loans were meant to serve as capital for them to start small businesses and, eventually, lift themselves out of poverty. But the women say the loans haven't turned into new income.

Sitting in a circle on the floor, some sound sad and others angry. One woman has started selling firewood, but others haven't started businesses at all. Instead, they say, the money helped them pay for urgent expenses, such as their children's school fees.

It's a scene that's repeated inside dark living rooms and on parched rooftops across India. Boosted by a government mandate to keep the startup funds flowing, lenders making these small, or microcredit, loans are dutifully throwing cash at shantytown borrowers.

Slideshows:
Most luxurious credit cards
Top ten hottest jobs in India

And the funds-- $1.3 billion was lent during the year ended Mar 31, up from $4 million in 1996--are often being used not as seed money for a new enterprise, such as buying a cow to sell the milk or setting up a fruit stand, but as handouts spent on consumption.

D S K Rao, the Asia organizer for the Microcredit Summit Campaign, estimates that only one-fifth of Indian microcredit borrowers start a business. Mathew Titus, executive director of New Delhi-headquartered Sa-Dhan, an association of microcredit lenders, says it's one-third, counting women who barter instead of selling goods.

Alas, there has never been a rigorous study in India or elsewhere of how successful the businesses become, whether any of the borrowers eventually graduate to the middle class or how many of the loans get repaid on time. "I don't know of any great studies--they haven't been done," says New York University professor Jonathan Morduch, who focuses on microcredit and economic development.

For much of its 35-year history microcredit has enjoyed glowing reviews, culminating last month when Muhammad Yunus--founder of Bangladesh's microlending Grameen Bank--won the Nobel Peace Prize. The tiny loans appeal to both idealistic aid workers and gimlet-eyed bankers.

They seem to help the poor in developing countries earn a livelihood without making them dependent on handouts. And they use market mechanisms, allowing banks to charge interest rates high enough to cover the risks of lending to people with no collateral and little credit history. In India rates average 30 per cent a year, though some lenders charge more than 45 per cent.

In the last decade microcredit has taken off. The 1997 Microcredit Summit in Washington, DC--sponsors included the World Bank, MasterCard and JPMorgan--vaulted microlending onto the world stage.

Participants pledged to lend to 100 million families worldwide within a decade. Pushing to maintain the momentum, the United Nations named 2005 the International Year of Microcredit. Today organisations from the US Agency for International Development to Citigroup trumpet the idea.

Slideshows:
Top ten richest Indians
Emerging global cities

Much of that attention is focused on India, which is home to a quarter of the world's poor but also one of the hottest economies. The country has seen an explosion of microfinance institutions, which are usually set up by nongovernmental organizations. Ten years ago 400 institutions boasted 200,000 customers.

Today there are 1,000 that, together with 300 commercial banks, lend to 17.5 million people, according to Sanjay Sinha, managing director of Micro-Credit Ratings International in India. Much of the lending is based on Yunus' original idea--making loans to groups of women and relying on peer pressure from the members to ensure repayment.

"India was very small in this field ten years ago," says Sam R Daley-Harris, director of the Microcredit Summit Campaign. "Now they're going like gangbusters, expanding ferociously."

The money to fund all these lenders comes mostly from India's biggest commercial banks, such as ICICI [Get Quote], State Bank of India [Get Quote] and Canara Bank [Get Quote]. But some is put up by foreign investors and banks making social-responsibility bets, and by overseas donors.

Technology entrepreneurs are particularly intrigued. The foundations of Bill and Melinda Gates and Michael and Susan Dell have each given $4 million or more to microcredit organisations. Vinod Khosla, a Sun Microsystems founder and venture capital star, has put at least $1 million into Indian microfinance institutions, as donations and investments.

"In their anxiety to expand very fast, they are pushing loans," says Rao, the microcredit campaign organiser. "They're not careful in vetting." At least, some groups of borrowers are first required to save money that they can put down as collateral; others must outline their business ideas to lenders.

The rapid growth also means that lenders are less likely to keep tabs on a borrower after a loan is made. In India, says Daley-Harris, borrowers sometimes "are on their own after a couple of years, whereas in Bangladesh there's a bank worker who often visits every week for a lifetime." In Bangladesh there is one microfinance staffer for 131 borrowers and in Afghanistan one for 54, but in India the ratio is one to 439, according to the Microfinance Information Exchange in Washington.

The women in India could probably use more guidance, because it's so difficult to run a business there: The country ranks 134th out of 175 that the World Bank studied this year for ease of doing business.

Slideshows:
The pros and cons of outsourcing in India
The world's largest companies

The result is that lenders often hand over money without taking the time to educate the borrowers, making sure they understand that they're supposed to start businesses and that there are penalties for not paying the money back, says Gowramma, secretary of the Bangalore branch of the All India Democratic Women's Association, who works with women in the slums. "The people receiving these loans do not know what this microcredit system is all about," she says.

This means that microcredit's other goal--to empower women in highly patriarchal societies--isn't always achieved, either. In Tumkur, a city outside Bangalore, a group of women received $155 loans. They stood in a circle, placed their hands on the money and pledged to use their loans to launch businesses. But afterward, each handed the money to her husband, and the men started the businesses. Because the women are responsible for repaying the loan, they suffer the consequences if their husbands squander the money.

Microcredit advocates often claim impressively high repayment rates, averaging 95 per cent, as proof that borrowers do use their loans to generate income. Otherwise they wouldn't be able to pay them back, they argue. "Can [microcredit] be done badly? Absolutely," Daley-Harris says. "Can it make a massive difference in people's lives when it's done well? Absolutely."

But in some cases women borrow from other sources to repay the loans. That often means resorting to moneylenders--or loans from friends and family--to pay back the microloan and vice versa, trapping them in a cycle of debt. The lucky ones have family members who have migrated to cities or abroad to work and send back remittances.

In Bangladesh, for example, 40 per cent of microcredit borrowers also borrow informally, says Thomas Dichter, who has evaluated microcredit programs in 20 countries, including India. "You're getting a large number of people borrowing from Peter to pay Paul," says Dichter, who wrote the 2003 book Despite Good Intentions: Why Development Assistance to the Third World Has Failed.

More important, Dichter and others who study microfinance doubt the repayment rates. There are no reliable figures on the percentage of loans delinquent for more than, say, six months. But it's telling that the vast majority of the microlenders are not "sustainable," meaning that so few loans are repaid on time and costs are so high that lenders need to keep tapping the easily available funds from banks, donors and the government to stay in business.

More than anything, the microcredit bubble is being inflated by government rules that all but force lenders to keep pumping out microloans. New Delhi requires banks to devote 40 per cent of the money they lend to a category of borrowers that includes small enterprises, with about half of that going to rural outfits. Banks have long struggled to meet this obligation, and lending to microcredit institutions that then lend to the groups of women offers an easy way to do it.

Sometimes microcredit does live up to its billing. Manjula, a 40-year-old seamstress who lives in Bagalur, a village 90 minutes away from downtown Bangalore, struggled to support her two children without help from her alcoholic husband. She depended on moneylenders, who charge up to 1,000 per cent a year.

Two $220 microcredit loans over two years enabled her to open a shop selling saris, which she had been peddling from her home. Since then she has increased her monthly income from $45 to $110, including the earnings from side jobs tailoring and teaching sewing classes. That's enough to pay off her loan and save $30 a month. "My dream is to convert this place into a big shop," she says proudly.

It was success stories like this that drew Khosla, an affiliated partner in Silicon Valley venture capital firm Kleiner, Perkins, Caufield & Byers, to microcredit in his native India. Three microfinance lenders that he funds--SKS Microfinance, Share Microfin and Activists for Social Alternatives--are on a top-ten list compiled by Micro-Credit Ratings International, and he becomes so passionate about microcredit that his eyes tear up when he talks about it. "This is the single most important tool for addressing poverty I have ever seen."

Might a careful study turn that article of faith into a provable proposition? "I don't need that study," Khosla says. "I won't spend a penny on it because I've talked to 200 women who are recipients of these microloans myself. I've seen it in the field."

Even Manjula's success story needs some cautionary footnotes. She started out with distinct advantages. Unlike most of India's poor, she is educated, speaks some English and started her business before receiving the loan. And Manjula hasn't been able to completely stop borrowing from moneylenders.

"The research is skimpy, but the research we do have shows increasingly that the main hope for microcredit--that it will generate millions of tiny businesses and these businesses will grow and pull their owners over the poverty line--is false," says Dichter, the author.

By the Numbers

Microcredit is quickly spreading in India, but the loans don't always lead to businesses; in some cases they are even used to pay off previous debts.

• 200,000  -- People getting loans in 1996.
• 17.5 million  -- People getting loans in 2006.
• $4 million  -- Value of loans granted in 1996.
• $1.3 billion  -- Value of loans granted in 2006.
• $75  -- Size of average loan.

Figures for year ending Mar. 31, 2006.
Source: Micro-Credit Ratings International.