Sunday, November 18, 2007

Microfinance, anyone?

Social networks or social capital is one of the topics have been discussed in this Cafe (even debated among its barristers). It is a topic that has been of a growing interest of some economists. Harvard's Economic Department even offers a course devoted specifically to this issue.

Roby, a sociologist, regular visitor of this Cafe, as well as close friend of the barristers, can be considered as the authority of this field. He sent us this piece about microfinance as a potential intersection for social scientists to study more about mechanism design problem. Although it is not the first time that microfinance is being discussed in the context of social network (this is one example, or here is a list of more serious stuffs), still it is interesting to see Roby's challenge will be answered in the Indonesian context.
Microfinance, anyone?
by Roby

I am one of the fans of this blog but so far I only contribute by writing rather negative comments. In this post I want to share something positive.

As you may know, I am not an economist but some of my colleagues here are economists or working on topics that are also of interest for economists. I must say that I like economics but can't stomach some economists especially when they start to pretend to understand something and come up with "explanations" that actually don't explain anything. Having said that, I do think economists have succeeded in creating some useful arsenals and insights to understand human behavior.

Recently my boss left his tenured position to join the microeconomic research group of a leading web company. Since his departure, I regularly come to his new office and meet a lot of interesting people with different background, including economists. I try to know more about the kinds of interdisciplinary work where computer scientists, physicists, mathematicians, economists, psychologists and sociologist are working together.

Social scientists and economists are hired by major internet firms to help those firms to understand user behavior, from buying or selling products to hanging out in social networking sites. Armed with the understanding, these firms hope they can find better ways to monetize the services they offer through their websites. From the scientific point of view, access to huge data of human interactions and relatively "unlimited" resources (in comparison to resources in
academia) give us the chance to study human dynamics in unprecedentedscale and detail.

All of these are interesting but these kinds of work are not very relevant to Indonesia. Although the Internet in Indonesia is growing, the scale of the usage – in terms of business or social – won't be even near the usage in the US anytime soon.

However, I think there is an area in particular in which similar basic ideas can be applied in Indonesia: the microfinance industry.

I know next to nothing about microfinance, but here is what I think. To my understanding, microfinance is about figuring out ways how to do business with people who are traditionally unbankable. They are usually poor people in rural area. They are untouched by traditional
finance system because it is too hard to assess their credit worthiness or it's simply too risky.

One way to overcome the problem is to deal them collectively instead of individually. The idea is to utilize some group mechanisms to minimize the risk: group members would make sure everyone to repay the loan and hence guarantee continued access to lenders.

We can approach the problem as an interdisciplinary mechanism design problem: social scientists and social psychologists could study the relevant properties of group dynamics in (rural) Indonesia, then, together with economists, design how incentives could be structured
in a particular setting.

The project, I think, is intellectually interesting and would be easy to gain support since everyone cares about the poor in Indonesia.

Is this possible?


  1. a collective deal may lead to a limited growth due to a limited size of loanable funds shared and success expectation. so individual deals may be better by selecting some good entrepreneurs. we don't expect all small entrepreneurs would be successful, it's an over estimation. take 50% of them by dealing individually.

  2. I agree to this so called collective credit giving. But the task of creating proper structure to minimize Non Performing Loan (unpaid loan or NPL) and maximize the "utility" of loan is really really tough IMHO, despite any successful well coordinated cross-disciplined work between economist, sociologist, and other related fields.

    A problem I'm aware of in effectively reducing NPL is the tendency of poor people in rural area to trust an institution employee instead of the instution (s)he represents. No trust means not much honest information sharing thus difficulty to assess the group credit risk. Building trust is slow and somewhat time consuming.

    With trust, the institution employee can direct the group to utilize the loan more effectively. But for the employee to be a trusted director in that group, the employee must be well versed in that group culture while able to withstand detrimental influence (detrimental for microfinancing perspective) by that very same culture.

    There should be a psychological screening to choose institution employees who are sensitive to cultural cues and comfortable in following the rules but detached enough to be not influenced by that culture if necessary. Probably people who can almost be categorized with anti social personality disorder are the best employees...

    History of living in similar rural area will help the employee to be accepted in the group. On the other hand, a
    a person from the same rural area he works with may be strongly influenced by his relationship to people in that area, despite of an individualistic psychological profile.

    A simpler way is to trust the group in managing their own finance, follow financing policy to the letters, and cut them of any financing if they screw up. Far lower cost and much more chance for the microfinancing to become profitable. But (I believe) lower utilitization of loan.

    If the microfinancing institution can hand picked the member of groups they are dealing with, the utilization of loan and NPL may be better even if we use the simpler way in paragraph above.

    Anyway, it's an ambitious and exciting project to be done in Indonesia but I don't know enough whether it will works, let alone profitable.

    Enough for now, I wrote too much. :-)

  3. amitz sekali: I think social network is appropriate to tackle problems involving trust.

    As you said, establishing trust is not easy. Instead of establishing new trusted relationships, one can use a friend whom are already trusted. i.e., A doesn't trust D, but A trusts B and B trusts C and in turn C trusts D. Therefore we established a indirect relationship where trust propagates.

    But still we have a mechanism design problem: what is the appropriate incentive structure such that the trust can propagate i.e. B is willing to help A, C is willing to help B and so on.

    This is a non-trivial problem.

    In a transaction, usually only the buyer or/and the seller bear the cost.

    In a 'network transaction', not only the seller and buyer, but also the intermediaries bear the cost.
    (this is why social networking websites such as facebook, friendster, or myspace may not be as useful as eBay).

  4. Ya lu maksud microfinance itu tengkulak ato bank perkreditan rakyat? Kalo konteksnya BPR, kebetulan gue berkecimpung dalam urusan bpr, walo hanya sebatas pengembangan sistem informasi. Gue sih gak tertarik dengan aspek ekonomi, keuangan ato sosialnya.

  5. muara, and later to roby:

    I don't know what exactly you do but from information system perspective, there are many information about poor debitors that indicate their ability to pay off their loan. Even if we don't know yet what information is more relevant, we can later do factor analysis to reveal factors that better predict the paying off of debt. So it's good if the system can store comprehensive entries.

    But what's challenging is how to get those information without unduly burdening the bank's agent in the field. Those important information is probably silly stuff like time between each data entry indicating time between each step of credit request, or the kind of information available and not available. The data mining is gonna be challenging especially if we have to consider the reasoning or common events leading to certain pattern of data..

    The second thing I'd like to mention is that tengkulak has an advantage over institution like BPR because they're not restricted with law. I'm pretty sure that the NPL of tengkulak is quite low even without taking the debtor properties by force, because they can employ many methods like intimidation or creating a ruckuss to make debtor feel malu (shame).

    Which brings to another point. Tengkulak do use culture extensively albeit unethically. They know relationship between family members are strong thus harassing family will make debtor pays. They know some life major events are very important thus debtor will pay before that event. They know if they demand payment in presence of others who are respected, debtor is more likely to pay. Even before lending, they're pretty accurate on identfying which debtor will pay (even if the debtor must borrow to others), given how bad the quality (5C) of debtor candidate.

    Which brings to another point. I think we should capitalize on what we already know about current social structure instead of experimenting with new characteristic. For example, rely on characteristic of peer pressure to make people pay instead of bothering with implementation of transitivity of trust.

  6. amitz sekali: that's a good point on exploiting existing social structure; and yes tengkulak2 are sophisticated-culture-operators.

  7. This comment has been removed by the author.

  8. This comment has been removed by the author.

  9. Micro-finance is not new. IFAD initiated its first micro-finance project in Indonesia w/ BRI (P4K) as early as 1978 (!!)

    One of the major failures of such projects were the pre-determined uniform utilization of loans. (e.g. every member had to utilize loans to purchase, for example, rice seeds. This ran against the wishes of others who wanted to utilize loans to purchase livestock, for instance. In this methodology, the initiative was lost on some farmers - and they were the ones having problems in repaying back the loans)

    New methodologies, assumptions etc were adopted along the way. The Indian experience has been a major source of cross-disciplinary rethinking to improve the schemes.

    I remember a BRI manager in some obscure Kabupaten (Grobogan - Central Java) saying that:

    "The repayment rates of poor P4K farmers are actually better than those middle-class people."

    It appears that repayment rates were better among farmer groups that were given the liberty to utilize credits according to actual NEEDS (e.g. sunatan, fertilizers, school fees) compared to previous projects where the utilization of loans for loans were strictly pre-determined (e.g. to buy seeds, fertilizers, etc.)

    It's of course debatable from an economic standpoint whether paying school fees constitutes a 'performing loan', but it shows that wider indicators beyond economics are needed in order to assess the impact of such schemes.