Wednesday, June 15, 2011

A Potential Model for Islamic Microfinance IV: First Working Paper posted

It has taken a long time to code and analyze the data. I have finally posted the first working paper (the paper is joint with Mohamed El-Komi, Dean Karlan, and Adam Osman) from the research program that I had summarized earlier here: Bank-Insured RoSCA for Microfinance: Experimental Evidence in Poor Egyptian Villages. Feedback welcome.

12 Comments:

Blogger Muhammad said...

Dear Professor El-Gamal,

Salam - this is a fascinating paper with really promising results from the perspective of the borrowers in terms of take-up and repayment. I was wondering about the incentives from the perspective of the bank making the guarantee? You mention that the fee is small and is actuarially fairly priced. In practice, given the required administrative/monitoring costs for the guaranteeing bank do you think that the fee ā€˜sā€™ may rise to a high enough level such that it significantly affects the predictions of the model? Would you envisage the bank actually making a profit/breaking even on the fee and their overall participation or would you expect this function to be performed by a non profit organisation or subsidised by external funding e.g. world bank?

9:55 AM  
Blogger Mahmoud El-Gamal said...

The main point is for this to emerge (our next phase, a file experiment) into a credit union, which is non-profit by definition. If the collected premia exceed losses, any profits are to be distributed to shareholders/members pro rata as dividends.

12:45 PM  
Blogger Mahmoud El-Gamal said...

I meant to write "a field experiment". By the way, I am aware of an "Islamic credit union" in Afghanistan, but I do not know how it structures its operations. If anyone has information about it, please let me know:

http://afghanistan.usaid.gov/en/USAID/Article/1225/Aqcha_Islamic_Investment_and_Finance_Cooperative_Turns_Three

12:47 PM  
Blogger ibster31 said...

Salam Dr. El-Gamal,

My name is Ibraheem Catovic, and I am a rising senior at Rutgers University's Honors College. I have read a number of your works, and this one is promising and potentially groundbreaking, Masha'Allah.

Just one question. For the Grameen layout, you mention: "It is also noteworthy that the dichotomy for good vs. bad equilibrium does not depend on p (page 9, par.2)." You later mention that the equilibrium depends on p only in the RoSCA layout. Looking at the Grameen tree (left branch, t=2, 4 blue boxes), it seems that if player 1 repays, player 2 doesn't repay and then he defaults, players 2ā€™s payoff is 2+2r-p. His good equilibrium payoff is 1+2r-i, which player 2 gets if he repays. Then, it would be better for player 2 to default if p<(1+i), meaning the equilibrium does depend on p.

Your results also seem to show that in the Grameen game, as p went up, the good equilibrium percentage also increased (Table 4). Let me know where my logic fails, or if I have misunderstood something.

I would appreciate your clarification. Thank you.

1:32 PM  
Blogger Muhammad said...

Thank you for the reply, inshAllah if this works it can be the beginning of something very significant. In particular, given the recent incidents in Andhra Pradesh and very small results from impact evaluations of traditional microfinance, it could be a viable alternative for non muslims and the beginnings of a small change in the direction for the industry away from the trend towards profit maximising models looking to make money for shareholders towards a more co-operative model where the unambigious objective is to benefit the communities and not some fuzzy 'double bottom line'

12:13 PM  
Blogger Ahmad.Munir.Blogger said...

I've a question. You describe riba as unbundled "sale" of credit. But is credit sold? When a lender lends money on interest, he gets the principal lent and interest back. When you sell something, like your car, you don't get it back. Compare it with unbundled sale of risk (your description of gharar). The seller of risk does not get the risk back. How do you then explain the concept of unbundled sale of credit?

9:58 AM  
Blogger Mahmoud El-Gamal said...

When you lend, you do not get the "credit" back either. You get the price of credit, which is interest (that depends on credit risk, period of credit extension, etc.).

11:10 AM  
Blogger Mahmoud El-Gamal said...

A monetary loan is renting money, and interest is the lease component. Hanafi jurists see all rent as the sale of usufruct of the property. In this case, the property is money, and renting it is equivalent to selling its usufruct (measured by the time value, benchmark risk less rate, etc). Because the lessee may not return the lent/leased money, one is selling him credit (that is the credit-risk premium over time value). I hope that this makes it clearer. For further discussion of the relationship between loan, `ariya, and lease, see my discussion of Al-Qarafi's distinction in his Al-Furuq, available in this paper: http://www.ruf.rice.edu/~elgamal/files/mutualize.pdf

3:39 PM  
Blogger Ahmad.Munir.Blogger said...

I've re-read your paper on mutuality (a good paper it is) and thought things over.

I am still of the view that lending money on interest is unlike any sale of good and service. In a sale, (a) you don't get back what you sold with some excess and (b) the ownership changes hand.

To your point about lessee, the lessee may not even return a real asset given to him on an operating lease, say a car. That does not turn the operating lease of a car into a sale.

In a nut shell, lending money on interest is not "sale of credit" because that's not what a "sale" is. Therefore, based on this reasoning, riba cannot be "unbundled sale of credit."

6:58 AM  
Blogger Asif said...

I really enjoyed reading this paper. The work is very comprehensive, and it is really nice to see a summary of this new thrust beginning from theory, educated by experiment, and tested by application. Has this been prepared for publication? If this is still in progress, however, you might consider splitting it into two papers with different themes.

Paper 1: a game-theoretic modeling of insured-Rosca as compared to the Grameen approach and as tested by the experiments.

Paper 2: a description of the field implementation of the insured-Rosca as applied to a real situation in the Egyptian context.

In the first part, the comparison with Grameen is very interesting. Although the Rosca mechanism has been published before, I still had to read the paper by Besley et. al. to get a plain description of what a Rosca actually is. The comparison on the basis of a game-theoretic formalism is useful, and particularly the idea of the Rosca having a larger 'basin of attraction' (my words) for the good outcome. The game presented is quite complicated, and the diagram is only suggestive of all the detail. It would be helpful to add some commentary, for example, where the essential difference in decision-making between the two schemes -- given the effort taken in setting up the game formalisms. In fact, the game description is clearer in the description of the experiment and not necessarily in the decision tree format. Perhaps a nomenclature section would resolve this, though it might be useful to present the experiment along with the decision tree so that they reinforce one another and to further tie the theoretical treatment with the experiment. Some useful questions in the theory aspect are as follows: Is there a tipping point where you go from a Nash equilibrium of unmet credit needs to sufficient credit for the community to eliminate attendant inefficiencies in resource allocation?

Somehow, the comparison to the Grameen approach (which seemed to be a primary theme of the theoretical/experimental analysis) is only mildly interesting in the field trial, perhaps as an experimental control though not necessarily the only or best comparison case. Far more interesting (to me) were the specifics of the implementation, the adaptations made for the Egyptian context, and the summary of how this new construct of an insured-Rosca actually performs. In fact, I would have liked to see more focus on the sociological aspect here, and although it is always nice to see hard data analysis, it sort of seemed like the story told in the data analysis seemed like only a fraction of what was observed and learned in the trial. For instance, there were several adaptations made for the local context (i.e. rural Egypt). What are these adaptations and how useful were they? What factors would drive additional or alternate adaptations (multiethnic context, income level, macro-economic factors, climate, etc.) What level of credit is necessary for such a community, and what fraction of it do insured-Roscas supply? Of course, these questions are not as amenable to statistical analysis as some of those answered, though I think that with field data it is appropriate to address them and present whatever evidence was collected to advance these questions.

1:30 AM  
Blogger nhusain said...

According to VA law:
As a defense against payment, a borrower can plead that the interest charged on a contract
was greater than allowed by state statute (i.e., usurious). If the court determines that the
contract is usurious, it will issue a judgment for payment of the principal sum only.

11:57 AM  
Blogger nhusain said...

The court can also intervene and halt any foreclosure action on property which is the
subject of a suit or action for excessive interest.

A creditor who can prove that it imposed or collected interest or other charges in excess
of those permitted as a result of a bona fide error in computation or similar mistake will
not be liable for penalties, but must return to the borrower the excess interest or other
charges collected.

12:04 PM  

Post a Comment

<< Home