Relative Portfolio Risk

[Guest]

.

[dokus]

Based on data collected from the KIVA site as off Jan 19, 2009 I figured the relative risks of lending to women vs men, lending to various regions, and lending to various sectors. Why is this important? If you are interested in maintaining a sustainable portfolio of loans to KIVA entrepreneurs, it is important to avoid giving loans to entrepreneurs that will not pay back, hence avoid defaults. If your portfolio runs less risks than the average KIVA portfolio, chances are that your capital can be redeployed many more times, meaning that over the long run it will be able to serve many more entrepreneurs. Hence, all KIVA Friends should be concerned about relative portfolio risks.

To obtain these relative risks I collected data (manually) from the KIVA site on the number of loans paid back so far in each of the categories (gender, sector and region) and each of their sub groups. I also collected data on the number of defaults in each class. I then computed the defaults as a percentage of the loans paid back in full. I also used the distributions of my own portfolio to figure how much less or more risky my portfolio is relative to the overall KIVA experience.

It turns out that based on some 31,996 loans that have been paid back as off Jan 19, 2009 there are 1,830 that have defaulted, which is 5.7%. Note that while this is a low number this is still a considerable higher risk than that acceptable by many MFIs (micro finance institutions, including the Grameen Bank in Bangladesh where the default rate is less than 2%). To avoid a 5.7% default rate one should make smart allocations between entrepreneurs, sectors and regions over time.

Between loans to women and loans to men there is so far very little difference in the KIVA data as off Jan 19, 2009: 5.67% of loans to women default compared to 5.89% of loans to men.

Among the regions the picture is different: relative to the number of loans paid back, the defaults are highest in Africa (13.6%) and in South America (9.9%). The Middle East (1.03%), East Europe (0.57%) and Asia (0.03%) have low default rates. North and Central America have none so far.

Among the sectors the differences are even bigger: While the average loan default is 5.7%, lots of sectors have less than 5% defaults. Two sectors have about 6-7% defaults (manufacturing has 6.3%, and education has 6.9% defaults). The outliner is however agriculture with 17.2% of loans defaulted. Note again I compared the total number of defaulted loans with the total number that were paid back in full in a particular sector, region or gender group. To be absolutely accurate, one would have to add the loans that defaulted to the ones that were paid back in full and then compute the percentage of defaulted loans to the total loans granted (the difference in results with the numbers provided here would however be small given the smallness of the numbers).

To get the relative risk of "my" own portfolio I took the current distributions of my portfolio (e.g. among gender, regions and sectors), which I obtained from a spreadsheet that I developed that provides these distributions,  and multiplied those percentages with the risks percentages obtained from the KIVA data. The summed amounts by gender group, sector and region then allows to compare the risk of a portfolio versus the KIVA experience too date.

When the sums in each group for "my" portfolio came out lower than the averages on the KIVA data, I concluded that "my" portfolio was less risky than the average of KIVA. So, my portfolio is in good shape, although being slightly lower than the 5.7%, still concerns me because it is about 3% higher than what the most efficient and best run MIFs are capable off.

Hope you found this interesting? Feel free to send me a message with your comments.  The provision of up too date, detailed default rates by gender, sector and region (and FP) would be a niece and welcome addition to the KIVA website. All KIVA Friends should become aware of relative portfolio risks if they are interested in maintaining a  sustainable portfolio of loans to KIVA entrepreneurs.

Guido
http://www.kiva.org/lender/guido8966

[wthepoo]

Guido,

first of all, welcome to KivaFriends - I hope you will stay around and post here more frequently.

Thanks so much for your interesting analysis; I have to admit, though, that I think it may be a little flawed in one respect:

The overwhelming majority of defaults on Kiva loans, so far, resulted not from the failure of an individual entrepreneur but were MFI-related (basically mismanagement and/or fraud) or due to the borrowers' death (leading to debt forgiveness with at least some MFIs). It breaks down like this:

total 1,830 defaults

MIFEX, WEEC, SEED (MFI-related): 1,804
Shurush (conflict/war-related): 7
GHAPE (death of borrower): 2

leaves a total of 17 defaults that may be due to the failure of the borrowers' businesses or their unreliability.

Of these, 4 IMHO are a little "doubtful", so far, because there is no explanation and the default happened "long before time".
Zene za Zene: 1
Assasah: 1
Ryada: 1
SPBD: 1

Of the remaining 13 loans that probably defaulted for borrower-related reasons, 7 were very early loans to Africa without the backing of an MFI (East Africa Beta) and 6 were loans with REDC (Bulgaria).


IMHO, this means that borrower-related criteria are - at the moment - not especially helpful in determining the Relative Portfolio Risk.


Just my 2 cents,
best wishes,
Wolfgang.

[dokus]

Hi Wolfgang

Thanks for your feedback. I do not know how you obtained the breakdown of defaults, but even if this is accurate the stats I presented remain the same, given that they are based on what KIVA provides as loans paid back in full and loans that defaulted. I did not make any attempt to find out why loans defaulted or what the underlying causes were. For example, even the death of a borrow would in the case of the Grameen Bank be handled by making the family of the borrower responsible for the payback of the loan. That there are risks associated with various Field Partners or MFIs is clear but so far KIVA provides only a very superficial analysis of those risks. The five star system on field partners is far from a good guide for making selections between borrowers and the MFI from which they obtained a loan. In sum, my contribution is some basic default risks is merely there to stimulate 1/ clients of KIVA to reflect on how they can make their portfolios more sustainable; and 2/ KIVA to start providing much more detailed and more accurate information on defaults and delayed payments. I trust we agree that both are worthwhile objectives. Again many thanks for your input.

Guido

[wthepoo]

Hi Guido,

thanks for your reply.

I do not know how you obtained the breakdown of defaults, ...

Basically by counting; only later I found out that Peter S. (probably in a much cleverer and less time-consuming way) had already compiled the (then current) information here: http://www.kivafriends.org/index.php/topic,2923.msg48110.html#msg48110. This means, though, that my break-down regarding the MFIs with defaulted loans is double checked and should be accurate. I cannot - admittedly - rule out that a couple of the MIFEX, WEEC, SEED-defaults are (also) borrower-related but I seriously doubt that this concerns more than one or two handful(s) of loans.

...but even if this is accurate the stats I presented remain the same, given that they are based on what KIVA provides as loans paid back in full and loans that defaulted. I did not make any attempt to find out why loans defaulted or what the underlying causes were.

Exactly. I trust your statistics are perfectly correct - but as long as they are not taking the reasons/causes of default into account, I suggest that their significance and their value to determine a "real Relative Portfolio Risk" is rather limited.

You are employing Kiva's historical stats for a prognosis of how likely "Active" loans are to default, divided into categories that are basically borrower related (gender and sector; "region" is both borrower and MFI related). One of your conclusions, e.g., seems to be that agricultural loans bear a significantly higher risk than loans to other sectors.  This may actually be true (for rather obvious reasons: impact of weather, infections, vermin...), but you cannot IMHO derive this from the stats when just under 1% of all agricultural defaults had their reasons in these specific risks.

...
The five star system on field partners is far from a good guide for making selections between borrowers and the MFI from which they obtained a loan.

I totally agree - see also: http://www.kivafriends.org/index.php/topic,3147.0.html

In sum, my contribution is some basic default risks is merely there to stimulate 1/ clients of KIVA to reflect on how they can make their portfolios more sustainable; and 2/ KIVA to start providing much more detailed and more accurate information on defaults and delayed payments. I trust we agree that both are worthwhile objectives.
...

Absolutely - the focus for now IMHO should be on the Field Partners, though, and their vetting.

Best wishes, take care,
Wolfgang.

[dokus]

Hi Judy & Wolfgang

Whether most of the defaults are "MFI related" or are "borrower related" is really not the purpose of a note on "relative risks". My main point was sustainability of capital lend through KIVA. Between a KIVA lender (you, me...) and a KIVA entrepreneur there are two different entities. KIVA, setup as a non-profit organization, selects the field partners (FP or MFIs) and provides the peer-to-peer lending capability; and the Field Partners (setup as for-profit or not-for profit organizations, cooperatives, or rural banks) who select the entrepreneurs that are looking for a loan.

If KIVA selects MFIs that perform poorly or end up producing high default rates, then after a while the relationship with these may be "closed", but the lenders who made loans through these MFIs are still going to suffer. Looking at the list of Field Partners you will see that there are currently 110 FPs of which about 15 are now "closed" (some had default rates going as high as 24%). Hence, so far about 14% of MFIs chosen by KIVA, did not work out!  One would hope that KIVA, by sharpening the MFI selection criteria,  will over time manage to reduce this failure rate.

M. Yunus in his newest book, makes a strong case for MFIs to be run like a social business; hence KIVA, which is like a global MFI, may eventually become a social business with  profit objectives so as to be able to pay a low return on capital lend -- just like a large Dutch MFIs already does -- or may adopt the objective of paying back in full the capital that was lend (at present KIVA provides zero interest loans and return less than 95% of the capital provided by lenders).

At the field level, among the MFIs with whom KIVA maintains contacts, there are some that perform very well (some even make profit) and others that don't (according to MIX about 565 out of 1300 of MFIs  around the globe make profit; some of the top ten best make median returns of 35%). The default rates of most MFIs is very low, but the default rate of some of the best MFIs (e.g. the Grameen Bank in Bangladesh) is a fraction of the average default rate currently posted by KIVA!

Out of 33,738 loans that were completed since 2005 there are 1830 loans that have defaulted (i.e. 5.4%) -- a number that should not be compared with any single portfolio because it is an aggregate over all  loans provided over the past four years--.

The reason why 1830 loans defaulted is not really important; lenders who made loans to these loans reduced their capital, and unless they replenish their capital. will end up with less capital to support future entrepreneurs. To be able to continue to support more entrepreneurs in the future, sustainability i.e. 100% replenishment of capital is essential.

In sum,  KIVA lenders who have the intention of supporting more entrepreneurs over time (and do not want to constantly replenish or increase the total capital they make available for KIVA loans) need to be concerned about relative risks.

Guido javascript:void(0);

[JoanW]

I want to make sure I am understanding your point. Since  a lot of the financial stuff is over my head, let me try to make an analogy to see if I get the concept here.

I think what you are saying we don't need to worry if it is raining or if there is a waterfall or a child with a squirt gun...or any of the other possible reasons why. All we need to know is:  "Stand here & you will get wet" or "you have X% chance of getting wet" if you prefer.

Is that what you are saying?

[wthepoo]

In sum,  KIVA lenders who have the intention of supporting more entrepreneurs over time (and do not want to constantly replenish or increase the total capital they make available for KIVA loans) need to be concerned about relative risks.

Absolutely, Guido,
(though not all KIVA lenders will necessarily prioritize this aspect but accept the odd default and rather choose loans by reasons of particular neediness, social impact, relevance etc.)

my only point really was that your statistics IMHO are of rather little value in assessing the risk of particular loans or even portfolios.

Best wishes,
Wolfgang.

[dokus]

Maybe a simple math example can stimulate you to see it in another way: if you started in 2008 with lets say $10,000 in KIVA loans and the default rate is 5.42%, then without any replenishments or additions your capital for lending to entrepreneurs will be reduced to less than $6,000 in ten years and less than $3,500 in twenty years! If the default rate is just 1.5% (which some of the best MFIs achieve), then you would still have over $8,500 in ten years and over $7,300 in twenty years. If the default rate were zero throughout and on top you earned the equivalent of say the US 10 Year Treasury Note, then in ten years you would have close to $13,000 to lend and in twenty years over $16,000! Which would you think would be best for stimulating entrepreneurship around the globe? Note that the median profit that MFIs are making according to MIX is somewhere around 13%. (2006)

[Limesarah]

I get that we should take a look at what loans are most likely to default, but your statistics don't seem to give us a good picture of that, given what we know about the why loans become default.  While agriculture loans have in the past a higher default rate than other business types, this was mainly due to one (or more?) particular large MFI which was found to be corrupt and which had a very high percentage of agriculture loans.  When managed by responsible MFIs, agriculture loans do not seem to have any higher a rate of default than other sectors.  So given past history it makes more sense to ask for better vetting of MFIs and more transparent representation of their policies than to focus on what sector an entrepreneur is working in if you want to maximize the amount of money you'll have to re-loan after a given number of years.  Though of course, a varied portfolio in ways other than MFI is pretty much always a good idea as well. 

[dokus]

The statistics were provided in my earlier posting of yesterday at 5:33 am.  You are right that "it makes more sense to ask for better vetting of MFIs and more transparent representation of their policies than to focus on what sector an entrepreneur is working in". The vetting of MFIs is however dependent on KIVA's selection of "field partners". As clients of KIVA we can only pick among the 94 FPs with whom KIVA currently has contacts. So far there is no information to the best of my knowledge as to where these 94 rank in the global MFI ranking by MIX! I am not sure I understood your point on  "a varied portfolio in ways other than MFI is pretty much always a good idea as well". A well diversified portfolio among regions, sectors and FP's is indeed best, but this does not eliminate the risk of getting stuck with an FP that eventually will be "closed", or with loans that will default. The ideal remains to keep zero defaults, keep the capital intact and keep on re-lending as long as possible. Thanks for your feedback.