What Percentage of Microfinance Loans Actually Go to Business Investment?


If someone with a clipboard came up to you in the street and asked you if you secretly harbor racist views, have stolen things in the past, had unprotected sex, or engaged in some other illicit behavior, how likely would you be to tell the truth?

Probably not very. This causes havoc for any researcher who wants to study behavior that may deviate from social norms in some way. A survey technique called “list randomization” allows researchers to calculate the average response to a question in a population, without being able to identify the response of any one individual. In theory this gives people the freedom to answer truthfully, knowing that even the interviewer won’t be able to tell what they answered.

This method has indeed been used to measure hidden racism and sexism among American voters, as well as all sorts of bad behavior by American teens.

In a paper, forthcoming in the Journal of Development Economics, Jonathan Zinman and I apply this approach to the question of how the poor spend their microfinance loans.

Now, to be clear, we are not for a moment passing a value judgment on how people use their loans, or comparing sneaky borrower behavior to racism or crime. But we do expect from other research that whilst borrowers are heavily encouraged or even required to spend their loan on business investment, often they just want to use it for everyday expenses.

So what we did was this:

We asked clients in Group A how many of these three statements apply to them:

  • I used part of my Arariwa loan to buy merchandise for my economic activity.
  • I used part of my Arariwa loan to buy equipment for my economic activity.
  • I shared my loan with another person.

Clients in Group B received these three statements with one additional statement:

  • I used at least a quarter of my Arariwa loan on household items, such as food, a TV, a radio, etc.

Given that we expect the number of answers to the first three statements to be the same on average in both populations, we can do a simple comparison of means to figure out how many people in group B said “yes” to that final statement.

So we did this with a few variations of questions, and we also asked people these questions individually and directly on a separate occasion, so we can compare the list randomization method with a regular survey.

When asked directly, less than 10% of people admitted to spending their loan on household expenses, health, or education.

Using the list randomization, this number jumped to between 30% and 20%.

Clearly borrowers feel a need to pretend to their loan officers, and even independent surveyors, that they are spending more of their loan on business investment than they actually are.

Research from other studies increasingly suggests that actually microfinance loans do not have that huge an impact on business productivity anyway, and that much of their benefit  is to help smooth out unpredictable income for day-to-day spending. Microloans can have a positive impact even without new business investment or dynamic entrepreneurism. Entrepreneurism is sexy in America. In developing countries, for most, it is synonymous with “I don’t have a job.”

Perhaps it is time for a bit more realism from lenders, and the donors and investors that fund them?


I like that more and more micro-lending is being done to specifically fund household purchases and/or income smoothing. As a first world borrower that is what most of my lending is for.

While I did have a loan to start my business it was quickly repaid from profits, now most borrowing takes the form of extremely short term (21 day grace period!) loans to purchase household goods/food/office equipment etc. or loans with a term of less than one year to buy bigger items (appliances/car/etc).

Or a business line of credit to smooth the cashflow through the fall when I don't make much money but still have expenses - more expneses really since I take vacation time in my slow period too1

I don't see why it would be different for people in the third world.

Joshua Northey

It would be different because to get an acceptable loss rate you need to charge high interest, and the people really cannot afford that interest. They would be better off just living within their means. Microloans are a good idea for a very small portion of the population, for most people it is just another way to fall into "debt slavery".

99% of the people in the world would be happier and have less financial trouble if they took out less debt not more.

Imad Qureshi

Just yesterday I was browsing kiva.org and I did not find even one person who I thought would actually spend money on new business (its my personal opinion). Most people were old with families to take care of. I just don't expect them to spend loan money on a risky new business while their families starve for basic necessities.

Brigitte B

Did your opinion, that the loan would be used for personal uses (in your own words to possible for "basic necessities"), deter you from making a loan? Kiva has a 98% repayment rate so why wouldn't you give it the benefit of the doubt? Even if the loans are used for personal or unanticipated expenses (medical, education, repair etc) isn't that why most loans are made in the US? As long as the loans are coupled with education on how loans, saving and repayment works who cares how the money is spent. Of course, there is room for improvement and the possibility of misuse, but it seems you are unimpressed by good because it's not perfect.

Imad Qureshi

Well, then lets call it charity. My point is to call it what it is.


Except it's not charity if it gets repaid.

Enter your name

School fees are another popular use for "microfinance" loans. Why not just let them pay school fees in installments?


Got a control where the 4th question is innocuous and known to be true only 5% or so?

Ed Kay

Some years ago I saw the results of a poll on whether respondents tell the truth on polls. One area of questioning was reading habits (this was long before the internet). 25% of adult Americans professed to regularly read LOOK magazine every week. But this was seven years after LOOK had ceased publication. There were other results like this.

I'm led to believe that as polls go, this type of poll is relatively inexpensive and easy to do. Yet since all those long years ago, not a single pollster or researcher has been willing to undertake this kind of poll (one where the results are easily verifiable), that would provide an unbiased view of the validity of their methods.

So yes, ridiculous poll results may not say much about the people being researched. But those polls may say a lot about the pollsters and researchers.


The question that should be asked and answered here is "why is it that doctor's waiting rooms are always stocked with really old magazines?"

LOOK and LIFE were great picture magazines, and I still have a stack of them in the basement. I haven't looked at them in a long while, but they were kept because they were interesting to look at and re-read.

So, I can't say that I am a "regular" reader, but, if asked, I would still classify myself as a reader of LOOK magazine, in spite of it having ceased publication long, long ago.

The point is that questions asked in polls can be interpreted differently by different people. How many students/people mess up on tests by not really reading the question and answering what is asked? Quite a few, what? Is it cultural or social background? Is it illiteracy? Is it ADD? Or, is it just mentally bouncing around freely associating (like I am in this comment)?

But, even so, the pollster's job is to eliminate these kinds of anomalies when designing their survey.



The issue with "asking the same question in a different way" surveys is that you will end up with confusing messages. And, trying to apply such simple arguments is plain and simple ludic fallacy.



When did unprotected sex become an "illicit behaviour"? Outside of dystopian novels, and in the absence of any STD's or forced coercion, I thought we could still file "unprotected sex' under "having a good time"?

Patrick McAllister

How does this help MFIs develop more diverse products for their customers? Let's say they use list randomization to discover that 30% of loans are going to places other than their business. Normally we would encourage them to do additional market research to determine which products are most in demand. But if the "borrowers feel a need to pretend to their loan officers, and even independent surveyors, that they are spending more of their loan on business investment than they actually are", how to get accurate product preference information? Could list randomization help?


This article makes the assumption that microfinance loans are to be used only for business purposes.
What is the big deal if they use it for personal use?