P.V. Viswanath
Rotating Savings and Credit Associations (ROSCAs)
are in the middle of a spectrum
 Informal
loans from family/relatives on one end and
 (relatively formal) moneylenders on the other.
Explains how group lending in microfinance works:
 How
groups can help to reduce costs
 Mobilize funds
 Improve monitoring
 Deploy informal community-based enforcement
Basic Structure
A group of individuals agree to regularly contribute
money to a common “pot” that is allocated to one
member of the group each period.
For example, 20 people may agree to contribute $15
each for 20 months, generating a monthly pot of $300.
At monthly intervals, the group meets to collect dues
and allocate the proceeds, with past recipients
excluded from getting the pot again until every member
has had a turn with the $300 pot.
Pot recipients can be chosen through a lottery or
participants may be allowed to bid to get the pot, with
profits shared by the entire group.
Basic Structure
ROSCAs are, thus, also in the middle of another
 Borrowing
(lumpsum inflow first –used for spending –
and then periodic outflows afterward)
 Saving and spending (periodic outflows first and then
lumpsum inflow afterward – for spending).
They take bits of surplus funds that come into
households and translate those bits into a large
chunk that can be used to fund a major purchase.
Clear beginning and end of the enterprise
Accounting is straightforward (one only has to keep
track of who has received the pot already and who
is in line to do so)
Storage of funds is not required since the money
that is collected each period goes directly to that
period’s recipient.
Neighborhood institutions – everybody knows
everybody else.
ROSCAs have been around for a long-time and the
proof that they are very useful and solve financial
problems for social groups is in their ubiquity.
Tontines in rural Cameroon
Hui in Taipei
Tanda in Mexico
Polla in Chile
Chit funds in India
Arisans in Indonesia
Loteri samities in Bangladesh
Kye in Korea
Susu in Ghana
Esusu in Nigeria
Upatu or mchezo in Tanzania
Chilemba or hiperegani in
Merry-go-rounds in Africa
How a ROSCA works – without
Suppose participants wish to buy a machine for $300 –
assume the money is needed in a lumpsum; hence
individuals need to wait until the entire sum is available
before buying the machine.
Assume monthly income is $50/month without the
machine, but $70/month with the machine.
$35/month are required for subsistence needs;
$15/month can be saved to buy the machine.
Without a ROSCA, the participant needs 20 months
before buying the machine and doubling her income.
How a ROSCA works – with
Suppose there is a ROSCA with 20 members and a
monthly contribution of $15.
Suppose the order of pot recipients is determined
through a lottery.
The probability of getting the pot in any given week is
1/20; the expected number of weeks to wait is 10.
On average, then the participants can increase their
consumption by $20 ten weeks earlier than without a
The worst case is the person who gets it last; this person
does not wait any longer to get the pot, than if she did
not participate in a ROSCA. Hence this ROSCA benefits
every participant.
ROSCA motives
The example earlier clarifies the “early pot” motive
to participate in a ROSCA.
There is also the “household conflict” motive;
participants, often women, seek to get money out of
the household and away from their husbands.
Then there is the “commitment to savings” motive –
ROSCAs present a clear, public, disciplined way to
accumulate funds.
What are funds used for?
Assumption 1. All individuals wish to buy an indivisible
durable good. If this were not so, then individuals can
buy the good in installments, as they save their money.
A study in Taiwan (Besley and Levenson, 1996) showed that
ROSCA participants are more likely to buy durables like
microwave overns, air conditioners etc than others.
 A study in Nairobi (Anderson and Baland, 2002) showed
that ROSCA participation is correlated with lumpy
purchases (school fees, clothing, rent, medical costs etc.)
 However, other studies (Gugerty, 2007) show that ROSCA
expenditures are often divisible – school fees can be paid
in instalments, food can be purchased in small quantities
When do they want the funds?
Assumption 2. They are impatient to do so – they
wish to buy sooner rather than later.
 Gugerty
(2007) shows that sometimes it’s not the speed
of access to the pot, but rather getting the pot when
there is a large expense, e.g. having funds available
during the harvest season to pay for harvesting
expenses and to be able to afford to hold the grain
until later when grain prices are higher.
This suggests that, more than the early pot motive,
ROSCAs are useful because they provide an
effective way to save.
Effectiveness of ROSCAs: Incentives
Assumption 3. ROSCA participation is enforceable –
those who win the pot will continue to contribute to
the pot until everybody has had a chance to get the
pot and purchase the durable good.
Q: In a random ROSCA, why should the last person
in line stay in the ROSCA? She has no benefit and
loses the flexibility of determining her own savings
Ans: Participants might not have alternative
effective ways to save.
Effectiveness of ROSCAs: Incentives
Why should early pot recipients stay in the ROSCA and
continue to contribute?
These are just like borrowers and the ROSCA has all the
enforcement problems with these people as all lenders
do with borrowers in more familiar arrangements.
One answer is to refuse absconders access to future
pots, i.e. it’s a multiperiod game. However, if the order
of pot assignment never changes, then the participants
are no worse off than going off on their own – they will
have to wait 20 months before the getting the next pot.
Effectiveness of ROSCAs: Incentives
Suppose the order is determined afresh at the end of every cycle.
This improves incentives for the participants who get the pot after
half the life of the ROSCA.
However, this makes the situation even worse for the first half of the
For example, the 9th recipient would have to contribute for another 20
months and would get the pot again in the 29th month if the order
remained constant, for a total of 20 more months before getting access
to the pot again.
If the order were refreshed, she would have to wait 11 months for the
current cycle to end and then, on average, wait another 10 months for a
total of 21 months, on average.
Similar arguments could be made for all the earlier recipients.
For the 11th person to get the pot in the first cycle, the average wait
drops from 20 to 19; similar arguments can be made for all the later
Alternative assignment mechanisms
Random assignment of the order of getting the pot
is considered fair even though the incentives to
cheat for early recipients are greater.
Alternatively, a fixed order could be used with
more creditworthy people going first.
Social sanctions could be used to ensure continued
payment into the pool – e.g. access to trade credit,
material inputs etc., or ostracism and exclusion from
social/religious events in the village.
Bidding ROSCAs
The pot assignment in a fixed order or a random
ROSCA is unrelated to the need of the recipient for
the pot. One way around this is to allow
participants to bid for the pot.
However, risky participants will be willing to bid
more, since they will actually pay less often. Studies
find that default rates in bidding ROSCAs are
One way around this is to limit the size of the bids
and to institute some kind of credit screening.
We have seen that one motive for ROSCA participation
is to have an effective way to save. ROSCAs only allow
for raising resources from the borrowers, themselves
and assume that savers are also borrowers.
Credit Cooperatives provide a way of including savers
who don’t necessary want to borrow to join the group.
Credit cooperatives are formal versions of ASCAs
(Accumulating Savings and Credit Associations).
The main advantage is that savers are no longer
required to borrow and the size of loans can vary with
need, but the disadvantage is that funds must now be
stored, and bookkeeping and management become
more complex.
Village Banks
One of the important characteristics of ROSCAs and ASCAs is
that the group is small, and the members of the group know
each other. This facilitates screening, monitoring and
Village Banks share this aspect. John Hatch in 1983 hit upon
the idea of making block loans, using USAID funds, to
cooperatives of about fifty villagers, instead of five or seven.
The group, not an outside banker, would allocate the credit to
members and keep the books. All could borrow immediately.
The banks would take savings, as well, working in an
informal setting using members to keep an eye on the use of
This technique has been used by the US nonprofit Foundation
for International Community Assistance (FINCA) and the
Mexican lender Comportamos.
Self-Help Groups (SHGs)
These were created in India in the mid-1980s by the Mysore
Resettlement and Development Agency (MYRADA).
These groups were homogenous, being united by caste, cred,
ender, profession, economic status, or kinship. Again, the aspect
of inexpensive self-monitoring can be seen.
SHGs encourage members to save into a group fund, rather than
obtain capital from outside.
Groups collectively decide on any lending to members.
These groups also serve as conduits for other services, such as
training in irrigation and agricultural practices.
Under the auspices of the National Bank for Agriculture and Rural
Development (NABARD), a semi-governmental apex bank, SHGs
have also borrowed from banks, after accumulating their own
capital base and establishing a track record of regular
Credit Cooperatives
One of the problems with ROSCAs and many of their
informal and semi-formal cousins is that loans are of short
maturity and repayments are frequent and often start soon
after the loan is made.
Such a credit schedule is convenient for shopkeepers and
other small businesses, whose inventory turns over rapidly.
However, for some borrowers, such a schedule is impractical.
For example, farmers usually have no cash inflows until the
harvest, which may only be once or twice a year. They also
need to make capital investments, which may not be
recouped for several years. A similar situation obtains for
Credit Cooperatives
Friedrich Raffeisen, in Germany, in 1864, started credit
cooperatives. Some key aspects were:
Members should belong to the same local parish.
Members would have unlimited liability to support borrowing from
outsiders. Defaulting members would lose their current assets, as
well as suffer social costs.
Low-income individuals could not be discriminated against and
should be given equal rights as members of the cooperative.
The cooperative also facilitated the purchase of inputs of production
for its members.
Members were to deposit savings as they repaid their loans, so that
the cooperatives could eventually free themselves from outside
The terms of loans, which were often made for farm investments,
could extend to two years or even more.
How Credit Cooperatives work
Consider a cooperative with two members. One of the two has an
investment opportunity and needs financing. The project is risky and
yields y with probability p and zero with probability 1-p.
The investment requirement of F can be financed partly by borrowing
from an outside lender and partly from an inside lender.
First assume that both members have zero wealth. The loan contract
specifies an amount b (= F in this case) lent and a gross interest rate
R, with R.b < y, which is paid if the project succeeds, and not
Suppose now the insider lends F-b and the outsider, b. The insider
acts as guarantor, possibly offering collateral that would secures the
outsider’s loan.
In this model, the inside lender has an incentive to monitor the (inside)
How Credit Cooperatives work
Since the borrower’s effort increases the probability of
success, for convenience, let p denote both the borrower’s
effort and the probability of success.
Let m denote the intensity of monitoring. The cost to the
borrower of exerting effort is given by 0.5(1/m)p2. The
marginal cost of effort is increasing, as indicated by p2, but
is decreasing in m. This latter point can be understood by
thinking of effort as the opposite of shirking – the higher the
level of monitoring, the costlier is shirking, for the borrower.
The timing is as follows: first the borrower contracts with
inside and outside lenders; then the inside lender choose
how much to monitor; then the borrower decides how much
effort to invest; finally project revenues are realized.
How Credit Cooperatives work
The borrower chooses effort, p, to maximize p(y-Rb) - 0.5(1/m)p2.
The optimal level of p is seen to be m(y-Rb), for a given level of
monitoring of m. Substituting this into the maximand, we see that
the gross return to the borrower, is (m/2)(y-Rb)2.
Since the benefits of monitoring would go primarily to the outside
lender, why would the insider monitor, and how much?
Suppose the inside lender has wealth w, that she uses as
collateral for the outside loan.
If the project fails, the insider loses her collateral. Hence she
would have an interest in monitoring the borrower.
Clearly the provision of collateral would reduce the interest rate,
R, that the outside lender charges. Hence the greater the amount
of monitoring, the greater the surplus available to be shared by
the insider lender and the inside borrower.
How Credit Cooperatives work
Since the insider would be able to monitor the borrower
cheaper than the outsider, this setup benefits all parties –
the outside lender and the two insiders.
How much monitoring will occur depends upon the contract
set up between the inside borrower and the inside lender.
It is, however, entirely possible that insiders will monitor too
much and punish borrowers too often relative to the social
In fact, in other social group borrowing contexts, such as the
Grameen bank group lending model, it has been
suggested that monitoring and punishment is excessive. This
might have led Grameen Bank to move away from its
group lending format.

similar documents