FEATURE ARTICLES<strong>MICROBANKING</strong> <strong>BULLETIN</strong>, Issue <strong>20</strong>, September <strong>20</strong>10<strong>Microfinance</strong> Default Rates in Ghana:Evidence from Individual-LiabilityCredit ContractsGerald Pollio 1 and James ObuobieAbstractIn this paper we present evidence on the factors that affect default probabilities in individual-liability creditcontracts. The data are drawn from a for-profit microfinance lender in Ghana. Our sample consists of nearly1,000 randomly selected loans approved between <strong>20</strong>02 and <strong>20</strong>07, three quarters of which were repaid; asdefault is relatively rare in microfinance, borrowers who failed to repay their loans were over-sampled. Westudy the impact of demographic, business and loan characteristics on default odds. We find that repaymentis affected mainly by the number of dependents in the household, years in business, use of proceeds, loanstatus, and frequency of loan monitoring. In contrast to other studies, we find no connection between theborrower’s marital status, gender or their savings’ behavior and the likelihood of default.<strong>Microfinance</strong>, however measured, has increasedrapidly in Ghana since the start of the present decade,growing by <strong>20</strong>-30 percent annually. <strong>Microfinance</strong>Institutions (MFIs) currently provide financialservices to an estimated 15 percent of the country’stotal population compared with 10 percent for thecommercial banking sector.Rural and community banks account for the lion’sshare of MFI activity in Ghana, representing more thanhalf the total number of microfinance borrowers anda similar proportion of the sector’s total loan portfolio(Aryeetey:<strong>20</strong>08). NGOs, by contrast, are comparativelyunimportant: the average loan size is roughly onethird that provided by rural and community banksand an even smaller fraction (25 percent) of theamount borrowed from savings and loan companies.On the other hand, loan repayment rates, at a reported99 percent, are considerably higher among financialNGOs than among other microfinance providers orgovernment-sponsored lending programs; their loanloss exposure is also relatively modest.Finally, and perhaps not too surprisingly giventheir heavy dependence on donors or officialsources of finance, financial NGOs have the worstrecord of achieving either operational or financialCorresponding author: LSC London, Chaucer House, White Hart Yard,London SE1 1NX, United Kingdom.self-sufficiency, surpassed only by governmentsponsoredprograms. The clear implication here isthat without substantial subsidy, interest rates onloans provided by both NGOs and state-supportedinstitutions would be significantly higher, withan attendant negative impact on repaymentobligations.Ghana’s commercial microfinance operations have,by contrast, broadly achieved a degree of operationalefficiency that compares favorably with mediumsized African financial institutions or worldwide MFIs;even so, in terms of financial sustainability, many stillhave a long way to go. The purpose of this paperis to investigate repayment rates among MFIs thatfollow the individual-liability lending model; sincethis model closely approximates to that pursued bycommercial banks, a close correspondence might beexpected between the two approaches. Our analysis derives from data on loan repaymentsand borrower characteristics provided by ProCredit(Ghana), a local microfinance institution, whichoperates on the basis of individual liability lending.Jha, et al. (<strong>20</strong>04) and Bank of Ghana (<strong>20</strong>07) provides an overview ofmicrofinance in Ghana.Owing to space limitations it was not possible to provide the full setof results. Readers interested in obtaining our findings can email thesenior author at the email address given above.<strong>Microfinance</strong> <strong>Information</strong> eXchange, Inc
<strong>MICROBANKING</strong> <strong>BULLETIN</strong>, Issue <strong>20</strong>, September <strong>20</strong>10Individuals and micro-entrepreneurs that apply forProCredit loans proceed through three stages prior toobtaining approval.i. Preliminary ScreeningIn this stage, loan applicants make contact withthe institution and are carefully screened andasked to answer specific questions regardingthe status of their business and householdaccounts, in order to establish whether theyqualify under ProCredit‘s eligibility guidelines.ii. Loan Proposal and Credit CommitteeLoan applicants are assigned to specific loanofficers. Applicants undergo a further reviewto verify the information taken at the initialstage, and a visit to the applicant’s businessesand household is arranged. The informationthus developed is organized into a formalloan proposal and presented to the lendinginstitution’s credit committee for approval.The loan amount and tenure are confirmedconditional on the adequacy of the cash flowsgenerated by the borrower’s business, sufficientpersonal collateral and guarantors agreeing toco-sign the loan agreement.iii. Monitoring and RepaymentAfter disbursement, the account officerfrequently visits the borrower’s business toensure that the proceeds are being used forthe specific purpose(s) for which the loanwas granted, and to remind borrowers oftheir next repayment date. Borrowers whomiss payments are pressured at this stage; ifthe arrears continue, legal action is initiatedagainst the borrower and guarantor(s) torecover any amounts owed, but usually afterthe designated collateral has been seized andliquidated.1. Loan Sample DataThe data used here are drawn from ProCredit’slending files. Six of the bank’s twelve branches,including one located outside the national capital,were selected for the study; the remaining brancheswere all established fairly recently and thus haverelatively small loan portfolios.The total sample consists of 960 loans made to localbusinessmen from the database of the institution’s sixbranches, five of which are located in Accra and onein Kumasi. The sample consists of loans granted andFEATURE ARTICLESrepaid (or not) between January <strong>20</strong>02 and December<strong>20</strong>07, and comprise 160 loans from each branch. Thesample thus consists of 7<strong>20</strong> repaid and 240 defaultedloans, with individual loans in each categorychosen randomly. To obtain a fair representationof the characteristics of defaulted borrowers, wedeliberately over-sampled this category, a decisionmotivated by the low actual default rate. The sampledataset was audited for errors and omissions toensure consistency and uniformity.Twenty-four borrower characteristics were extractedfrom the data and grouped into four main categoriescross-classified by borrower status (Table 1).(1) Individual borrowers’ household characteristics(gender, age, marital status, household income notgenerated from either the business or earnings ofdependants). Dependants consist of the number ofpeople in the household who rely on the businessincome. Households with fewer dependants have asmaller claim on their business income, which shouldserve to reduce the default rate. The borrower’smarital status is also expected to lower the likelihoodof default; working spouses generate an independentincome, thus increasing the financial resourcesavailable to service the loan, in contrast to borrowerswho are single, divorced or widowed, where thereare no supplementary earnings. We also expect theprobability of the loan being repaid to increase ifthe borrower is a woman, in keeping with empiricalevidence to that effect.(2) Savings behavior (default and non-defaultborrowers’ saving behavior during the term of theirrespective loans). Borrowers who save during the termof the loan build up a cash reserve that can be used toservice the debt during periods when the business isfacing liquidity difficulties. The presence of savings isexpected to increase the probability of the loan beingrepaid.(3) Business characteristics: (business type, age ofthe business, location of the business [Branch]).The number of years the borrower has been in thesame business should increase the probability ofthe loan being repaid; there is again ample evidenceshowing that established businesses are less proneto experiencing financial distress than are newlycreated ventures.(4) Loan characteristics (loan amount in Ghana cedis,loan purpose, loan monitoring, collateral type andvalue in Ghana cedis, term of loan, loan status, numberof guarantors [co-signers]). Each loan is unique in<strong>Microfinance</strong> <strong>Information</strong> eXchange, Inc