“Please don’t remove the loan officers from the credit cycle!” exclaimed the CEO of a Tunisian MFI during a presentation of MUSONI’s concept of “the first 100% mobile microfinance institution” made at the Convergences2015 event in Paris this September.
Musoni BV used its knowhow in mobile money transfer services to establish Musoni Kenya - a cost efficient MFI. Musoni Kenya leverages on the success of existing mobile money transfer services like Safaricom’s M-PESA to process all loan disbursements and loan repayments over the mobile phone. The concept supposes the elimination of the cash from the entire loan cycle and with it the fatty chunk of costs that entails cash management. The successfully implemented idea won the “Most Innovative Use of Technology” award during the 2011 Microfinance Investment Summit in Geneva.
According to Cameron Goldie-Scot - Musoni’s COO - the MFI has 8,000 clients and disbursed over 20,000 loans using mobile money transfer services. Musoni BV is now looking to establish a second MFI in Uganda. The company is offering its innovative technology platform to non-Musoni MFIs.
Hyperinflation can easily transform your wallet to a purse, then to a plastic bag and eventually to a sack. Although not every micro borrower needs a suitcase to carry home the loan, most envy the Eurozone residents who have the 500 euro bill that makes counting and handling cash (including dirty money) safer. It is obvious that handling bills and coins generates costs to the borrower and to the MFI. The former has to put together the money required to repay the installment, bring it to the branch, queue and then deliver the money to the cashier. The cashier has to count the money, make sure no fake bills were squeezed-in, eventually report each counterfeit, store the cash in the till at first, than transfer progressively the surplus to the vault, from where the bills continue the migration. As cash changes hands, mandatory receipts have to be generated, signed and then stored in a different place. In addition, each such operation bears the risk of robbery and miscount.
Musoni BV used its knowhow in mobile money transfer services to establish Musoni Kenya - a cost efficient MFI. Musoni Kenya leverages on the success of existing mobile money transfer services like Safaricom’s M-PESA to process all loan disbursements and loan repayments over the mobile phone. The concept supposes the elimination of the cash from the entire loan cycle and with it the fatty chunk of costs that entails cash management. The successfully implemented idea won the “Most Innovative Use of Technology” award during the 2011 Microfinance Investment Summit in Geneva.
According to Cameron Goldie-Scot - Musoni’s COO - the MFI has 8,000 clients and disbursed over 20,000 loans using mobile money transfer services. Musoni BV is now looking to establish a second MFI in Uganda. The company is offering its innovative technology platform to non-Musoni MFIs.
Hyperinflation can easily transform your wallet to a purse, then to a plastic bag and eventually to a sack. Although not every micro borrower needs a suitcase to carry home the loan, most envy the Eurozone residents who have the 500 euro bill that makes counting and handling cash (including dirty money) safer. It is obvious that handling bills and coins generates costs to the borrower and to the MFI. The former has to put together the money required to repay the installment, bring it to the branch, queue and then deliver the money to the cashier. The cashier has to count the money, make sure no fake bills were squeezed-in, eventually report each counterfeit, store the cash in the till at first, than transfer progressively the surplus to the vault, from where the bills continue the migration. As cash changes hands, mandatory receipts have to be generated, signed and then stored in a different place. In addition, each such operation bears the risk of robbery and miscount.
McKinsey & Company in 2007 estimated that the cost of handling cash for a retailer is 1.3% of the purchase price. For a cash-based financial institution this percentage could more than double since cash is handled at disbursal first, then when principal and interest is collected in installments. Obviously a big part of collected money is disbursed in new loans in the same day, shortening the costly circuit of money, but 1.3% of each handled amount could be the correct estimation of the cost of cash for the MFI. This cost ultimately is transferred to the borrower, who pays a higher interest rate. In net present value terms, if imputed to the micro borrower, the cost of handling the cash in the MFI is responsible for increasing the cost of borrowing by 6 - 7%[1]. To this we should add the cost of handling the cash for the borrower himself. In addition to the opportunity costs born by the borrower when he comes to the branch and waits in line to repay his installment, we should include the cost of the risk he takes when carrying the cash to and from the MFI.
Rosenberg et al. in a 2009 CGAP paper[2] estimate administrative costs - the largest single contributor to interest rates in an MFI - at about 11% of the loan portfolio. If cash were to be eliminated, administrative costs can drop by 2 x 1.3%. The annual interest rate can drop by 2.6% in consequence. Mobile banking can help reduce the cost of borrowing for the poor as long as the fee per mobile transaction is significantly below the cost of handling cash and infrastructure allows a convenient use of dematerialized money.
Khan and Ashta[3] found strong positive correlation between the cost per borrower and average loan balance for MFIs in Bangladesh. The cost of handling more cash can possibly explain a part of this diseconomy of scale.
What about the cost of the loan officers? They play a key role in micro lending as they select the borrowers and manage the portfolio. Many attribute the success of microfinance to the close personal connection between the loan officer and the client. Musoni also places the loan officer at the core of its financing model.
My question is: Can Musoni through technology improve significantly the efficiency of the approval process to further reduce the cost of borrowing for its clients? Grameen bank was amongst the first to challenge the status of the loan officer by ‘subcontracting’ to the self-help groups of borrowers the procedure of screening newcomers and enforcing loan collection, but this works for group lending only. Musoni’s borrowers use mobile phones to process the repayments, automated SMS are generated to improve communication with the clients, and loan officers use tablet devices to operate independently from the branch. In the future, algorithms and computers could probably replace the loan officers in financing standard applications for individual micro loans received through mobile phones.
Transaction cost can be reduced. On the applicant’s side, the loan demand can be made simply by sending an SMS containing the information about the requested amount and maturity. The applicant can do that without leaving his store or his workshop, or at home after sending the kids to bed. There is little cost for the MFI also. No need to keep the branches open for applicants or send the loan officers to prospect the neighborhoods.
The question is how to assess the creditworthiness of the applicants with little implication from the loan officers? IT and credit scoring give the answer, but where can the MFI find relevant data and make sure it is reliable?
For repeat borrowers the procedure is simple. The lender already knows how they behaved when repaying previous loans. Past credit behavior is by far the most predictive information of future repayment behavior. In the same spirit credit bureau reports can help screen applicants based on their past credit behavior with other institutions, but examples of credit bureaus that cover extensively micro borrowers are rare. Cameron Goldie-Scot added that Musoni Kenya played a key role in setting up a credit bureau for micro borrowers in partnership with AMFI - the local body representing Kenyan MFIs.
For new applicants the question is more delicate. Credit scoring algorithms usually use socio-demographic, business-demographic, financial data and loan characteristics to generate a score that estimates the probability of the loan to go delinquent or to default. Loan officers are the ones that collect this information by visiting the applicant at his workplace or home. Micro borrowers usually operate in the informal or semi-formal sector, so their capacity to generate or provide reliable documents that would make the visit redundant is limited. The cost of evaluation of the loan application is high in microfinance. Each client covers the cost of appraisal of his application and also contributes to offsetting the costs of assessment of refused applications.
The mobile phone company can provide some valuable statistics about the use of the mobile services by the potential borrower and his payment behavior with regard to these services. How many different contacts do you usually call in business and leisure time can indicate about the depth and width of your business and social network. Do same contacts call you back or others? Do you have friends or associates abroad?
How often do you charge the mobile phone account is also an indicator of your capacity to make money and eventually repay loans. Is your mobile phone subscription new or old, or it is pre-pay, could also be a valuable information. Through data mining, one could identify patterns of use of mobile phone services specific to creditworthy borrowers and ‘bad’ borrowers. These patterns are incorporated in credit scoring algorithms that can estimate the creditworthiness of the applicant with certain accuracy.
For more precision, the applicant may be required to provide more information about his family, business, financial position, etc. A smartphone application would be more convenient for answering questions, but the same procedure could be considered using SMS or mobile chat – services supported by all mobile phones and most providers of mobile phone services. Such information would be self-declared and thus prone to manipulation, but relevant statistical links could be made in order to get more accurate estimations of the credit risk. All this makes sense if the long term costs of the new IT infrastructure will be below the current costs.
Replacing the loan officer with a digital interface raises some business and ethical questions. The removal of the human interaction with applicants and borrowers could reduce the knowledge about the clientele. Many MFIs rely upon the loan officer to provide additional services such as training and assistance to micro-entrepreneurs in building successful businesses. From an ethical point of view, IT can make jobs redundant and this doesn’t go along with the microfinance mission. On the other hand, high interest rates keep away many micro borrowers or catch them into a survival trap where increased revenues are spent on interest and loan related fees. Minor personal or professional shock can lead them to overindebtedness. It could be that reduction of the interest rate by few percent points is translated in more job creation than losses on the MFI side.
It is still a taboo issue, but often loan officers are responsible for fraud. Some approve all loan applications to cash the variable bonus and leave before the portfolio deteriorates. Some look for potential borrowers willing to default and share the loan amount. Some charge the borrower a ‘personal’ fee of few percent for making the disbursal possible. The higher the risk of the borrower, the higher the fee is. This fee can increase further if the interest rate is subsidized or part of disbursed money is grant.
In theory, provided a strong implication of the stakeholders, the partial exclusion of the loan officer from the credit cycle is possible and often necessary. Experienced loan officers will be always required in the MFI, as their skills are crucial in assessing non-standard loan application or in launching new loan products and providing additional services such as coaching for clients. The question is where to put the cut-off between the loan officer and the machine and how to move it progressively towards the less expensive resource?
[1] Simulations based on a 12 months loan of 300 euro, bearing simple annual interest rates of 12% and 24%. 1.3% of the amount of the transaction (at disbursal and at repayment) is added as a fee.
[2] Rosenberg, R., Gonzalez, A. and Narain, S. (2009) The New Moneylenders: Are the Poor Being Exploited by High Microcredit Interest Rates, CGAP.
[3] Khan, S. and Ashta, A. (2012) Cost control in Microfinance: Lessons from the ASA case, Cost Management, 26(1).
Vitalie Bumacov
associate researcher with the Banque Populaire Chair in Microfinance at the Burgundy School of Business (Groupe ESC Dijon-Bourgogne), France and PhD candidate at Oxford Brookes University, UK.
Rosenberg et al. in a 2009 CGAP paper[2] estimate administrative costs - the largest single contributor to interest rates in an MFI - at about 11% of the loan portfolio. If cash were to be eliminated, administrative costs can drop by 2 x 1.3%. The annual interest rate can drop by 2.6% in consequence. Mobile banking can help reduce the cost of borrowing for the poor as long as the fee per mobile transaction is significantly below the cost of handling cash and infrastructure allows a convenient use of dematerialized money.
Khan and Ashta[3] found strong positive correlation between the cost per borrower and average loan balance for MFIs in Bangladesh. The cost of handling more cash can possibly explain a part of this diseconomy of scale.
What about the cost of the loan officers? They play a key role in micro lending as they select the borrowers and manage the portfolio. Many attribute the success of microfinance to the close personal connection between the loan officer and the client. Musoni also places the loan officer at the core of its financing model.
My question is: Can Musoni through technology improve significantly the efficiency of the approval process to further reduce the cost of borrowing for its clients? Grameen bank was amongst the first to challenge the status of the loan officer by ‘subcontracting’ to the self-help groups of borrowers the procedure of screening newcomers and enforcing loan collection, but this works for group lending only. Musoni’s borrowers use mobile phones to process the repayments, automated SMS are generated to improve communication with the clients, and loan officers use tablet devices to operate independently from the branch. In the future, algorithms and computers could probably replace the loan officers in financing standard applications for individual micro loans received through mobile phones.
Transaction cost can be reduced. On the applicant’s side, the loan demand can be made simply by sending an SMS containing the information about the requested amount and maturity. The applicant can do that without leaving his store or his workshop, or at home after sending the kids to bed. There is little cost for the MFI also. No need to keep the branches open for applicants or send the loan officers to prospect the neighborhoods.
The question is how to assess the creditworthiness of the applicants with little implication from the loan officers? IT and credit scoring give the answer, but where can the MFI find relevant data and make sure it is reliable?
For repeat borrowers the procedure is simple. The lender already knows how they behaved when repaying previous loans. Past credit behavior is by far the most predictive information of future repayment behavior. In the same spirit credit bureau reports can help screen applicants based on their past credit behavior with other institutions, but examples of credit bureaus that cover extensively micro borrowers are rare. Cameron Goldie-Scot added that Musoni Kenya played a key role in setting up a credit bureau for micro borrowers in partnership with AMFI - the local body representing Kenyan MFIs.
For new applicants the question is more delicate. Credit scoring algorithms usually use socio-demographic, business-demographic, financial data and loan characteristics to generate a score that estimates the probability of the loan to go delinquent or to default. Loan officers are the ones that collect this information by visiting the applicant at his workplace or home. Micro borrowers usually operate in the informal or semi-formal sector, so their capacity to generate or provide reliable documents that would make the visit redundant is limited. The cost of evaluation of the loan application is high in microfinance. Each client covers the cost of appraisal of his application and also contributes to offsetting the costs of assessment of refused applications.
The mobile phone company can provide some valuable statistics about the use of the mobile services by the potential borrower and his payment behavior with regard to these services. How many different contacts do you usually call in business and leisure time can indicate about the depth and width of your business and social network. Do same contacts call you back or others? Do you have friends or associates abroad?
How often do you charge the mobile phone account is also an indicator of your capacity to make money and eventually repay loans. Is your mobile phone subscription new or old, or it is pre-pay, could also be a valuable information. Through data mining, one could identify patterns of use of mobile phone services specific to creditworthy borrowers and ‘bad’ borrowers. These patterns are incorporated in credit scoring algorithms that can estimate the creditworthiness of the applicant with certain accuracy.
For more precision, the applicant may be required to provide more information about his family, business, financial position, etc. A smartphone application would be more convenient for answering questions, but the same procedure could be considered using SMS or mobile chat – services supported by all mobile phones and most providers of mobile phone services. Such information would be self-declared and thus prone to manipulation, but relevant statistical links could be made in order to get more accurate estimations of the credit risk. All this makes sense if the long term costs of the new IT infrastructure will be below the current costs.
Replacing the loan officer with a digital interface raises some business and ethical questions. The removal of the human interaction with applicants and borrowers could reduce the knowledge about the clientele. Many MFIs rely upon the loan officer to provide additional services such as training and assistance to micro-entrepreneurs in building successful businesses. From an ethical point of view, IT can make jobs redundant and this doesn’t go along with the microfinance mission. On the other hand, high interest rates keep away many micro borrowers or catch them into a survival trap where increased revenues are spent on interest and loan related fees. Minor personal or professional shock can lead them to overindebtedness. It could be that reduction of the interest rate by few percent points is translated in more job creation than losses on the MFI side.
It is still a taboo issue, but often loan officers are responsible for fraud. Some approve all loan applications to cash the variable bonus and leave before the portfolio deteriorates. Some look for potential borrowers willing to default and share the loan amount. Some charge the borrower a ‘personal’ fee of few percent for making the disbursal possible. The higher the risk of the borrower, the higher the fee is. This fee can increase further if the interest rate is subsidized or part of disbursed money is grant.
In theory, provided a strong implication of the stakeholders, the partial exclusion of the loan officer from the credit cycle is possible and often necessary. Experienced loan officers will be always required in the MFI, as their skills are crucial in assessing non-standard loan application or in launching new loan products and providing additional services such as coaching for clients. The question is where to put the cut-off between the loan officer and the machine and how to move it progressively towards the less expensive resource?
[1] Simulations based on a 12 months loan of 300 euro, bearing simple annual interest rates of 12% and 24%. 1.3% of the amount of the transaction (at disbursal and at repayment) is added as a fee.
[2] Rosenberg, R., Gonzalez, A. and Narain, S. (2009) The New Moneylenders: Are the Poor Being Exploited by High Microcredit Interest Rates, CGAP.
[3] Khan, S. and Ashta, A. (2012) Cost control in Microfinance: Lessons from the ASA case, Cost Management, 26(1).
Vitalie Bumacov
associate researcher with the Banque Populaire Chair in Microfinance at the Burgundy School of Business (Groupe ESC Dijon-Bourgogne), France and PhD candidate at Oxford Brookes University, UK.
