This is the second in this year’s job market series.
Increasing access to bank loans for small and medium enterprises (SMEs) in low- and middle-income countries is considered vital for economic development. Interventions to boost SME lending exist in most countries, and they are commonly implemented through commercial banks as partial default guarantees or loans with promotional interest rates. These two interventions are widespread even in countries where the fewest SMEs have bank loans, yet the percentage of registered firms that apply in low- and middle-income countries (LMICs) remains lower than anticipated.
Many would declare that promotional interest rates need to be pushed even lower. However, interpreting the low number of applications as low demand for loans could be a mistake for three reasons. First, in countries where credit information is scarce, SMEs cannot submit their own applications—loan officers need to spend time gathering information on the prospective borrowers under the guidance of the bank’s underwriting team. Second, loan officers report hesitations to screen many prospective SME borrowers after initial discussions, because SMEs are unwilling to discuss how their financial statements were generated and struggle to acknowledge their business’s industry and idiosyncratic risks. Third, offering promotional loans is the loan officer’s decision. Banks can always lend at whatever interest rates they deem appropriate, meaning that internally, promotional lending schemes are a way for banks to incentivize loan officers to offer lower interest rates to certain borrowers. Lending interventions might offer incentives for commercial bank head offices to want to lend, but these incentives might not translate into benefits for loan officers. Taken together, low applications might be driven by firms never having arrived at the bank, by firms being unable to secure acceptable loan terms, or by the loan officer denying them. Firms that the bank would want to consider for promotional loans might not be asking for them, or loan officers might not be properly incentivized to offer. Critically, these nuances would not be captured in bank administrative data, as the first time firms appear is when applications are submitted.
In my job market paper, we study what drives the low number of application submissions for small business bank loans in Pakistan, where all small business bank loans to new borrowers are partially guaranteed against default by the central bank, and every commercial bank has lending schemes advertising lower-than-market interest rates for small business borrowers.
We do this by recruiting 702 firms across Pakistan, referring them to loan officers, and documenting where they drop out of the application process. Along the way, we randomize two interventions to firms: (i) information on the existence of lending schemes at commercial banks (“Every registered SME was eligible to be considered for a 9%, physical-collateral free loan under the State Bank of Pakistan’s Asaan Finance Scheme”), and (2) preparation for the first meeting with loan officers (via a mock interview, specific feedback, and quick tips).
Registered SMEs in our experiment were recruited via the Small Medium Enterprise Development Authority of Pakistan, local trade bodies, and local commerce chambers. Respondents were told the study was about how SMEs in Pakistan finance themselves, and relative to the average SME, they were more likely to have audited financial statements and accept digital payments from customers. Over 90% of the firms were in manufacturing or services (IT, software, accounting, education) and less than 10% were in retail.
How far do firms make it? Do they drop out before or after they arrive at the bank?
We find very little aversion to the bank, as measured by 80% firms requesting a referral to the bank and by 51% having a real interaction with a loan officer. Instead, we find the main point of dropoff to occur after interacting, but before advancing to the screening stage of the application process, where loan officers visit the business and regenerate parts of the accounting statements from transaction level data to verify validity.
We find that of the SME-loan officer pairs that interacted but did not result in applications, 63% explicitly said that interest rate or collateral stopped the process, and 27% said more generally that the SME was no longer interested. Taken together, this means that 90% of dropoff was driven by SMEs giving up on the bank loan, and not by loan officer denials.
