Inside the Production Function: The Effect of Financial Contracts on Growing Firms' Technology Use. Evidence from a field experiment in Uganda.

We examine how key aspects of the most common form of financing-debt-may inhibit young firms' expansion. Starting a business entails learning and risk taking, implying that project returns to investment can start low but increase over time (in other words, be "backloaded") or be uncertain. Also, indivisible start-up costs often require large investments. Meanwhile, standard debt contracts available for micro-entrepreneurs from the formal or semi-formal financial sectors of many developing countries (such as microfinance) stipulate a constant repayment stream and caps on the initial loan size. The interaction of such features of the loan contract and the firm's production technology, may distort investment toward inputs that involve less learning, less uncertainty, and smaller projects; hampering firm growth. To shed light on the extent to which these theoretical mechanisms limit the effectiveness of microloans, we plan to collaborate with BRAC Uganda's Small Enterprise Lending Program to study the effect of the credit terms on starting firms' input use, profits, and repayment performance. As such, our project contributes to the DFID-ESRC Growth Research Programme's focus on Finance and Growth in Low Income Countries.

Small and medium-sized firms are the engines of the Ugandan economy, comprising over 90 % of the private sector and BRAC Uganda has been lending to such firms since 2008 through its Small Enterprise Lending Programme. The loans range from 2.5 million to 13 million Ugandan Shillings (630 to 3,300 GBP) and are repaid monthly with a maturity of 12 months at an annual interest rate of 25%. The research project will select, among firms applying for BRAC loans from mid-2014, a representative sample of 1600 firms to be part of a randomized experiment.

In order to investigate whether standard contractual terms in microloans from formal or semi-formal sources are restrictive for firms that face indivisible costs and/or are characterized by backloaded or uncertain project returns, we will (randomly) implement the following interventions for different groups of firms by: (i) changing the repayment frequency to distinguish the effects of uncertain project returns from those of backloaded returns; (ii) offering subsidies to ease the purchase of indivisible goods; and by (iii) offering consultancy services to shorten the learning process about the use of certain inputs to alleviate problems of backloadedness. We will survey these firms at baseline and upon completion of the loan-cycle (1-year after) to measure the change in firms' production and profits. In addition, we will use detailed and high frequency firm data to trace how the financial structure, a firm's learning curve, or the ease with which an indivisible good is acquired affects the use of machines and labor and how this in turn impacts profits and repayment performance. The project will provide unique evidence on the constraints caused by the interaction of financial structure and technology use that complements the recent emphasis on access to finance.

The results of our research will be relevant for four key stakeholder communities.

First, members of BRAC's Small Enterprise Lending and Microfinance Programmes are likely to benefit and use our findings to improve the design of credit products to better meet the needs of business owners. As described in our Case for Support, the research aims to shed light on what contractual features enable greater benefit, in terms of firm growth and profitability, for borrowers of the programme. These findings will be communicated to BRAC program officers, which will be facilitated by our existing networks within BRAC. BRAC currently operates in Bangladesh, Pakistan, Sri Lanka, Uganda, Tanzania, South Sudan, Liberia, Sierra Leone and Haiti. Its Small Enterprise lending Program is expanding rapidly throughout its regions of operations. As a result, there is great potential for the research findings to shape BRAC's lending operations both in Uganda and in other countries where BRAC operates. In particular, to continue building our relationship with BRAC and to make sure that the findings are communicated to them at every stage of the project, we plan to undertake the following steps: (i) travel to Uganda to meet and work with programme management at the start of the project; (ii) conduct regular visits to Uganda and to Bangladesh to meet with programme management; (iii) conduct presentations and prepare reports on our findings to BRAC management; and (iv) provide training for program officers.

The second community of stakeholders comprises of civil society and policymakers, in particular those are working on expanding credit access to the poor. The output of our research will provide insights on how to design microfinance products to achieve maximal effectiveness in terms of firm growth and profitability. As such, microfinance institutions and policymakers are among the key stakeholders that may benefit from this research. To ensure that our results are disseminated within this community we will organize stakeholder workshops, participate in relevant events, and use our existing network connections with the civil society and policymakers working on microfinance.

The third community to benefit from our work will be local researchers and members of the research team in Uganda that will directly or indirectly benefit from our capacity building efforts. These efforts will contain different modes of training and experience opportunities for local researchers. These include training that will be provided to the local research manager, the local research team, enumerators, and a training session on developments in microfinance research that will be offered to local researchers and students.

Last but not least, the fourth stakeholder group to benefit from our findings is existing or potential entrepreneurs in search of better credit opportunities to establish or expand their businesses. As the end users of BRAC's and other MFIs' loan products, these individuals will benefit directly from improvements in the design of loan products resulting from the findings of our research. We will pay particular attention to improving credit access to female entrepreneurs. For this, in addition to investigating how household-level constraints affect the success of female versus male entrepreneurs and disseminating our findings to BRAC and other MFIs (as described above), we will collaborate with BRAC to conduct a publicity campaign for Small Enterprise Loan among SMEs, targeting female entrepreneurs.