Research & Policy Briefs From the World Bank Chile Center and Malaysia Hub No. 23, June 2019 Financial Innovation and Additionality: The Power of Economic Analysis and Data Analytics Facundo Abraham Sergio L. Schmukler José Tessada As public and private financial institutions innovate to expand the range of financial products that households and firms use, questions about the additionality of different services have become central. Additional financial services are those that are not already provided by the private sector and that create value added to the overall economy. To identify and measure the contribution of financial services, precise analytical techniques and data are essential. Measuring additionality is challenging but, to the extent that it can be done, it is helpful to assess the multiple effects of financial innovation on the consumers of financial products, the financial service providers, and the economy as a whole. Introduction overall economy (Forbes 2019). Public sector banks, DFIs, and other state-owned entities want to understand which services Innovation is a staple of the financial sector. From ATMs, credit expand access to finance so they can design more effective policy cards, and securitization to branchless banking, mobile interventions. Private sector firms want to know whether new payments, and crowdfunding, new financial services are services are generating new businesses or are simply leading constantly transforming the way in which households and firms clients to substitute between services (for example, when firms access and use financial products. These new services help cannibalize their own products without boosting profits). complete markets and, in some cases, integrate marginalized Moreover, many private firms are interested in demonstrating groups and firms into the financial sector. how their innovations yield positive externalities and benefits to society. Academics also want to understand the effects of new Both the public and private sectors participate in this process financial services. For example, they might want to ascertain how of innovation. For the public sector, providing new financial different financial services offered by the public and private services is a way to try to increase access to finance. The financial sectors affect users as well as the economy as a whole, or sector fosters investment, technological innovation, risk whether new firms are complementing or substituting traditional diversification, and consumption smoothing, among other financial institutions. benefits. As a consequence, several governments around the world have established national strategies to deepen access to Evaluating the “additionality” of financial services can help finance. These strategies have led to a resurgence of answer such questions. Measuring additionality requires a new development financial institutions (DFIs) (The Economist 2019a). framework or approach that exploits data in the financial system and applies new statistical/econometric methods and economic In the private sector, both incumbents and new entrants also analyses, with the aim of providing a more accurate evaluation of produce financial innovations when searching for ways to access the impact of financial services. new segments, increase market share, and reduce costs. For example, rapid advances in information technology, cloud What Is Additionality? computing, and big data have opened the door to innovative financial services (The Economist 2019b). New financial A financial service is considered additional if it creates value for technology (fintech) firms (such as Akulaku, Cumplo, Go-Pay, the economy. In some cases, financial institutions offer new Jirnexu, Klarna, Lending Club, Qudian, and Viva Republica) are products that already exist in the market. In these cases, even if disrupting banking by offering easier and cheaper access to consumers use them, no value is created. Consumers are simply traditional financial services, as well as completing markets. substituting among similar products. But in other cases, financial Large, nonfinancial corporations have also ventured into the institutions innovate and provide new services that did not exist financial sector. E-commerce platforms, such as Amazon, before or offer existing services at lower costs and/or better Alibaba, Mercado Libre, and PayPal, now offer loans to small terms, increasing welfare for users. New financial services can suppliers, assessing firms’ risk based on their transaction history. also produce spillovers (positive or negative) to nonusers and Traditional financial institutions have responded to increased other financial intermediaries. Overall, financial innovations competition by offering innovative services, as well. For example, create additionality when their net effect on the overall economy traditional banks have started to become more digital by closing is positive, benefiting both consumers and producers of these branches and focusing on online services (The Economist 2019b). products (EP 2008). The continuous emergence of new financial services has The concept of additionality was initially conceived for the generated great interest in measuring how these services are public sector, particularly to guide interventions by DFIs and affecting the bottom line of financial intermediaries and the multilateral development banks (MDBs). The idea is that public Affiliation: Development Research Group, the World Bank; Pontificia Universidad Católica de Chile (PUC & Finance UC); and JPAL. E-mail addresses: fabraham@worldbank.org, sschmukler@worldbank.org, jtessada@gmail.com. Acknowledgement: We received very useful comments from Robert Cull, Aart Kraay, and Norman Loayza, as well as helpful edits from Nancy Morrison. The World Bank Chile Research and Development Center, the FCI group at the Malaysia Global Knowledge and Research Hub, and the KCP and SRP programs provided financial support for this brief. Objective and disclaimer: Research & Policy Briefs synthesize existing research and data to shed light on a useful and interesting question for policy debate. Research & Policy Briefs carry the names of the authors and should be cited accordingly. The findings, interpretations, and conclusions are entirely those of the authors. They do not necessarily represent the views of the World Bank Group, its Executive Directors, or the governments they represent. Global Knowledge & Research Hub in Malaysia Financial Innovation and Additionality: The Power of Economic Analysis and Data Analytics interventions in the financial sector should add resources beyond growth, more jobs, increased innovation, new markets, and those already in the market and should not crowd out the private higher education. sector (MDBs 2018). In this context, additionality refers to the provision of financial services that the private sector is not How Can Additionality Be Measured? offering (such as when public banks provide microfinance products to low-income households that are not served by Measuring additionality is not straightforward. Because it entails traditional banks) or to the mobilization of the private sector to measuring the extent to which a financial service produces a net provide services that otherwise would not be provided (such as positive outcome to the economy, it requires identifying and by reducing risks and/or transaction costs for the private sector) quantifying the financial and economic effects of a financial (OECD 2016). For example, when a government establishes a service on its users, nonusers, and financial intermediaries. public credit guarantee scheme, the public sector might not Whereas measuring additionality comprehensively can be provide any financing directly, but instead could promote private intricate and time-consuming, several approaches help identify lending by reducing risk and the need of collateral for guaranteed the possible implications of financial services and provide borrowers. insights as to their additionality. Examples of these approaches, which are reviewed next, illustrate some principles of what Although additionality has been commonly used in the constitutes a proper approach to start measuring additionality. context of public interventions, it is possible to extend this concept to the private sector. In the search for profit To be clear, additionality is typically not properly measured by maximization, the private sector can produce new value for the indicators that only quantify the scale of a financial service (such economy (The Economist 2019b). For example, when traditional as the number of clients served, the total disbursements made, banks and fintech firms (such as Destacame, Lendingkart, and or the revenues from those services). Many times, such Tala) use nontraditional data to assess clients’ risk and provide or quantitative indicators are used because they are simple to facilitate access to lending, they can generate additionality by calculate and communicate. However, they do not provide a enabling credit-constrained individuals to access financial good metric of the actual contribution of a service to the resources. Similarly, online platforms that bring together small economy. They offer no information about whether the service is firms, their large suppliers, and private banks to promote increasing the overall use of finance or is displacing other factoring operations (such as ASYX, InvoiNet, Orbian, Prime services and/or institutions. They also do not provide information Revenue, and Taulia) can produce additionality by allowing small about whether the financial services are actually necessary for firms to obtain working capital. Many times, the private sector consumers, how they are being used, or what their economic can produce additionality by processing existing data on their impact is. Given these limitations, alternative analytical own operations to better understand their clients and identify approaches can be used to address these issues. gaps in the market, which are then served with new products. 1. Case Studies The public and private sectors can create additionality in One approach is to conduct a case study. This method is mainly similar ways. In essence, additionality takes place when qualitative and tries to answer basic questions about a financial institutions address existing market failures that constrain access service. For example, it could start by identifying what causes the to finance. Lack of information about borrowers, weak creditor problem of access and who is affected. Then, it could review how protection, coordination failures, or high costs of serving specific the service is designed and how it addresses market failures. geographical areas, among other frictions, can limit the provision Moreover, a case study could examine whether the targeted of financial services. Finding innovative ways to mitigate these group is effectively using the service and whether the service is frictions can lead to new resources being generated or new value sustainable over time. Comparison with similar services in other being created—whether this is done by the public or private countries could also be informative. For services offered by the sector (de la Torre, Gozzi, and Schmukler 2017). In fact, in many public sector, a case study could also examine whether the cases a new additional service provided by the public sector private sector participates, how incentives are aligned, and what could have been equally provided by the private sector and vice different contributions the public sector offers (de la Torre, Gozzi, versa. A good example is credit information schemes, which can and Schmukler 2017; CGD 2019). Case studies can be an be established by the public sector (credit registries), the private illustrative way to understand how a financial service is designed, sector (credit bureaus), or both simultaneously. how it addresses gaps in the financial sector, and how it could potentially create value. However, they lack quantitative analysis Additionality can occur in multiple ways (EP 2008). For of the actual impact of the service in practice. example, a financial service can improve overall finance by targeting excluded groups, reducing costs, or improving terms. 2. Financial System Data Analyses Another important aspect is whether the service is used for its intended purpose. For instance, various studies of microfinance A more data-driven approach is to examine data for the entire have found that even though microentrepreneurs gain access to financial sector, including public information from central banks, credit, they often use the new financing for consumption instead supervisory authorities, and credit bureaus/registries, as well as of investment (Attanasio et al. 2015). Similarly, payday loans data from private vendors. These data can offer an overview of might facilitate lending to low-income households, but if this the financial sector and how different financial institutions are comes at the cost of customers becoming highly indebted and providing valuable services. For example, analysis of loan and bankrupt, such loans cannot be considered to be generating credit bureau/registry data could help determine whether the additionality (Liberty Street Economics 2015). Another way of emergence of new fintech firms allow new individuals and firms looking at additionality is to observe whether the service to obtain loans (Jagtiani and Lemieux 2017). Moreover, data on generates an economic impact, for example, in terms of higher how individuals and firms enter and exit financial institutions 2 Research & Policy Brief No.23 Figure 1. Difference-in-Differences Figure 2. Discontinuity Analysis Examine just before and just after cut-off With Banco Azteca Number of informal businesses Municipalities without Banco Azteca Effect Academic performance of Banco Azteca Effect of program Municipalities with Banco Azteca No Banco Azteca 2002 Before Banco Azteca After Banco Azteca Cut-off score Exam score This figure shows the effect of loans to low-income households granted by Banco This figure shows the effect of a college loan program in Colombia on academic Azteca on the number of informal businesses in Mexico. Starting operations in 2002, performance. The program granted loans only to those students who scored above a Banco Azteca was initially present in some municipalities but not others, allowing for minimum score in their high school exit exams (Melguizo, Sanchez Torres, and a before-and-after comparison (Bruhn and Love 2009). Velasco 2016). could be used to show whether specific intermediaries, such as 2009); comparing the removal of subsidized credit for exporters DFIs or microcredit institutions, are “entry points” to the financial of yarn but not for other exporters in Pakistan (Zia 2008); and system. In particular, the data could reveal whether customers examining the effect of access to payday loans by exploiting the used other financial institutions before becoming clients or assignment of military personnel in the United States to different whether clients move to other institutions over time after being states with and without payday loans (Carrell and Zinman 2014). served by a DFI or a microcredit institution. For the public sector, an important line of investigation is whether public participation b. Discontinuity analysis. In other cases, the eligibility to in private operations can lead to better outcomes. For example, access a specific financial service depends on a clearly defined some studies examine whether participation of MDBs on cutoff, such as income level, firm size, credit score, or a test syndicated lending can increase the volume and maturity of score. For example, only households with income below a loans and whether infrastructure projects supported by MDB threshold might receive financial aid or only firms below a certain loans have lower cancellation rates (Broccolini et al. 2015; size could apply for subsidized lending. In these cases, individuals Marcelo Gordillo and House 2016). “just below” the cutoff would be very similar to those “just above” (in essence, one group barely does not make it and the 3. Impact Evaluation other group barely does). In such cases, the only fundamental A more focused approach based on data analytics consists of difference among the two groups is that one uses the financial quantifying the effect that a financial service has on its users. The service by qualifying for it, while the other does not. Thus, it is idea is to compare users of a financial service (the “treated group”) possible to compare outcomes between them (Figure 2). For with a similar group of nonusers (the “control group”), and identify example, a student credit program in Colombia that grants credit whether users benefit in a way that is not observed for nonusers. based on a test score allows researchers to examine the effect of Whereas this approach can be very useful when trying to student loans on academic outcomes (Melguizo, Sanchez Torres, understand the direct effects of a financial service on users, it does and Velasco 2016). Other studies have examined mortgage not necessarily provide insights on the spillovers to nonusers and lending support in the United States to households below a other financial intermediaries. For example, an observed increase determined income level (Bhutta 2011) and the provision of in lending to the treated group could come at the expense of credit guarantees in Italy to firms based on balance-sheet reduced borrowing by third parties that are not being considered observables such as leverage or cash flows (de Blasio et al. 2014). in the analysis. Thus, one has to be cautious when interpreting results. This approach, which is commonly known as impact c. Randomized control trials. Sometimes, financial services evaluation, can be applied using different methods (Gertler et al can be tested through a small-scale intervention before being 2016). Some of these methods are reviewed next. broadly implemented. In particular, the financial service can be randomly assigned among households, firms, or geographical a. Difference-in-differences. Some financial services are areas. This random assignment means that users and nonusers restricted to specific geographical areas or sectors. In these will not have different characteristics on average, and thus it is cases, it is possible to compare trends before and after the possible to compare the difference in outcomes among both service is introduced for both users and nonusers. If after the groups (Figure 3). In many cases, these experiments are introduction of the new financial service, users experience a conducted by private and public institutions jointly with positive change in trend while the trend for nonusers remain researchers. For example, in Malawi a government-owned unchanged, then it is possible to conclude that the service has microfinance institution randomly required farmers applying for beneficial effects for users (Figure 1). Examples include studying agricultural loans to provide fingerprints. The goal was to the impact of introducing loans to low-income households in examine whether borrower identification reduced moral hazard some Mexican municipalities but not in others (Bruhn and Love and adverse selection, increasing repayment rates (Giné, 3 Financial Innovation and Additionality: The Power of Economic Analysis and Data Analytics way they think about financial innovations: to abandon the Figure 3. Randomized Control Trial narrow view that financial services should increase finance per se Treatment group in favor of considering how financial services affect market Repayment participants and how successful they are at accomplishing a Sample rates particular expansion or development goal. Measuring additionality also requires increased data analyses 2. Finger printing 1. Randomly split and coordination. Assessing the additionality of financial services sample 3. Comparison can involve processing and analyzing a substantial body of data. In some cases, financial institutions would need to analyze data Control group from their own operations. Many times, these data are already available, but financial institutions do not have the technical Repayment rates capabilities and expertise to process and exploit them. The academic community could play a useful role in the design and This figure shows the design of a borrower identification experiment in Malawi. evaluation of financial services and in the validation of results. In Farmers applying for agricultural loans were randomly fingerprinted. Their loan repayment rates were compared with borrowers who were not fingerprinted (Giné, other cases, measuring additionality would require looking Goldberg, and Yan 2012). beyond a financial institution’s clients and examining trends in the financial sector as a whole. But doing this would require Goldberg, and Yan 2012). In Senegal, a microfinance institution increased coordination among financial sector participants to arbitrarily encouraged new clients to open an account at either a share and match relevant information. banking agent or a branch. Then, researchers examined how account usage changed depending on how clients opened the The discussion and methodological approaches introduced in account (Buri et al. 2018). In Ghana, a private bank randomly this brief can help the public and private sectors start measuring offered a new savings product to clients to analyze how this additionality. But these approaches provide only a partial view of product affected financial behavior (Buehren et al. 2018). the additionality of new financial services. As financial sector stakeholders become more familiar with the concept of What Are the Challenges in Measuring the Additionality additionality, it would be convenient to develop new, more of Innovations? comprehensive gauges of additionality. Achieving a good balance Measuring additionality facilitates better evaluations of the between a measure that is general enough to capture multiple impact of financial innovations and thus helps improve the effects but that, at the same time, is not too overly complicated overall allocation of both public and private resources and to calculate would be challenging. Academics have already services in the financial sector. However, measuring additionality started thinking about the next steps in evaluating additionality is not easy. It requires analysts and practitioners to change the (Taylor and Filipski 2014). References Attanasio, O., B. Augsburg, R. De Haas, E. Fitzsimons, and H. Harmgart. 2015. “The Forbes. 2019. “The Fintech Revolution Is Here. Can It Help Build A Better Impacts of Microfinance: Evidence from Joint-Liability Lending in Mongolia.” Economy?” February 22. American Economic Journal: Applied Economics 7 (1): 90–122. Gertler, P., S. Martinez, P. Premand, L. B. Rawlings, and C. M .J. Vermeersch. 2016. Bhutta, N. 2011. “The Community Reinvestment Act and Mortgage Lending to Impact Evaluation in Practice, 2nd edition. Inter-American Development Lower Income Borrowers and Neighborhoods.” Journal of Law and Bank and World Bank. Economics 54 (4): 953–83. Giné, X., J. Goldberg, and D. Yan. 2012. “Credit Market Consequences of Improved Broccolini, C., G. Lotti, A. Maffioli, A. F. Presbitero, and R. Stucchi. 2015. Personal Identification: Field Experimental Evidence from Malawi.” American “Mobilization Effects of Multilateral Development Banks.” IMF Working Economic Review 102 (6): 2923–54. Paper 19/28, International Monetary Fund. Jagtiani, J., and C. Lemieux. 2017. “Fintech Lending: Financial Inclusion, Risk Bruhn, M., and I. Love. 2009. “The Economic Impact of Banking the Unbanked: Pricing, and Alternative Information.” Working Paper, Federal Reserve Bank Evidence from Mexico.” Policy Research Working Paper 4981, World Bank. of Philadelphia. Buehren, N., M. Goldstein, L. Klapper, T. Koroknay-Palicz, and S. G. Schaner. 2018. “The Limits of Commitment: Who Benefits from Illiquid Savings Products?” Liberty Street Economics. 2015. “Reframing the Debate about Payday Lending.” Policy Research Working Paper 8633, World Bank. October 19. Buri, S., R. J. Cull, X. Giné, S. Harten, and S. Heitmann. 2018. “Banking with Agents: Marcelo Gordillo, D., and R. S. House. 2016. “Effects of Multilateral Support on Experimental Evidence from Senegal.” Policy Research Working Paper 8417, Infrastructure PPP Contract Cancellation.” Policy Research Working Paper World Bank. 7751, World Bank. Carrell, S., and J. Zinman. 2014. “In Harm’s Way? Payday Loan Access and Military MDBs (Multilateral Development Banks Working Group). 2018. Multilateral Personnel Performance.” Review of Financial Studies 27 (9): 2805–40. Development Banks’ Harmonized Framework for Additionality in Private Sector Operations. CGD (Center for Global Development). 2019. “Case Studies on Access to Financial Services.” Center for Global Development. Melguizo, T., F. Sanchez Torres, and T. Velasco. 2016. “Credit for Low-Income Students and Access to and Academic Performance in Higher Education in de Blasio, G., S. De Mitri, A. D’Ignazio, P. Finaldi Russo, and L. Stoppani. 2014. Colombia: A Regression Discontinuity Approach.” World Development 80 (C): “Public Guarantees to SME Borrowing. An RDD Evaluation.” Università degli 61–77. Studi di Padova. OECD (Organization for Economic Co-operation and Development). 2016. de la Torre, A., J. C. Gozzi, and S. L. Schmukler. 2017. Innovative Experiences in “Understanding Key Terms and Modalities for Private Sector Engagement in Access to Finance: Market-Friendly Roles for the Visible Hand? Latin American Development Co-operation.” OECD. Development Forum, World Bank. Economist, The. 2019a. “National Development Banks Are Back in Vogue.” March 7. Taylor, J. E., and M. J. Filipski. 2014. Beyond Experiments in Development Economics. Oxford University Press. ---------. 2019b. “Special Report: A Bank in Your Pocket.” May 2. Zia, B. H. 2008. “Export Incentives, Financial Constraints, and the (Mis)allocation EP (English Partnerships). 2008. Additionality Guide: A Standard Approach to Assessing of Credit: Micro-level Evidence from Subsidized Export Loans.” Journal of http://documents.worldbank.org/curated/en/795161508437504068/Corporate-debt-maturity-in-developing-countries-sources-of-long-and-short-termism the Additional Impact of Interventions, 3rd edition. English Partnerships. Financial Economics 87 (2): 498–527. Global Knowledge & Research Hub in Malaysia