Micro finance, pandemic and role of open source technology

In India, rural communities are historically the people for whom services offered by high-street banks represent an elusive source of liquidity, product innovations and financial democracy. This article explores the role of open source technology in the Micro finance sector.

This post is written by S.J. Stapleton, University of Copenhagen, November 2020.


In India, rural communities are historically the people for whom services offered by high-street banks represent an elusive source of liquidity and financial democracy. Thanks to advantageous borrowing terms of micro finance, the means for achieving long-term daily sustenance becomes a veritable possibility. Very low-income customers can now enter into loan agreements as one of many in large groups of local people who, together, combine minute assets into a potentially meaningful cash-flow.

However, COVID-19 is threatening to dissolve the good practices and standards that have been introduced to India’s poorest people over the last decades. The pandemic has raised doubts regarding the continued sustainability of microfinance terms. It appears that the industry’s customers face new looming choices; between nourishment and repayment deadlines.

This article considers the evolution of an industry, yet in its infancy, and asks what open-source technologies can do to reach those without access to finance, while reducing the fixed costs. It suggests that an initiative supported by Apache Free Software Foundation, such as the Apache Fineract, currently used by more than 250 financial institutions (including credit unions, asset-based lenders, and SME lenders) that share 3.5m+ borrowers in 37 countries.

Another open source project named Finscale is responsive to poverty alleviation, strategic goals and economies vulnerable to natural disasters.

What is microfinance?

Microfinance is typically characterized as financial services for clients excluded from mainstream banking provisions. It does aim at debt solutions and poverty alleviation for the ‘neglected’ populations who survive on very low household incomes, whose financial shortfalls are occasionally met by moneylenders. Few challenges for microfinance organizations expressed by their borrowers themselves include social attitudes towards debt repayments, geographic restrictions on the collection of loan repayments and dependencies on collections’ agents for relationships; and, the extended family’s financial requirements.

Whereas Paruthi, Frias-Martinez, and Frias-Martines (2016) argue for further research that analyses micro-lending platforms to comprehend community borrowing structures, extensible to MFI profitability (Ibid.: 2180-2181) and attributable to trust and borrower-power (Ibid.: 2181), the rating model also warrants further consideration if its conceptual development is to become subject to refinement (Addi and Souissi 2020: 380, 384). For Terko, Zunic, and Donko (2019), credit risk modeling should be undertaken only where the nature and specificity of the data are robustly comprehended and therefore, they propose that other techniques could be more appropriate (Terko, Zunic and Donko 2019: 1).

It has been a challenge to calculate appropriate loan pricing for community groups without legitimate documentation in the form of national identification numbers and therefore credit histories. Microfinance firms have developed specialized predictive modeling solutions. The Propensity Score Matching utility is a probabilistic determiner according to observed characteristics, attitudes, and psychological indicators. A similar metric, ‘Kiva’, provides sports groups, religious congregations, schools, and other common interest groups with finance according to the terms of a framework that uses machine learning to validate social cohesion and repayment pdfs.

The MFI sector in India comprises one part of a diversification strategy in high capital entities’ portfolio management. The Indian banking industry shared 5,475,530 borrowers in 2017 and operates using information management and communication strategy. These strategies harness customer relationship management and high national competencies. “Since 2006, MFIs have undergone the first stage of transformational change from the spreadsheets towards open-source fintech solutions. They have begun to move towards personalized and gamified, innovative financial platform products that combine synchronicity with aesthetic interfaces.” suggests Muellners RnD.

Field officers collect and disseminate loans which serve as a useful stimulus for the industry. This combines a business mgmt. strategy with relationship mgmt. The sector requires a personalized communication, frequent interaction, gradual involvement in the orchestration of business fundamentals that represent both the importance of the loan and the measures necessary to ensure sustainable repayments. Maintaining and controlling system infrastructure has been a balance between anticipation of future regulatory practices influenced by GDPR(2018), as well as multi-faceted due diligence. Evidenced by poor functional control, the unreproducible results and distinct omissions in system requirements, such as comprehensive BI, methods for early verification and validation of systems, complex systems’ suitability for service-orientated processes, could nonetheless operate interdependently and consistently.
Real-time batch processing, transparency, and agile workflow patterns are considered by Muellners to provide the required ‘modularity, decomposition, deployment’ of open-source technologies. The importance of digital identification and traceable records is a persistent challenge for real-time data analysis models, which have not supported insurance provisions and have excluded many from Covid-19 aid packages.

It is important to note that Muellners’ experiences in the industry have indicated that responses to poverty alleviation are more effective wherever digital solutions are implemented. Although the challenge thus becomes the low rates of smart mobile ownership and unfamiliarity with app-based services that organize everything from travel itineraries to buying a home. These knowledge competencies and a bridge across the digital wasteland are required to build trust in technological services, for example, receipts from credits and debits, balance statements, transaction histories, all necessitate the increasing appeal of digital media. The additional incentive for the industry is that low transaction charges must be maintained during the Covid-19 response and to accomplish this sustainability initiative, the technological functions of online services must offer the equivalent sentimental experience and management as the in-person field officers provide. As the advertising conduits that appeal to women in India are founded on social forums, embraced via app installation, loan incentives that endorse and enable mobile adoption in suitable geographic areas should be considered. Others appreciate an app competency that permits geotagging of livestock, equipment, or soil conditions supported by alerts from text messages, and ‘tracking and tracing’ could provide the industry and the client with required efficiency and scalability.

Errors of the past

Critics of open-source technology solutions in the financial sector argue that it has failed to support and maintain microfinance capital. After all, the IBM sponsored e-tablets scheme that was intended to revolutionize microfinance and introduce rural communities to smart e-banking systems has generated low adoption rates. The industry is also supported by relatively few well-intentioned executives, developers, and volunteers who are ill-equipped to meet the constantly evolving landscape. These two factors may have led to the demise of projects initiated by both SKS India and Compartamos Mexico (Parekh 2006: 225). And then there are additional challenges faced by several microfinance institutions, such as mobile computing implementation costs that reach figures between $23,000 – $88,000 (Parekh 2006: 226) because good practices and information dissemination do not always occur. However, technical solutions were also implemented successfully; (i) SafeSave Bangladesh achieved employee efficiency, rapid loan applications, financial due diligence, and data accuracy that augmented trust; Basix India observed lower transaction costs, improved accountability, consumer trust from synchronicity with mobile computing and centralized MIS (Parekh 2006: 225-226).

Cohesion between MFIs

For Parekh (2006), the future of microfinance institutions is dependent on specialization, standardization, and systemization (Parekh 2006: 231), which Ma (2018) argues can be implemented by measures of dependency, code reuse, and knowledge flow networks in open source software (Ma 2018: 1) that use APIs to implement financial products extensible to non-financial activities and that access LAT functions at banks using a tech platform (IMF 2019: 42). Certain steps thus become vital to implementation:

Since Parekh examined three determiners of social change in microfinance; information exchange in remote settings, institutional management and processing, and collection and transfer of funds to rural clients (Parekh 2006: ), Lui et al.’s (2019) research has further advanced the field to include machine learning to predict Kiva’s lending activity and thus, concluded that smaller loans and applications from women are most rapid (Paruthi, Frias-Martinez and Frias-Martnez 2016: 2181), a phenomenon Gelb and Mukherjee attribute to a national ID number (Gelb and Mukherjee 2020: 10).

Mendez et al. (2018: 1013) researched gender diversity in OSS for microfinance organizations with the surprising conclusion that, ‘tools and infrastructure are complicit in newcomer and gender-biased barriers’, resulting in persistent barriers to entry and Nafus et al.’s ‘epistemological pluralism’ (multiple solutions to problems). The recipients of loans in India can bypass restrictions on the number of loans payable due to legislation that serves as an impediment, in the vein of GDPR(2018) and the German Bundestag v Facebook (2015), to share databases, in addition to eligibility based on financial status (Gelb and Mukherjee 2020: 15), and worsened by data portability uptake and data and privacy protection (Ibid.).

Critics have voiced doubt concerning the ability of the MFI to fulfill loan commitments and the viability of loans unavailable from banks with familiar names and reputations, which have generated trust and respect throughout generations. These institutions use strategic borrowing tactics that incur high maturities across loan durations for borrowers in addition to supporting their business models with guarantees that secure investments. However, the World of Code (2019) data pipeline, built on open-source technology, offers high-performance capabilities in the form of data mining and aggregation without substantial loss of time to platform latency. It has brought a competitive dimension to service delivery on equal footing with the established credit repositories such as PROMISE, Moose, Black Duck OpenHub, Sorcerer DB (SQL), which can perform predictive modeling, and Boa, which performs search queries across multiple repositories. Driven by shareholders’ investments, many familiar banking houses are exploring the potential for open-source technology in core business activities.

The migration of records from paper to computerized systems is an impediment to the acceptance of online banking services by communities for whom their relationship with the collections’ agent serves as the single point of contact with the MFI. Whilst field officers maintain the trust of communities, cultivating aspirations to financial security and independence, they can capitalize on casual inquiries to build stronger relationships and integration of the culture of an MFI enterprise. This close contact encouraged by strategic goals is hindered by the global pandemic and integrated solutions are essential to curb the spread of the virus and maintain business performance.

The field ‘loan’ officers were utilized to take the MFI to the people from the early days of their inception (Parekh 2006: 224) and nowadays those same officers perform millions of small transactions each month in tens of thousands of villages. They have recorded millions of personal data pointers, credit assessment reports, and scoring systems through regular conversations, community presence, and relationship building processes. The Internet platforms used should demonstrate an easily comprehensible interface and insight into the challenges faced by borrowers themselves. Moreover, the unique structure of collaborative loan partnerships should become readily identifiable to borrowers to contribute to the sustainable objectives of both parties, achieved through specialist product offers and evaluation metrics, standardization in decision-making. The rationale behind systemization should supplement the philanthropic idealism and philosophy of accessibility.

Algorithmic implementation of credit scoring decisions aggregates data for accurate modeling purposes, harnessing the power of the significant data resource now available to MFIs and sponsoring organizations. It has demonstrably improved person-centered financial capacity, building attitudes with an accuracy result of 81.8%. This is based on multiple observations relevant to specific features (Terko, Zunic and Donko 2019: 5). However, the subjectivity of field officers remains pertinent, the data elicited and recorded by personal interactions, frequently from designated community spokespersons. Furthermore, this technique offers valuable clarity to comparative data matrices that could also prove influential in database schemas.

The integrity and unanimity of loan decisions can be supported by multiple elements of an open-source technology environment that are advantageous for a new generation of micro-finance customers. A determiner of loan services is inherently biased towards the microfinance borrower and lending organization. In addition to the conventional measures of account usage patterns, employment history, and propensity for debt, MFIs also use social media data repositories, payment default predictions using sentiment analysis, and macroeconomic indicators of liquidity. MFIs are also known to use household occupancy data as the basis for their loan decisions. Overall, this presents business intelligence bottlenecks and shortfalls in financial management. To circumvent this challenge, MFI business models practice stakeholder consultations to synchronize distributed ledgers and balance sheet reconciliations (IMF 2019: 24). The distributed ledger’s capacity to support updates affects its implementation where an inordinate number of systems are combined and its stability, feasibility, and therefore, the directionality of open-source change is shifting towards flattened real-time structures.

Hence, alternative sources of finance, most specifically crowdfunding and alternative lending, are considered to offer significant productivity potential (Quinn, Guo and Castro 2016: 5). Globally financial institutions’ interest in web-supported databases of microfinance borrowers’ assets has soared as they are looked on as potential customers; for instance, the International Monetary Fund predicts ‘significant gains’ from financial technology solutions in the areas of payments, clearing, and settlement (IMF 2019: 1). This development has a duality of purpose that leads to either fruitful and prosperous financial relationships for banks with MFIs or direct competition between them. The regulatory structure is overtly conventional and ill-equipped for fintech’s challenges (IMF 2019: 2) in microfinance organizations, especially information asymmetry, unverifiable identities, and low range of products (IMF 2019: 31), to the advantage of the MFI.

The fact remains that microfinance is an evolutionary poverty alleviation strategy that has increasingly proved its worth as a fintech innovation on API platform infrastructures, as exemplified by Apache Fineract (AF). The vision is ‘to provide widespread access to community-driven open source banking technology’ with three goals; (i) sustainability, (ii) inclusion, (iii) advocacy. Thus, the characteristics of AF are consistent with declarative knowledge artifacts demonstrating network dependency, code reuse, and knowledge base dependent operations (Singh, Bhar and Sen 2018: 458) that are responsive to Parekh’s (2006) call for a definitive value chain standard (Parekh 2006: 233). Its’ three-layered architecture is built on a foundation of reporting, accounting, organization, and CRM/KYC. There is an audit layer that supports a portfolio applications process and a loan default estimation layer that uses logistic regression. The competitive advantage of Apache Fineract seems to rest with a comprehensive UI, alleviating maintenance, and control issues on a MySQL/Java Client API arranged around labels and icons operating on modules. This has enabled managerial practices common to other financial institutions of this size, like localized data and geo-tagging specific to the regions’ roles and permissions assigned to be ascribed.

According to Alaoui and Tkiouat (2018), customer satisfaction and the uptake of digital solutions is determined by the customer’s willingness to learn how to use it and trusting it sufficiently to allow it to replace previous modes of representation, in this case, the dominance of field officers. Furthermore, India is primarily distributing loan solutions in cash, a somewhat worrisome area that should be tackled by technological innovation (Gelb and Mukherjee 2020: 5), as well as the practice of taking loan payments by field officers (Muellners 2020), which is affecting the structure, regulation, investment, and participation in microfinance products.

For Anand and Chandramouli (2016: 264), repayments maintained during economic recessions increase the overall national economy and this notion was substantiated by the research literature. In M-CRIL (2020) advisory note, a Nepalese investigative research group identified that there is currently a liquidity shortfall of $400m+, a situation shared amongst 75% MFIs arising from 70% disbursements and anticipated 90%+ increase from the return of migrants; a situation alleviated by a 30% increase in borrowings over the previous fiscal year (M-CRIL 2020: 3). As the aforementioned challenges affirm, MFIs are undergoing issues of instability, processing, and credit referencing models that an integrated open-source system that is simple to use and maintain could alleviate and, further, assist in the sustainability of the sector.

Exit mobile version