Legacy credit assessment doesn’t work for MSMEs. As the backbone of societies everywhere they contribute to local and national economies and to sustaining livelihoods, MSMEs account for 90% of businesses, 60 to 70% of employment and 50% of GDP worldwide.
Despite this, MSMEs face difficulties in obtaining loans from financial institutions that use a conventional approach to creditworthiness. These difficulties include a lack of financial and operational data, regulatory gaps, and ineffective risk management capabilities of the financial institution. Additionally, conventional credit assessment is also considered inefficient and costly for loan processing, especially when it involves low-value loans.
To address these issues, various types of information from third-party data providers, known as “alternative data” or “modern data” can be used to determine financial solvency, willingness to repay, and creditworthiness of productive units.
This data provides additional information that allows a holistic view of the applicant for the decision to grant a loan. There are several ways to implement modern data, and the use cases depend fundamentally on the objectives of the financial institution. For example, lenders may want to:
The variety of modern data sets on the market are constantly changing, making it difficult to analyze them and select the ones that are truly useful. Four useful criteria for determining the value of data sources are:
In this article, we will look at 7 modern data sources, describe how they apply to credit analysis in micro businesses, and what advantages and disadvantages they might offer.
Social networks are popular among SMEs as tools to promote themselves to the public and acquire new customers.
Platforms such as LinkedIn and Facebook collect profiles, report information about businesses, their most recent initiatives, as well as their interactions with their customers. This data is used to generate models that estimate the value of a company.
Advantages
Social Networks offer a high volume of information due to the vast data points resulting from activity on accounts.
Some success stories in consumer loans such as Vkontakte’s for credit scoring, improving repayment prediction by 18%.
Can be used as a proxy to identify payment capacity by inferring purchasing and consumption behaviors.
Disadvantages
Technological and compliance barriers can make it difficult to access this data.
No use cases in microbusinesses lending so there are no records on best practices nor standards.
The use of social data could be fragmented and would require complex models and eventually with different vendors.
Events such as those that occurred in the Cambridge Analytica case; where as a result of this scandal there were fines and regulatory changes in the use of network data, significantly decreasing functionality, predictive power and in general the diminishing relevance for credit assessment use cases.
Psychometrics is based on the fact that people have an enduring personality disposition that partly predicts what they will do. There is a science behind personality and how personality is measured and what it can predict (Nettle, 2020).
In the context of credit assessment, psychometric tests or assessments can be used to understand an individual’s willingness and ability to pay. The focus on willingness to pay makes it possible to determine whether the applicant has the propensity to honor their debts.
Advantages
As a data source, psychometrics has the widest coverage, this is because all people have a personality that can be evaluated (Nettle, 2020).
Multiple success stories in loans to individuals and MSMEs without credit history and informal sector (Mondato, 2023)
Feasibility was demonstrated with improved predictive power (Arráiz, Bruhn, Ruiz, & Stucchi, 2018)
The assessment can be applied directly to the applicant to ensure specific data of the individual.
It is the only source that contributes to the analysis of willingness to repay.
Disadvantages
The data collected is unique and depends on the capabilities and expertise of each vendor.
There may be friction in the process if the evaluation is unpleasant or too long.
This source does not provide financial information for payment capacity analysis.
Open Banking enables FIs to access customers’ banking and non-banking transactional data using APIs. Around the world, this trend is rapidly evolving.
There are different models in which FIs can develop new products based on data that they don’t normally have access to, but that FI’s and retailers with financial operations now share through APIs.
Advantages
Growing industry with new data aggregators generating new agreements with financial and non-financial institutions to share data.
Regulations in different countries are driving the implementation of Open Banking.
Cases show an increase in prediction accuracy between 10%-20% when combined with traditional models (Birch, Cummins, & Shin, 2018), (Equifax, 2019)
Disadvantages
Limited coverage because access to data depends on agreements between data providers and aggregators.
Some microbusinesses segments are still cash intensive so may posess little operational information.
It does not solve the problem of coverage of microbusinesses without credit history in the information centers.
The micro-entrepreneur can decide to share data only from certain institutions at his/her convenience. This can intentionally skew the analysis.
Through the registration of the customer’s telephone number, it is possible to know and evaluate the behavior that the applicant has had with their telephone line.
Accessible data could be prepaid/postpaid, data usage, geolocation, recharge history, call patterns and SMS (among others)
Advantages
Pilots have obtained accurate credit scores (Dietrich, De Souza, & Guerreiro, 2020).
Particularly useful for generating collection strategies.
Complementary source to enrich data for fraud detection.
Disadvantages
Difficulties in establishing agreements with telco operators.
Limited availability due to local data protection and privacy regulations.
Decreasing availability as lower tier segments and younger population often switch to substitute and cheaper alternatives to phone calls (e.g. WhatsApp call and voice message)
Coverage is only available for users and micro-businesses that use a specific operator.
Refers to trackable data from a mobile device, which is accessed through an App installed with the user’s permission.
Accessible data may be location, coarse location, personal information, messaging usage and patterns, media library, calendar, contacts, activity in other apps, search history, etc.
Advantages
Users are adopting banking apps more and more, implying wide availability.
Immediate response time.
High adoption of smartphone devices.
Smartphone digital footprint can be used for behavioral analytics and personality proxy as well.
Increased capabilities of vendors to develop device SDKs.
Disadvantages
Device memory. Some population segments are prone to delete Apps due to lack of memory on their devices.
Difficulties in the specificity of the information since the activity of the device is not segregated between business use and personal use of the device user.
Vendor requires behavioral economics expertise to interpret data and develop actual usable models for lending.
The use of email data has become widespread because entering an email is part of the minimum data required of credit applicants. This mechanism makes it possible to assess the consistency of the user’s data.
Advantages
Offers a high volume of information due to the vast data points resulting from activity on accounts.
Enables fraud detection (Center for Financial Inclusion at Accion & Institute of International Finance, 2018).
Disadvantages
Difficulties in the specificity of the information, not all micro-businesses have a business account, sometimes the personal account of the micro-entrepreneur is used.
Depends on a high volume of data, hence increasing technological efforts.
There are no concrete use cases in credit risk for microbusinesses.
Some pilot use cases but for fraud prevention mainly.
This is a G2B type of data source that refers to records of microbusiness activities. Data subsets may include utility bills, health, pension and risk contributions, and tax information.
They usually include payment-related information (cash flow) and non-money-related information.
The behavioral trends generated by the analysis of this information can be used to evaluate the revenue and most recent financial status of a microbusiness.
Advantages
During pandemic several countries dispersed economic aid, subsidies and incentives for microbusinesses which generated large volumes of financial data.
In some regions there are some aggregators and open data vendors that provide this data.
Government databases can be used to enrich operational risk / fraud policy.
Disadvantages
Microentrepreneurs in emerging economies often exist ‘off the grid’ from government tax, health and pension programs.
It is common to find inconsistencies between account holders and the actual microbusiness that uses the service (e.g, Merchant pays rent and all utility bills under the owner’s name)
Difficulties in establishing inter-institutional agreements (e.g. utilities and data vendors)
Limited availability due to local data protection and privacy regulations.
In today’s data-driven world, microbusiness owners and individuals alike increasingly recognize the value of their data and seek control over how it’s utilized to achieve personal and business goals.
Financial inclusion aims for a “democratisation of credit,” achievable through the “democratisation of data,” reflecting a growing trend of individuals asserting consumer power to decide who accesses their data and how it’s leveraged for more meaningful financial products.
Modern alternative data sources offer a powerful means to this end, providing insights beyond traditional metrics for assessing creditworthiness. Despite their potential advantages, challenges such as data availability, privacy concerns, and technological barriers persist therefore selecting the right data source and the right partner requires careful consideration to enhance financial inclusion, mitigate credit risks, and empower microbusinesses and individuals in their financial journeys.
Arráiz, I., Bruhn, M., Ruiz, C., & Stucchi, R. (2018). Are Psychometric Tools a Viable Screening Method for Small and Medium Enterprise Lending. Private Sector Series, TN(5).
Birch, D., Cummins, W., & Shin, C. (2018). How Open Banking Can Drive Innovation in Financial Services. Harvard Business Review, 96, 18-19.
Bocconi, S. B. (2015). Social glass: a platform for urban analytics and decision-making through heterogeneous social data. In WWW2015Companion, WWW2015Companion (pp. 175–178). New York, NY: Association for Computing Machinery.
Center for Financial Inclusion at Accion & Institute of International Finance. (2018). Accelerating Financial Inclusion with New Data. MAINSTREAMING FINANCIAL INCLUSION: BEST PRACTICES SERIES, Part 4.
Dietrich, L., De Souza, F., y Guerreiro, A. (2020). Development of credit scores with telco data using ML and agile methodology in Brazil, 4831. S Global Forum 2020.
Equifax. (2019). Knowledge Centre. Retrieved from Open Banking & Credit Risk: Friend or Foe?: https://www.equifax.co.uk/resources/open-banking-credit-risk-friend-or-foe.html
Mondato. (Nov 15, 2023). mondato.com. Retrieved from Psychometrics: What makes a reliable borrower?: https://blog.mondato.com/psychometrics-reliable-borrower/
Nettle, D. (2020). Personality: What Makes You the Way You Are. HighBridge.
Galvanizing MSMEs worldwide by supporting women and youth entrepreneurship and resilient supply chains: https://www.un.org/en/observances/micro-small-medium-businesses-day
SaaS technology that provide character-based credit scores for Banks, Micro Finance, Digital Lenders, Neo Banks, BNPL and Asset Financing.
