Automated, scalable, and proven psychometric risk measurement tool for SMEs.

Competition Finalist

This entry has been selected as a finalist in the
The G-20 SME Finance Challenge competition.

By assessing an entrepreneur’s fundamental intellectual and psychological characteristics, this tool predicts credit risk and upside potential with the same accuracy as traditional credit scoring models, but without requiring any credit history or collateral. This statistically validated tool makes a large portfolio of small SME bank loans economically viable.

About You

Organization: Entrepreneurial Finance Lab Visit websitemore ↓↑ hide↑ hide

About You

First Name

Bailey

Last Name

Klinger

Your Organization

Entrepreneurial Finance Lab

Country

Peru, LI

About Your Organization

Organization Name

Entrepreneurial Finance Lab

Organization Website

Organization Phone

+5114478865

Organization Address

Diego Ferre 387 #J

Organization Country

Peru, LI

Organization Type

Private Institution

The information you provide here will be used to fill in any parts of your profile that have been left blank, such as interests, organization information, and website. No contact information will be made public. Please uncheck here if you do not want this to happen..

Your solution

read more↑ hide↑ hide

Name Your solution

Automated, scalable, and proven psychometric risk measurement tool for SMEs.

Describe Your Solution

By assessing an entrepreneur’s fundamental intellectual and psychological characteristics, this tool predicts credit risk and upside potential with the same accuracy as traditional credit scoring models, but without requiring any credit history or collateral. This statistically validated tool makes a large portfolio of small SME bank loans economically viable.

Country your work focuses on

n/a

If multiple countries, please list them here. If your solution targets an entire region, please select it below

Currently Peru, Colombia, Chile, South Africa, Tanzania, Kenya & Rwanda. Tool applicable world-wide

Region(s) your solution focuses on:

Africa, Latin America and the Caribbean.

Range of turnover in your target firms, in USD

Less than $1 Million, $1-5 Million, $6-10 Million, $11-20 Million.

Average turnover in USD of your target firm

1,200,000

Number of employees in your target firms

Fewer than 5, 5-24, 25-49, 50-74, 75-99.

Average number of employees of your target firm

20

Specify the size, average and range of expected loans or investments in each target firm

For bank partners: 1,000-100,000USD, $15,000 average. For SME VC partners, $150,000-$500,000, $225,000 average.

What stage is your solution in?

Operating for 1‐5 years

Innovation

read more↑ hide↑ hide

What makes your innovative solution unique?

1) Novel Approach to financial Screening.

Our tool measures future upside potential and willingness to repay, whereas traditional screening methods examine only current wealth (collateral) or past performance (credit history). Instead of only lending to small minority that possess these indicators, banks can lend to high potential, growth-generating SMEs, allowing unbanked and informal entrepreneurs to begin building formal credit history.

2) Demonstrated Scalability

a. Highly scalable within countries. Risk evaluations by experts are too expensive and time-consuming compared to the size of the loan. EFL’s automated low transaction cost tool makes a large number of smaller SME loans profitable for banks.

b. Highly scalable across countries. Our pilot tests across 7 countries and 8 languages show that the core test modules are very consistent across countries and cultures; only minor customization is needed to continue to spread the EFL tool worldwide.

3) Proven.

The EFL tool, initially developed at Harvard’s Center for International Development, modifies existing psychometric tests used successfully in research and pre-employment screening for decades. Over the past 4 years we have pilot-tested the tool on over 2,000 entrepreneurs and statistically validated its predictive power.

4) Versatile.

While our focus is on unlocking scalable bank lending to SMEs, the EFL tool can also be used to expand microfinance and venture capital for SMEs, and better target business training and support programs on high-potential entrepreneurs.

How does your proposed innovation leverage public intervention in catalyzing private SME finance?

Latent Supply and Demand for SME Lending:

Developing countries have a large number of microenterprises and some large firms, but few SMEs. In developed economies, formal SMEs account for around 57% of total employment, but only 18% of employment in low income countries (Ayyagari Beck & Demirguc-Kunt 2003). These firms are not absent only because of inefficient business environments and the costs of formality: new research shows very robustly that they have returns to capital well beyond bank’s cost of capital (e.g. Banerjee & Duflo 2004, McKenzie & Woodruff 2008). There is supply of and demand for finance; they just can’t be matched.

Blocked by a lack of information:

With little credit history, collateral, and financial statements, banks have few low-cost methods to select among applicants. EFL uses psychometric assessments as a viable low-cost, automated screening tool to identify high-potential entrepreneurs and evaluate risk and future potential.

Unlocked with a small public intervention:

Before EFL, these psychometric assessments have never been adapted for risk analysis in finance, nor has their predictive power been proven to the statistical requirements of risk departments in emerging market banks. In order to perform this adaptation, demonstrate that this solution works, and get banks to take it on in a new country, some public support is necessary, and has spurred the adoption needed for systemic change.

Enabling self-scaling:

Once banks have a low-cost way to identify the large number of SMEs with high returns to capital, lending expands massively. Understanding risk enables risk-management, and innovations such as quasi-equity contracts and risk-based pricing. A relatively small well-targeted amount of public support will allow this dormant capital and latent entrepreneurial potential to connect in a way that is profitable for those involved, and therefore self-scaling.

EFL’s catalytic effect has already been demonstrated:

Each of our first round testing partners is in some stage of deploying the assessment, (9 institutions spanning 7 countries with a total loan books of approximately $63b,). Moreover, within only 3 months of the first round tests being completed, Standard Bank has signed an agreement with EFL to launch pilot implementations in Kenya, South Africa, Ghana, Nigeria and Uganda over the next twelve months. The bank is opening new branches where every business applicant must take EFL to be accepted, and they will be lending from $1,000 to $10,000 to 20,000 SMEs that, under their existing tools and policies, would have been rejected. This means every dollar of public support invested in spreading the EFL assessment to Africa and Latin America has unlocked $20 of private capital in only its first few months, in this one partner alone.

With public help, the growth can be accelerated and amplified; the EFL tool can be adapted and spread to other countries, and billions of incremental dollars of private capital could be profitably lent to SMEs across the developing world.

What barriers does your proposed solution address?

Asymmetry of information, Lack of collateral, Lack of financial capacity, Lack of SME access to skills / knowledge / markets, Lack of institutional capacity of financial intermediaries, High transaction costs for financial intermediaries to serve SMEs, Lack of financing to women entrepreneurs, Specific barriers to fragile and weak states.

If you checked any of these barriers, describe how your solution addresses them

[Lack of institutional capacity of financial intermediaries & High transaction costs for financial intermediaries to serve SMEs]

A key barrier that our solution overcomes is the absence of low transaction-cost tools for financial intermediaries to serve SMEs. Human-based due diligence is usually too expensive compared to the loan size. The only low transaction-cost methods available to banks are currently based on crude proxies for risk, such as long credit histories and collateral, which many SMEs in developing countries do not have. Our tool solves this problem by bringing in other sources of information which every entrepreneur possesses, in a low transaction-cost way.

[Asymmetry of information & Lack of collateral]

Financial institutions cannot determine which SMEs they should lend to, in part because of asymmetric information (they lack the knowledge that the SMEs have about their profitability and likelihood of repayment). Asymmetric information also prevents lenders from utilizing risk-based pricing, quasi-equity contracts, and other basic methods to stratify and manage risk for those lacking collateral, documentation, and verifiable history. The EFL assessment solves this information asymmetry because it is highly resistant to gaming, meaning the questions are non-transparent and entrepreneurs can’t manipulate their responses in an attempt to ‘appear’ more entrepreneurial. It therefore captures both hard to measure, and reliable information about the applicant’s risk and potential.

[Lack of financing to women entrepreneurs & specific barriers to fragile and weak states]

The impact of this contribution is even greater for individuals that have little available information for banks to use, such as immigrant communities and women entrepreneurs, and in environments where the information infrastructure is almost non-existent, such as fragile states where banks have almost no trustworthy data on which to evaluate risk. For this reason, we have entered in discussions with organizations seeking to promote financial access for SMEs in conflict, and post-conflict environments such as Iraq and Afghanistan, (ICF-SME, USAID-Tijara, CHF International).

[Lack of financial capacity & Lack of SME access to skills/knowledge/markets]

Finally, our solution can be used not only to better direct lending and investment capital, but also to better direct training and capacity building for SMEs. This increases the effectiveness of these programs in their aims to overcome the other constraints to SME growth, such as financial or technical skills. One of our first-round test partners was TechnoServe, and that organization has already begun to use the EFL assessment to select the highest-potential entrepreneurs to admit to their training programs and business plan competitions across Latin America, in order to maximize their impact. We are exploring similar implementations to improve the impact of government training programs, incubators, and SME support services.

Impact

read more↑ hide↑ hide

Provide empirical evidence of your proposed solution's success/impact at present. If your project is in the idea phase, please provide evidence that speaks to its potential impact

Born from Academic Research & Empirical Results

We have spent almost 4 years developing and testing this solution. After a preliminary small-scale pilot in South Africa, our research project based in Harvard’s Center for International Development was selected as one of Google.org’s first SME initiatives to support, which allowed us to develop an assessment and test it in Africa and Latin America. Working in close partnership with private psychometric testing companies as well as academic researchers, we completed a new assessment in 2008, assembled a network of 9 organizations, (primarily banks, but also microfinance organizations, one venture capital firm, and one training organization,) and tested the prototype across 7 countries and in 8 languages. Entrepreneurs were given prizes based on their performance on the test to create a high-stakes implementation.

With the results of these tests, we can statistically compare the scores of the over 2,000 entrepreneurs that were tested to their actual business performance and loan repayment history. The results show that our assessment can predict default with an AUC (a common summary metric to evaluate credit scoring models) of 0.752 (equivalent to a Gini of 0.5), with all individual relationships significant at the 95% level. To translate this into default rates, in one sample with overall default rate of 12%, the top 5th of scorers on the EFL assessment have a default rate of just under 3.4%, while the bottom 5th have a default rate in excess of 26%. Out of time and out of sample testing reveal that the results are very stable, and have stood the test independent evaluations by risk departments at third party financial institutions.

Demonstrated Impact

Traditional corporate risk scoring models use formal financial statements and past borrowing history, which aren’t available for most SMEs. Our model does not depend on any of that information, but the resulting predictive power is equal to, and even superior to, credit scoring models for corporate borrowers.

Our current record of adoption among our pilot trial partners is demonstrated by the fact that just months after the completion of the pilot, all of our first-round partners are at some stage of implementing the tool in their operations, and illustrates the catalytic effect of this tool on access to finance for SMEs. At the same time, new partners in these regions are now seeking to roll out the tool, such as one of Africa’s largest banks, Standard Bank, which has signed an agreement with EFL for a pilot implementation across 5 African countries. So with only a few months and limited exposure to the market, we have already achieved significant impact, which could be increased exponentially with public support to expand the testing to other regions.

How many firms do you expect to reach?

Applying the current bank penetration level into the SME segment found in South Africa (1.04%) to the developing world in whole, there are approximately 127 million high-potential, unbanked SMEs. Enabling banks to identify even just the top 5th percentile of the unbanked SMEs creates direct opportunities for over 6.3 million businesses.

What is the volume of private SME finance you aim to catalyze?

In developed economies, 16% of startup capital for SMEs is financed by commercial banks. Comparing this to the current level in developing countries suggests at least $87b of dormant private capital. Catalyzing just a small percentage of this opportunity would impact millions of lives directly, and even more indirectly.

What time frame will be required to reach these targets?

The first EFL implementations have required 12 to 24 months to localize the tool, validate its predictive power, and integrate into existing bank processes and policies. This can be shortened in subsequent implementations using lessons learned and our skilled and experienced field team. And due to EFL’s decentralized structure, scaling across multiple markets can occur in parallel; the timeframe is a function of the resources and network of implementers we can reach in new countries. If EFL can add 4 new partners each with an incremental customer target of 70,000 per year, 6.3 million is achievable in just over 6 years.

Does your solution seek to have an impact on public policy?

Yes

What would prevent your solution from being a success?

1. Regulatory delays. Regulations that govern the use of psychometric assessments, as well as those that define which types of information can be used for extending credit could, in some jurisdictions, limit implementation. We have overcome these barriers with partners in our current countries of operation, but they may be more problematic in other countries and take longer to resolve.

2. Speed of adoption in new countries. In our current countries of operation, partner banks have taken measures to adjust their policies and procedures to take full advantage of the EFL tool in new market segments. But, in general, banks tend to be conservative and slow-moving institutions, and it is possible that partners in new regions or countries might not move as quickly as our current partners in Latin America and Africa.

3. Limited resources. Our current span of operations has been limited by our financial resources, our network of contacts, and our contacts through Harvard’s Center for International Development. If we are unable to connect with financial institutions and obtain financial support to adapt and validate the tool in new countries, this could slow global scale-up of the tool and limit us to our existing pilot countries.

Describe the social impact of your innovation. Please include both numbers and stories as evidence of this impact

Macro Impact

The low transaction cost, highly scalable nature of EFL enables rapid global adoption and therefore, significant social impact. If the distortion in the firm size distribution in developing countries was removed and the distribution regularized, thus “filling” the missing middle, estimates suggest GDP across developing countries increasing by over $3.6 trillion dollars annually. If you estimate the number of SMEs ‘missing,’ this shift would create millions of new SMEs in a handful of our countries of operation alone: 2 million in Nigeria, and over half a million in each of Kenya, South Africa, and Colombia. Filling in the missing middle SMEs would have a significantly larger impact than that of the microfinance revolution, given the much larger employment and growth contribution of SMEs compared to informal microenterprises.

Impact to Date

Up until early 2010, our focus was on calibrating the psychometric assessment, pilot testing its effectiveness, and data gathering; the tool was not used for decision making but rather was tested on existing bank clients. However, soon after the pilot concluded, rapid adoption began. Starting this August, 4,000 SMEs in Kenya will be approved for term loans of up to $10,000 based on the EFL assessment, even though they don’t meet the bank’s traditional requirements for borrowing history and collateral. Similarly, in South Africa starting in September, 4,000 SMEs will be offered overdraft accounts at account opening, based on the EFL assessment, instead of the traditional requirement of a 6 month waiting period to establish account history. These 8,000 newly-funded entrepreneurs demonstrate the tremendous social impact that the EFL innovation can have if expanded: banks relaxing traditional policies of SME lending based on new predictive power, massively expanding access to finance, and creating macroeconomic and employment growth.

Sustainability

read more↑ hide↑ hide

List all the funding sources that are required for the sustainability of this solution

This solution only requires some initial public support in order to adapt the tool to the new country and culture and to demonstrate its effectiveness though small pilot trials, followed by efforts to communicate the potential and share the pilot results with local financial institutions. Once this is done, we have demonstrated in our current countries of operation that private funding is catalyzed, as financial institutions themselves are now funding further pilot implementations and scale up there. Only a small amount of external funding for the first 12-24 months is required to enter new countries, before similar results of unlocking $20 of private bank funds for every $1 can be expected within the first few months alone.

Demonstrate how your proposed solution has the capacity to graduate from dependence on public finance. What is the time frame?

Demonstrated Graduation from Research Stage

EFL was initially a research project at Harvard University supported by public funds. This support was necessary to acquire licenses to various psychometric assessments, translate them into 7 languages, acquire the necessary hardware and software, build a network of pilot partners and implement the tests. Although financial institutions are very hungry for a profitable solution to open up lending to SMEs, they were not willing to make these upfront investments, both because of uncertain benefits of this unique approach, and also because those benefits are spread over the financial system as a whole rather than accruing only to them.

But once these pilot trials are completed, individual financial institutions have proven very willing to not only fund further pilot testing, but also to make significant amounts of capital available for lending to SMEs based on the tool. For example, in the next few months Standard Bank is placing a portfolio of $16 million USD at risk for new overdrafts to previously rejected SMEs in South Africa and over $4 million USD in term loans for previously rejected SMEs in Kenya based upon our trial results. In only the first months following the pilot, each dollar of public finance invested in the project has already catalyzed private capital more than 20 times greater. In Latin America and Africa, our organization has graduated from a research project completely dependent on public finance to a private company in Peru spun off of the University, with the majority of our funding coming from the banks themselves.

Accelerated Expansion and Refinement with Public Support

Each incremental data point will ensure greater test accuracy, more versatile applications of the results, and lower per-test costs. Lower costs and increased accuracy will directly create larger social impact. Greater volumes of test and performance data allows for robust research and development into shortening the test, incorporating more test modules, and collaborating with both public and private institutions alike to revolutionize the way risk and opportunity are measured. We could therefore generate greater impact if we could replicate our approach in a larger number of countries simultaneously using public seed funding, and our track record suggests that only 12-24 months of public support is necessary before a new region becomes self-sustainable. Additional public funding now will allow for geographic expansion, more accurate results, and a more sustainable organization.

Demonstrate how your proposed solution will survive a potential loss of its largest private funding source

Our organization was initially supported by a two-year grant from Google.org and SNV in 2007, which has since expired. Although we have obtained some other small research grants over the past year, our largest funding source is now the financial institutions themselves in our test countries that are seeking to implement and scale up the tool. We are therefore not dependent on any single funding source, either private or public, and while our global growth trajectory would be truncated, the loss of any one of our pilot partners would not have a major effect on our operations in other countries.

Please tell us what kind of partnerships, if any, could be critical to the greater success and sustainability of your innovation

The most important partnerships to scale up the success of our tool are with more implementing organizations: financial institutions (including microfinance institutions and SME venture capital funds, but primarily commercial banks) and other capacity builders (for example, incubators and entrepreneurship training programs). These organizations would generate additional data that we could use to validate and improve the assessments, and customize them for new countries, business segments, and applications, and through scale up successful pilot trials they would unlock larger amounts of private capital for SMEs.

In addition to partnerships with implementing organizations, we could accelerate the roll-out of our solution to a wider set of countries if we could partner with public and private funders. These could both be public resources to fund the initial testing and research in new countries, and private investment in our organization to more rapidly and aggressively market the commercializable product to financial institutions.

Are there non-financial issues that could threaten the sustainability of your proposed solution?

Test resilience over time

One concern when launching this project was the sustainability of a psychometric screening tool over time: that the test could eventually be learned and ‘gamed’, thereby losing its predictive power. As such, we limited ourselves from the outset to psychometric tests that are non-transparent and ‘game-proof’, meaning that even if a single test taker took it multiple times, or colluded with a loan officer, they would not be able to game the test to mimic target profiles and inflate their score. Despite this, in order to assure sustainability, we are focused on continually testing new and different psychometric tests. This will ensure that the power of our test is increased, rather than reduced, over time.

Regulatory Environment

Related to this challenge are regulatory requirements. In order to keep the test non-gamable, it is necessary to protect test content and scoring algorithms, which creates some tension with regulatory aim of increasing model transparency under the Basel II guidelines. In partnership with the legal departments of our testing partners, we have been examining these requirements and are confident of our tools’ consistency with Basel II and the national regulatory regimes in our countries of operation.

Please tell us if your proposed solution aims to scale up through a high growth sector, expand immediately to multiple sectors, and/or scale up geographically

The tool is immediately applicable across SME sectors and countries as well as across institutions that could use it: microfinance institutions, commercial banks, SME venture capital funds, and capacity builders. Our sample crossed countries, cultures, and all major economic activities of SMEs in developing countries, including commerce, services, production, and agriculture, and the results reveal only minor changes across these dimensions. The major customization required is across business sizes: the risk factors and the psychometric profile of a successful owner/operator of a small business with 3 employees and turnover of under $250,000 per year is rather different from that of business with 25 employees and turnover of $10 million per year. Armed with our first round test results, we have developed the customizations required for business size.

Therefore key dimension for future expansion is geographic. We now have a major presence in Peru, Colombia, South Africa, and Kenya, with smaller-scale testing completed in Chile, Tanzania and Rwanda. Additionally, EFL has been contracted to expand to Nigeria, Ghana, and Uganda before the end of 2010. Beyond scaling up in these countries, we hope to expand to new regions. We have already begun searching for opportunities in the Middle East and North Africa region, particularly in Iraq and Afghanistan, given the augmented value added of our tool in post-conflict environments with poor information infrastructure. We have also been in preliminary discussions with commercial banks (most notably Banco Santander) and capacity builders (most notably Endeavor) in Brazil. In the medium term, we hope to expand to Pakistan India and Bangladesh, as well as China and East Asia.

AttachmentSize
eflevidence.jpg182.69 KB
eflimplementation.jpg204.63 KB
eflsolution.jpg186.97 KB
179 weeks ago Matt Guttentag said: On October 27, 2010, the judges reviewed entries for the Changemakers G-20 SME Finance Challenge competition and would like to pass on ... about this Competition Entry. - read more >
181 weeks ago Automated, scalable, and proven psychometric risk measurement tool for SMEs. has been chosen as a finalist in The G-20 SME Finance Challenge.
190 weeks ago updated this Competition Entry.
191 weeks ago J. Skyler Fernandes said: Hi Bailey, Spoke with DJ not to long ago regarding EFL. Was great to learn about the work you are doing. It is intriguing to see if ... about this Competition Entry. - read more >
192 weeks ago Renee Lee said: Psychometric testing measures always intrigue me. We are investigating the establishment of a social development bank, with new ... about this Competition Entry. - read more >
193 weeks ago updated this Competition Entry.
193 weeks ago submitted this idea.