Check your financial bias: How your lending business may be less accessible to women & how to fix that

In 2021-22, 8.59 lakh MSMEs run by women were registered on the Udyam Portal. (File photo of then MSME Minister Kalraj Mishra at an MSME Financing Conference in Kolkata in 2014, via Wikimedia Commons)

In India, the percentage of micro, small and medium enterprises (MSMEs) run by women in 2021-22 is 17.7 percent, as registered on the Udyam Portal. While 4.9 lakh women-led MSMEs registered on the portal in 2020-21, the number rose to about 8.59 lakh units during 2021-22.

While this is substantial growth for women-led businesses post-pandemic, there lies a massive opportunity to spur this growth. Women who run businesses largely point out to access to capital as their biggest constraint, with nearly 70 percent claiming that their financing needs go unmet. There is a global financing gap in access to finance for women-owned businesses, and this is estimated at USD 20 billion in India and USD 1.7 trillion globally.

As per information provided to the Parliament

The government has taken several steps towards facilitating women’s access to formal credit through schemes such as PM Employment Generation Program, under which women are entitled to higher rates of subsidy, and Credit Guarantee Scheme for Micro and Small Enterprises, which ensures women receive additional 10 percent guarantee coverage.

Also read: PMEGP extended till FY26 with Rs 13554 crore outlay

Banks and other financial institutions, too, have bolstered women entrepreneurs’ access to formal credit. Public sector banks have been advised to earmark 5 percent of their Adjusted Net Bank Credit for lending to women.

However, socio-economic-cultural and financial biases continue to make it difficult for women entrepreneurs to access credit, which is pertinent to scaling their businesses, despite evidence that women are:

- Strong savers

- Prudent borrowers with lower loan-to-deposit ratios

- And have a lower share of non-performing loans.

Bias in lending: How credit is less accessible to women

Typically, a loan application goes through a multi-step process. A customer learns of a credit product and she fills out an application to avail of it. The Financial Institution (FI) processes her application through a series of steps that

- Verify the accuracy

- Assess creditworthiness

- And then determine loan terms.

The process ends in a decision, which may go through a step of human scrutiny, and it is finally communicated to the applicant.

At every stage of this process, there could emerge biases in lending to women. A lack of women applicants creates imbalanced data from which an institution can learn over time what kinds of customers are creditworthy. Then, the very questions that appear in application forms may benefit men, as women own fewer properties which limits their creditworthiness. Finally, the decision-making process – whether human or algorithmic – may contain unconscious biases, resulting in loan terms being more favourable to one gender.

Apart from an unconscious bias, available credit products may be unsuitable for women. Other reasons that prevent women from accessing credit include:

- Limited access to information about Financial Service Providers (FSP)

- Insufficient support in accessing government schemes

- And discomfort with borrowing from formal sources.

Further, a lack of gender-aggregated data in the lending and MSME ecosystems restricts collective abilities to build greater credit opportunities for women business owners.

How lending institutions can fix gender biases

Checking for bias is the first step to uncovering gender imbalances in credit decisions:

- Using a simple data query, an FSP can understand where in the lending process women are most likely to be rejected.

- To check their credit process bias, an FSP can start by establishing a clear definition of fairness, such as by ensuring women applicants have an equal probability of getting credit approved or rejected on the same terms.

- Then, they can establish relevant control variables which should be indicative of creditworthiness, regardless of gender.

- Finally, evaluating for bias and assessing its magnitude can make their lending process more gender intentional.

Answering questions, like how much a bias cost your company or how many people a bias excludes or disadvantages unnecessarily, can reveal the business opportunity an FSP can unlock in lending. Women’s World Banking has devised a set of tools to detect gender bias in lending. The scorecard identifies six dimensions of fairness that FIs can check and see how they measure.

One of the issues is that women are more likely to be “thin file” customers, lacking a formal credit score. There are new ways of using alternative data to create a proxy score.

If the issue is the algorithm is prioritizing highly gendered data, there are ways to instruct the algorithm to ensure no gender bias indicators are used to determine creditworthiness.

If the issue is an unconscious bias among loan officers, we can use measures of efficacy of loan officer training and test these to improve gender fairness.

In conclusion

This is a story of opportunities for fairness and unlocking a greater business potential for India’s lending ecosystem. Given the global gender credit gap, we know that there are women highly deserving of credit who, if given the chance, can be excellent customers. Investigating biases can help diversify an institution’s portfolio by bringing in more women applicants, and more women repeat borrowers can help reduce overall acquisition costs. The trick is to know where to look for them – data science can help us do this.