Why millions of creditworthy borrowers can’t get loans

By Luca Terragni, Co-Founder and Chief Revenue, Prestatech

Most narratives on the ongoing loan rejection crisis are framed around risk paired with unforgiving constraints and interest rates. However, a deeper look reveals undeniable infrastructural issues that ultimately lock banking and finance in a downward spiral.

Although the crisis is expected to worsen, underwriters and lenders continue to lean on outdated, manual processes to make decisions. By doing so, they restrict themselves to backward data as part of their assessments of prospective risk. That misalignment has led to a growing portion of SMBs and individuals being excluded as credible borrowers, and this must be addressed now by the industry.

The current system makes more borrowers ‘invisible’ 

Credit bureau scores, balance sheets, and financial statements: these make up the backbone of traditional underwriting and lending decision-making, but they reflect a system that was created for a bygone era. A mismatch between hopeful borrowers’ financial health and lenders’ assessments of them is the biggest driver behind the rise in loan rejections. Within the market dynamics of today, though, lending decisions increasingly need to be forward-looking and dynamic.

Because of this, insights and opportunities on borrowers and loans slip through the cracks. This setup also means that underwriters and lenders are dealing with static data to make decisions in a dynamic, often volatile, market. The models are designed for well-established, well-documented borrowers, and more borrowers slip under the radar in today’s world, where SMBs scale fast, gig work is rising, and entrepreneurs are growing in numbers. The result is a system that’s structurally biased for rejection, where smaller loans are dismissed as less profitable to write.

The data is there, but it’s not being acknowledged

Data needs have evolved in line with market behaviors. Most people now have multiple bank accounts. Consider the rising trend of ‘soft switching’, where account holders are quietly changing banks. This has significant implications from a data standpoint. Money is being made and moving that banks fail to pick up on. So, the aforementioned indicators — balance sheets and credit bureau scores — used to assess borrowers and their associated risk actually fail to keep up with their behaviors.

There are other roadblocks to accurate insights shaping lending decisions. The first is fraud, which is the trend dominating nearly every conversation with banks that we work with. Financial institutions are more risk-averse than ever before, with worries magnified because fraud is becoming sophisticated at a rate that far outpaces banks’ detection systems. Within this environment, banks are spending enormous amounts of energy and budget on fraud prevention, which makes sense, but it also creates a chilling effect on legitimate lending.

The second barrier is infrastructural. Lenders now pay a lot more attention to what the money will be used for. There’s a clear gravitation toward purpose-driven lending, green projects, asset-backed financing, and supply chain investment. Lenders are more scrupulous in their assessments of how funds will be used instead of focusing on whether a borrower can repay the balance or not.

That has created a misalignment between risk appetite and confidence, affecting what banks are able to assess. To know whether a borrower can repay a debt in say 12 to 16 months, lenders need to understand their cashflow dynamics, income stability, expense patterns, concentration risk, and seasonal cycles. And these insights should be in real time, not from a document that’s already six months old. The bottom line is that lending decisions must be forward-looking, and a traditional underwriting model fails to facilitate that.

Banks struggle to act on recognised constraints

Ironically, most financial institutions are aware of the shortcomings of their current assessment practices and models. The industry is getting called out for its ‘technological inertia’ that’s holding it back when it comes to credit scoring and loans.

Moreover, finance is notorious for having a large number of legacy systems, which make changes difficult to implement at scale. Banks, especially larger ones, run on legacy core systems that are often decades old, sometimes written in languages like COBOL. Their credit models are deeply embedded in these architectures, so a new analytics layer cannot just be bolted on and expected to function smoothly.

There is also the added constraint of decision-making processes that are ingrained and tightly governed across all operations and workflows. Of course, this is necessary for an industry constantly battling fraud and cyberattacks, but it also bottlenecks modernisation and digital transformation. Alongside this, external regulatory frameworks prioritise auditability and consistency, meaning that adopting new technologies becomes a daunting undertaking. It’s a headache many organisations are not willing to take on.

Arguably, however, the harder challenge is cultural. There is a widespread misconception that automation means fewer people. Industry stakeholders are resisting the integration of automation into workflows due to fears of regulatory and workforce pushback, particularly in markets with strict labour protection laws.

Change is necessary

Fundamental changes in approaches and attitudes to lending processes are necessary to alleviate this crisis. Fortunately, there are already lenders taking steps in that direction, including how they analyse and assess potential borrowers and measure their financial health. A broader view of financial behavior and metrics beyond static snapshots from outdated indicators is becoming more prevalent — but this adoption is still in the early stages across the industry.

Alongside that, institutions must nurture awareness of the true nature of automation’s role within modernised models: that of increasing capacity and accuracy, but not replacing people. Technology should be used to augment underwriters’ and lenders’ performance, not to usurp their much‑needed role in the process.

Crucially, when underwriters and lenders are able to look beyond static metrics to more minute, real-time indicators such as cashflow consistency and transaction patterns, fewer potential borrowers are missed. Because of the nature of these insights, the data used is more accurate and more explainable. Lenders must be able to interrogate, verify, and justify outputs when making decisions.

The goal is not to expand risk around borrowing and lending. What matters is enhancing the accuracy of assessment for better alignment between credit decisions and the reality of each financial situation. This is what will even the playing field for borrowers and lenders alike.