The subtle tax on good borrowers

In credit risk, adverse selection is usually discussed in the context of underwriting. The classic story is simple: lenders cannot perfectly distinguish good borrowers from bad ones, so riskier borrowers are more likely to seek credit, get approved, and dominate the portfolio. But there is another form of adverse selection that receives far less attention, even though it is just as damaging and often more subtle. It happens in loan pricing. 

This occurs when good, consistent borrowers are charged higher interest rates than their true risk warrants, while riskier borrowers are charged lower rates than they should be. On paper, the pricing structure may still look “market based.” Commercial teams may even call it competitive. But beneath the surface, the lender is quietly distorting its borrower mix which gradually drives away the best clients and attracts more of the weaker ones. That is the hidden danger of getting pricing wrong: the lender does not merely misprice loans; it misprices who it wants to keep. 

While institutions rarely admit it, this practice is widespread, fueled by various strategic maneuvers. Traditional banks seeking market entry, digital lenders chasing rapid volume, and private lenders targeting underserved segments all face pressure to align with external benchmarks. Consequently, rates often set according to “market appetite” rather than internal risk data, leading to management’s standard justification that their pricing merely reflects current market rates. The problem is that markets do not automatically produce risk-sensitive pricing. They often produce commercially convenient pricing, which can mask a dangerous cross-subsidy. 

Strong borrowers, i.e., those with stable repayment behavior, low default propensity, and consistent credit performance, are usually more price-sensitive than lenders assume. Good borrowers tend to have alternatives. They can compare offers, refinance, negotiate, or simply walk away. If they sense that the offered rate is not aligned with their quality, they leave quietly. They rarely complain or explain. They simply migrate to a better-priced lender. 

On the other hand, weaker borrowers behave differently. They may accept a rate that still looks affordable on the surface, even if it is too low relative to the actual risk being transferred to the lender. Some have fewer alternatives. Others value access more than price. Either way, the lender unintentionally becomes more attractive to the wrong segment. This is adverse selection through pricing. 

Over time, the dynamic becomes self-reinforcing. The best borrowers exit or never enter while marginal borrowers stay and accumulate. Delinquencies creep upward, and collections become heavier. In the end, lifetime profitability disappoints. The lender then responds with broader repricing or tighter underwriting, which can further punish the remaining good borrowers. By then, the institution faces not just a pricing issue, but a portfolio composition problem created by its own pricing logic. This is exactly where pricing and expected credit loss (ECL) should meet, but too often they do not. 

In my years of experience developing and validating ECL models, I have often observed a structural disconnect. The ECL model is treated as a risk measurement framework for accounting, provisioning, and regulatory governance. Pricing, meanwhile, is treated as a commercial decision informed by the market. Both may be technically sound in isolation, yet they often fail to connect meaningfully. This separation is a mistake. 

At the most basic level, loan pricing should reflect the economics of risk-taking. A lender’s rate consists of several building blocks: the funding or risk-free base, operating costs, required return on capital, and the credit spread. That credit spread should not be based primarily on instinct, legacy rate cards, or the vague claim that “this is what the market charges.” It should be grounded in loss expectations. 

The most disciplined internal estimate of expected loss sits in the ECL framework. The ECL model, despite its accounting purpose, already contains many ingredients relevant for pricing: probability of default, loss given default, exposure dynamics, segmentation, and forward-looking risk assessment. ECL outputs should not be copied mechanically into pricing because competition, product strategy, customer experience, and business objectives still matter. But the credit spread should at least be anchored to the institution’s own view of expected loss and risk differentiation. If pricing ignores this anchor, good borrowers end up subsidizing riskier ones without anyone clearly acknowledging it. 

Banks often resist risk pricing by arguing that loan rates must follow the market, especially in competitive consumer and SME segments. This is fair up to a point. But “market-based” can become an excuse for weak pricing discipline. It can hide the fact that the institution has not properly tested whether its rates are earning an adequate risk-adjusted return across borrower types, score bands, channels, and tenors. 

The right question is not whether pricing is market-based but whether market-based pricing is economically sensible for the lender’s own risk profile. Such distinction matters. 

A lender may match competitors on price and still underprice risk if its borrower mix, underwriting standards, recovery experience, or servicing model differ materially from the market. Similarly, it may overcharge low-risk borrowers because its pricing architecture is too blunt to distinguish quality properly. In both cases, the lender is not really pricing risk; it is pricing convenience. 

This is dangerous because the consequences of pricing mistakes do not arise immediately. Decisions are made at origination, but evidence of their soundness emerges later through delinquency trends, cure rates, restructures, vintage loss performance, and profitability. By the time management sees the pattern, the portfolio has already been built. This is why strong loan growth should never be mistaken for proof of sound pricing. Fast take-up can signal an attractive niche, but it can also mean the lender is too cheap for the risk it is absorbing. In credit, popularity is not always strength. Sometimes it is a warning. 

The ECL and pricing models are different tools describing the same credit reality from different angles. One estimates the expected cost of credit risk, while the other determines whether the lender is being paid enough to bear that risk. If those two views remain disconnected, adverse selection will not stay theoretical. It will appear in borrower mix, portfolio performance, and eventually profitability. 

That is why pricing discipline matters more than many lenders realise. When pricing is detached from risk, the lender may think it is winning market share. In truth, it may be doing something far more dangerous: asking its best borrowers to pay for its weakest ones. 

That strategy rarely ends well. Sooner or later, the good borrowers notice. And when they do, they exit quietly, leaving behind the true cost of getting pricing wrong.

As published in The Manila Times, dated 22 April 2026