Adverse selection

Topic: Banking Regulation, NPAs and Financial Stability · NCERT: Beyond NCERT

Meaning

Adverse selection happens when one side of a deal knows more about its own risk than the other side, and this hidden information pulls in the riskiest people. For example, the borrowers most eager to take a loan, or the people most eager to buy insurance, are often the ones most likely to default or make a claim. The key point is that the problem exists before the deal is signed. It matters because it can make lenders charge more, lend less or refuse good borrowers. This is why small firms and MSMEs often cannot get credit even when interest rates are low.

Explanation

How it works: hidden information before the deal

• Asymmetric information means the borrower knows more about their own risk than the bank does.
• The bank cannot tell a safe borrower from a risky borrower, so it charges both the same average rate.
• The average rate drives out the safe borrowers:
• For a safe borrower, the average rate is too high, so the loan feels like a bad deal.
• For a risky borrower, the same rate is cheap, because they may not repay anyway.

• So the pool of people applying gets riskier. The more eager someone is to borrow, the more worrying they are.

• Origin: Akerlof, "market for lemons" (1970).

• Buyers of used cars cannot tell good cars from bad ones ("lemons"), so they offer only an average price.
• Owners of good cars then refuse to sell, and mostly lemons remain.

• The market can shrink, or even collapse.

• Insurance version: people who know they are unhealthy are more eager to buy health insurance. Premiums rise, healthy people drop out, and the pool gets sicker still.

Stiglitz-Weiss (1981): why banks ration credit

• Credit rationing means the bank refuses some borrowers outright instead of simply raising its interest rate.
• The chain of effects:
• The bank raises its interest rate.
• Safe borrowers, whose projects earn modest returns, drop out.
• Risky borrowers stay in.

• Defaults rise, so the bank's actual income can fall.

• Worked example:

Loan rate	Share of loans repaid	Expected return per Rs 100 lent
10%	95%	0.95 × 110 = 104.5
14%	88% (safe borrowers have left)	0.88 × 114 = 100.3

• The higher rate earns the bank less.

• So the bank keeps the rate lower and turns some applicants away.

• Who suffers: small firms and MSMEs. They have short credit records and little collateral, so the bank cannot prove they are safe.

What reduces it: fixes that reveal or screen risk

• Collateral (an asset pledged to the bank): the borrower has something to lose, so risky borrowers are less keen to apply.
• Credit scores: a record of past repayment lets the bank tell safe borrowers from risky ones.
• Relationship lending: the bank gets to know the borrower over many years.
• Lending against someone else's credit: the bank looks at a stronger party's record instead of the unknown borrower's. Letters of credit and invoice discounting work this way.
• What makes adverse selection worse: thin credit data, no collateral, very high interest rates and borrowers who are new to the bank.

In India

• Credit information companies (CICs) collect borrowers' repayment histories from lenders and produce credit reports and credit scores. This reduces hidden information.
• The law behind them is the Credit Information Companies (Regulation) Act, 2005. The RBI licenses CICs.

• There are four CICs: TransUnion CIBIL, Equifax, Experian and CRIF High Mark.

• TReDS (Trade Receivables Discounting System) is an electronic platform where MSME invoices on large buyers are auctioned to many financiers.

• It avoids adverse selection because financiers rely on the large buyer's creditworthiness, not the small MSME's.
• It began with RBI guidelines in 2014, and the platforms have operated since 2017 [3]. The platforms are RXIL, M1xchange and Invoicemart.
• Buyers with turnover above Rs 250 crore (earlier Rs 500 crore) must join TReDS, as announced in the Union Budget 2024-25 [2].
• CPSEs must now settle all MSME invoices through TReDS [3].

• Invoices discounted rose from Rs 40,000 crore (2021-22) to Rs 3.47 lakh crore (2025-26) [3].

• Collateral rules as a screen: loan-to-value (LTV) caps make borrowers put in their own money, so they have something to lose.

• Under the RBI's 2025 gold and silver loan directions (effective 1 April 2026), the maximum LTV is 85% for loans up to Rs 2.5 lakh, 80% for Rs 2.5-5 lakh and 75% above Rs 5 lakh [1].

• Letter of credit (LC): the importer's bank promises to pay the exporter. The exporter relies on the bank's credit instead of an unknown foreign buyer's.

• Factoring: a firm sells its receivables to a financier at a discount. This is governed by the Factoring Regulation Act, 2011, which was amended in 2021.

Don't confuse with

• Moral hazard: adverse selection is about hidden risk before the contract (who applies). Moral hazard is about risky behaviour after the contract, for example a borrower taking wild risks with a loan once it has been received.
• Asymmetric information: this is the broad cause, meaning one side knows more. Adverse selection is one result of it, and moral hazard is another.
• Credit rationing: this is the bank's response, refusing some loans instead of raising rates. Adverse selection is the reason for it (Stiglitz-Weiss, 1981).
• The idea that "a higher rate always means more bank income": this is false when adverse selection exists. A higher rate can lower expected income by driving out safe borrowers.

Prelims Hooks

• Akerlof (1970), "market for lemons" is the origin of adverse selection. Stiglitz-Weiss (1981) explains credit rationing through adverse selection.
• Adverse selection is a problem before the contract. Moral hazard is a problem after it. This is a classic swap trap.
• Under Stiglitz-Weiss, raising the loan rate can reduce a bank's expected return, because safe borrowers leave and defaults rise.
• The CIC (Regulation) Act, 2005 governs credit information companies, which the RBI licenses. The four CICs are CIBIL, Equifax, Experian and CRIF High Mark.
• Fixes for adverse selection in lending are collateral, credit scores and relationship lending.
• TReDS: onboarding is mandatory for buyers with turnover > Rs 250 crore (earlier Rs 500 crore) [2]. It lends against the buyer's credit, not the MSME's.

Mains Points

• The MSME credit gap: Stiglitz-Weiss rationing explains why MSMEs cannot get credit even when interest rates are low. Banks cannot judge their risk, so they refuse loans instead of pricing them. Fixes that need no collateral include CIC data, TReDS (invoices discounted rose to Rs 3.47 lakh crore in 2025-26) and factoring. These tools use the buyer's credit instead of the MSME's assets, so better information, not just cheaper money, is the answer [3].
• Inclusion vs stability: collateral and strict screening reduce adverse selection, but they shut out the poor, who have few assets and no credit history. The 2025 gold-silver directions show this balance. Small borrowers get a higher LTV (85%) to support inclusion, while bullet-repayment loans face tighter rules to protect lenders and consumers [1].
• Data as public infrastructure: wider credit reporting and digital records of cash flows can turn "invisible" borrowers into scored ones. This reduces adverse selection, but it raises questions of data privacy, errors in credit reports and how borrowers can get mistakes corrected. These are useful GS-III points on financial inclusion and regulation.

Related concepts

• CASA ratio
• Credit-deposit ratio
• Base rate
• Marginal Cost of Funds based Lending Rate
• External Benchmark Lending Rate
• Loan-to-value ratio
• Credit information company
• Letter of credit
• Factoring
• Trade Receivables Discounting System

Read more

• Bank business: funding, the price of credit and credit delivery