A law for bad loans is not the same as a market for them

At the end of 2025 a number moved in the right direction. Bangladesh's non-performing loans, which had hit Tk6.44 lakh crore, or 35.73% of all lending, in September, fell to 30.60% by December, a drop of Tk87,298 crore in three months.

Not one taka of that came from a loan being sold, recovered, or written off against a buyer. It came from rescheduling. Around 1,300 borrowers made their loans current again with a two percent down payment and a decade to repay the rest. The loans did not get better; they were relabelled. And the relabelling did not hold: by March the ratio had climbed back to 32.26%, Tk5.89 lakh crore, up Tk31,487 crore in a single quarter, as recovery stayed weak and rescheduled loans defaulted again. 

The next wave is already forming: special mention accounts, the early-distress category just before default, rose by nearly Tk29,000 crore that quarter, to Tk1.32 lakh crore. The number fell, then rose; the problem never moved.

That distinction sits at the centre of the reform now being finalised. The Financial Institutions Division has drafted a Distressed Asset Management Act, due around June, to license private firms to buy and trade bad loans, without a state-funded bad bank. It belongs to the same package meant to bring bad loans below eight percent by June 2026, a target missed by a factor of four, and now moving the wrong way. After two years of stalled reform it is a genuine step, rightly aimed: bad loans must leave bank balance sheets for specialists to work out.

But a law that permits trading is not a market that produces trades. I spent much of my career on the other side of this problem, including recapitalising the Greek banks after their crisis. Distressed-debt markets do not stall for lack of legislation. They stall on price, and price is exactly what the law cannot legislate.

The mechanism is brutal in practice. A bank carries a bad loan far above what it is worth, because writing it down crystallises a loss it cannot afford; system capital is already negative, and the gap between provisions held and provisions needed runs past Tk2 lakh crore. 

A distressed-debt fund buys only at a price reflecting what it can recover. The asking price is anchored to a fiction, the bid to reality. When the gap is wide enough, nothing changes hands. The law can license a hundred asset management companies, and not one loan will trade if the seller cannot afford the only price a buyer will pay.

Bangladesh has a particular handicap here. A buyer values a distressed loan by discounting expected recoveries, and that rate needs a reference point for the country's credit risk. Bangladesh has never issued an international sovereign bond. For years, the government has declined to establish an international yield curve, treating its absence as a badge of honour. With no traded benchmark for sovereign risk, buyers price recovers off guesswork and a wide risk margin, lowering every bid.

The recovery side compounds it. The price a buyer pays is the cash it expects to extract, and in Bangladesh, that runs through the Money Loan Courts, where banks recover around a quarter of what they are owed, slowly, with much stuck in pending suits. A buyer expecting twenty-five paisa in the taka, late, bids accordingly.

Greece is a useful case, not because it resembles Bangladesh but because it thawed the same frozen market without a bad bank. Rather than buy loans, the state guaranteed only the senior, safest slice of the portfolios banks sold. The guarantee was priced at a market rate, off Greece's own sovereign spread, so it counted as no subsidy, and the deals paid for it. Taking the risk out of the senior tranche narrowed the gap enough to transact; investors took the junior pieces at their own price, and the market cleared. 

Greek bad loans fell from roughly half of all lending to under four percent in a few years, through sales, not re-labelling. Bangladesh cannot copy it wholesale, and such a guarantee is harder where no benchmark trades. But the principle holds: guaranteeing the senior tranche makes the price meet without the state overpaying outright, the trap that sank Spain's bad bank, which paid too dear and left taxpayers the bill.

Governance is the harder constraint, and others have made that case better than I can: until wilful defaulters face consequences, no structure fully works. But governance and pricing are not rival explanations. A market clears on price only once book values are honest, and that is where the design around the law decides whether it bites.

The law will only bite if the design around it does three things. Bangladesh Bank should require the asset quality reviews underway to publish standardised valuations, so buyers and sellers argue from the same data, and should tighten provisioning until book values fall toward real worth. The Finance Ministry should pair the framework with a narrow, market-priced guarantee on senior tranches, modelled on Athens rather than Madrid. And the courts that enforce collateral need to recover far more than a quarter, and far faster.

None of this is a reason to slow the law down, only to be clear about what it does. Passing it makes trading Bangladesh's bad loans legal. Making them trade is a separate task, and it turns on the question the law leaves unanswered: at what price.

Sketch: TBS

Fahim Chowdhury is an investment banker who has raised over $200bn and executed 500+ capital markets transactions in more than 30 markets. He is currently Managing Director at RetailBook and was previously at Citi. [email protected]

Disclaimer: The views and opinions expressed in this article are those of the author and do not necessarily reflect the opinions and views of The Business Standard.