Case 081FinancialsHard

An NBFC with Rs 20,000 crore of loans sees gross NPAs rise from 2% to 4.5% and keeps provision cover at 55%. What does that do to credit cost and ROE this year?

The situation

Nivrita Finance lends to small businesses and has assets under management of Rs 20,000 crore, funded with Rs 4,000 crore of equity. Its net interest income is 7% of loans, operating costs are 2.5%, and in a normal year its credit cost is 0.8%, Rs 160 crore, which replaces loans written off and keeps gross NPAs steady at 2%. Tax is 25%.

This year gross NPAs rise to 4.5%. Management says it will hold provision cover at 55%. Loans that turn bad stop paying their 14% yield; assume they slip evenly through the year, so on average they earn nothing for half of it.

Your task

What is this year's credit cost, what happens to ROE, and which line should an analyst watch most closely?

Quick check

Roughly where does Nivrita's ROE land this year, from 13.9% last year?

Worked solution

Try it on paper, then open one step at a time.

30-second answerThe answer to give first

Credit cost rises from 0.8% to about 2.17%, and ROE falls from 13.9% to about 8.1%. Gross NPAs rise by Rs 500 crore. Holding 55% cover needs Rs 275 crore of provisions on top of the normal Rs 160 crore, and the bad loans stop paying interest. After tax that is 5.8 points of ROE. The line to watch is cover: dropping it to 45% would flatter ROE by about 1.7 points.

Step 1How does a higher NPA ratio become a profit number?

Through provisions. A shopkeeper who sells on credit and hears that a regular customer has lost his job does not wait to be sure: he sets money aside in his head for the part he will not get back. A rise in bad loans hits profit through the provisions it forces, not through the headline NPA ratio. Gross NPAs go from Rs 400 crore to Rs 900 crore, Rs 500 crore of new bad loans. Keeping 55% provision coverProvisions set aside divided by gross NPAs: the share of bad loans already written down in the accounts. means provisions must go from Rs 220 crore to Rs 495 crore, a charge of Rs 275 crore on top of the normal Rs 160 crore.

From new bad loans to lost ROE, one step at a timeNew bad loansGNPA 400 to 900Rs 500 crProvisions neededat 55% coverRs 275 crProfit lostafter 25% taxRs 206 crROE loston Rs 4,000 cr equity-5.2 ptsx 55%x (1 - 25%)/ 4,000The NPA ratio moved 2.5 points; ROE moved 5.8 points. The P&L feels provisions, not the ratio.
Rs 500 crore of new bad loans at 55% cover forces Rs 275 crore of provisions, which after tax removes Rs 206 crore of profit and 5.2 points of ROE from Nivrita's Rs 4,000 crore equity base.
Step 2What else does a bad loan take away?

Its interest. A loan that stops paying is taken off the income line. At a 14% yield on Rs 500 crore, for half a year on average, Nivrita loses Rs 35 crore of interest income, 0.7 points of ROE after tax. Lost interest is real but small; the provision is five times larger and is the number to model first. Credit cost, provisions over loans, goes from 0.8% to 2.17%.

Rs croreLast yearThis year
Net interest income1,4001,365
Operating costs(500)(500)
Provisions: normal(160)(160)
Provisions: new NPAs at 55% cover(275)
Tax at 25%(185)(107.5)
Profit after tax555322.5
Credit cost, % of loans0.80%2.17%
ROE on Rs 4,000 crore13.9%8.1%
Nivrita's profit falls from Rs 555 crore to Rs 322.5 crore and ROE from 13.9% to 8.1%, almost entirely because Rs 275 crore of extra provisions lifts credit cost to 2.17%.
Nivrita's ROE, per cent: provisions do almost all the damage13.9Last year-0.7Interest loston new NPAs-5.2Extraprovisions8.1This year
Nivrita's ROE falls from 13.9% to 8.1%: lost interest on the new bad loans takes 0.7 points and the extra provisions take 5.2 points.
Step 3Which number should you watch most closely?

Provision cover, because management chooses it. If Nivrita let cover slip to 45%, it would need only Rs 185 crore of extra provisions, and ROE would print 9.8% instead of 8.1%. A lender that lets cover fall while NPAs rise is borrowing profit from next year. Check net NPAs as well: after provisions they rise from Rs 180 crore to Rs 405 crore, from 4.5% to 10.1% of equity. That is the unprovided loss still sitting on the balance sheet if recoveries disappoint.

State the limit of the calculation. It assumes the new NPAs all arrive this year and none recover. In practice some accounts cure, collateral on small business loans recovers part of the loss, and the pace of slippage decides how much falls in this year's accounts. The questions for management are which segments slipped, what the collection efficiency is, and whether the cover of 55% is held segment by segment.

Where candidates lose it

Candidates stop at the ratio: NPAs went from 2% to 4.5%, still a small number, so profits are fine. They never convert the new bad loans into provisions, and so never notice that a 2.5 point move in NPAs removes almost six points of ROE on a lender levered five times.

The second miss is ignoring cover. An analyst who models reported provisions without checking cover will praise the lender that under-provides.

What the interviewer asks next

  • If half of the new NPAs are secured by property worth 70% of the loan, how would you set cover?
  • What does Nivrita's book value per share do this year, and why does it matter for a price-to-book multiple?
  • How would the ROE hit change if equity were Rs 3,000 crore instead of Rs 4,000 crore?

Company names and figures are illustrative.