Why your bank shares stopped paying dividends: the Rs 70 arba profit that became a 2% payout

A friend who bought his first bank shares in 2017 sent me a screenshot last week. Kumari Bank, board meeting of 30 Bhadra: proposed dividend 2.1053%, all cash. Under it, a one-line message. "I remember when banks gave 30, 40 percent. What happened?"

Most of the twenty commercial banks he could have bought back then paid between 30% and 48% a year. This season the best announcement so far is 20%, the median is around 10%, and four banks are arithmetically unable to pay anything. The odd part is that the banks made more money than ever. Twenty commercial banks earned Rs 69.78 arba in FY 2082/83, up 32% on the year. The profit is real. It just is not the number a dividend is paid from.

The number that decides your dividend is not net profit

Every dividend headline is a percentage of paid-up capital, and every dividend rupee has to be traced back to a line NRB calls "net profit available for distribution". That line is what remains after a bank's net profit has been run through two sets of deductions: the statutory appropriations required by the Bank and Financial Institution Act, and the regulatory adjustments required by NRB's Unified Directive 4. A bank can report a record profit and still arrive at zero, or below zero, on that line.

The fourth-quarter numbers for FY 2082/83 make the gap visible. Of twenty commercial banks, sixteen had positive distributable profit and four had negative balances, which means accumulated deductions exceed everything they have earned and retained.

Sources: ShareSansar's Q4 aggregate and Ratopati's per-bank breakdown. Unaudited; the audited figures NRB approves can differ.

Kumari is the case to hold in your head. Net profit up 306% to Rs 7.39 arba, earnings per share of Rs 28.17, and yet distributable profit of Rs 791.66 million, which is 3.02% of its Rs 26.22 arba paid-up capital. The board proposed 2.1053%, leaving a little headroom. Everything between Rs 28 of earnings and Rs 3 of distributable profit went into reserves the bank is not allowed to hand out.

Where the profit goes before it reaches you

The first layer of deductions is old and uncontroversial. The Bank and Financial Institution Act (BAFIA) 2073 and NRB circulars require every bank to set aside fixed slices of profit before considering shareholders. None of this is new, and none of it explains the collapse in payouts on its own, but it sets the floor.

Section 47 of the same Act adds the gate: no bank may declare a dividend without NRB's prior approval, and none may pay until previous years' losses are recovered, capital requirements are met and loan-loss provisions are made as directed. That last clause is where the second, larger layer begins.

The regulatory reserve: profit you earned but cannot touch

When Nepali banks adopted NFRS accounting, their reported profits started including items that were not cash and might never become cash. NRB's answer was the regulatory reserve. Its own definition, in the format of financial statements every commercial bank must follow, calls it the amount set aside from profit "which shall not be regarded as free for distribution of dividend (cash as well as bonus shares)".

The Unified Directive 2082 prescribes the exact statement. After the appropriations above, the bank must deduct:

• Interest income recognised on an accrual basis but not yet received in cash. A borrower who has not paid for eight months still generates "income" on the books; NRB refuses to let that income be paid out.
• Any gap between the loan-loss provision NRB's directive demands and the lower impairment NFRS may allow.
• Deferred tax assets recognised in the year.
• Goodwill and bargain-purchase gains booked in a merger or acquisition.
• Actuarial losses on staff gratuity and leave liabilities.
• Short provisions on investments and on non-banking assets (property the bank has seized from defaulters).

Add opening retained earnings, subtract dividends already paid, and what is left is the total distributable profit. The Procedure for Approval of Financial Statement Publication and Dividend Distribution, 2082, issued in Bhadra 2082 to replace the 2077 version, tightened the plumbing: draft accounts go to NRB's supervision department before the board signs them, interest capitalised during grace periods also goes to the reserve, and the bank must still clear the 11% total capital ratio after paying. Reserves created by a merger's capital shortfall cannot fund a cash dividend at all.

Two things follow for a shareholder. First, the reserve is not lost. When the interest is eventually collected or the provision reverses, the amount moves back to retained earnings and becomes distributable in a later year. Second, the size of the reserve is a direct readout of how much of a bank's profit is paper. A bank whose regulatory reserve is growing faster than its retained earnings is telling you, in a footnote, that its borrowers are not paying.

The caps NRB tried, and dropped

The regulatory reserve is not the first tool NRB has used to hold dividends down; it is the one that stuck. In October 2020 an amendment to the dividend working procedure barred any cash dividend, beyond the slice needed for tax, from a bank whose distributable profit was below 5% of paid-up capital. The July 2021 Unified Directive went further and capped cash dividends at the lower of 30% of distributable income or the bank's average deposit rate, a rule that tied what shareholders got to what depositors got. NRB later removed the ceiling, the 5% test and the deposit-rate link.

Those caps were blunt, and the banks lobbied them away. What replaced them is quieter and harder to argue with. A cap says "you may not pay more than X". The distributable-profit statement says "you have not earned it yet", and for a bank with a growing pile of uncollected interest that is simply true. The result is that the ceiling moved from NRB's rulebook into each bank's own loan book, which is why the payouts now vary so widely between banks that all report growing profits.

The bad-loan connection

The regulatory reserve did not grow on its own. It grew because bad loans did.

NRB's July 2026 Macroeconomic Report puts the commercial-bank non-performing loan (NPL) ratio at 1.81% in the third quarter of 2016 and 5.41% in the same quarter of 2026. The monthly statistics for Asar-end 2083 show A-class banks at 5.56%, with development banks at 5.77% and finance companies at 9.89%. Watchlist loans, the ones a step short of default, went from 6.7% of the book in mid-2023 to 11.1% by mid-April 2026. The same NRB report spells out the mechanism in one sentence: NPLs "reduce profits, weaken capital through higher provisions, raise risk-weighted assets, and pressure capital adequacy ratios."

Per-bank figures from the banks' fourth-quarter reports as compiled by NEPSE Trading and ictframe; Everest's is from its first-quarter filing and may have moved.

A bad loan hits distributable profit twice. The provision the bank must book against it is deducted directly. And the interest it was accruing on that loan, which had been counted as income, now has to be moved to the reserve because no cash arrived. Prabhu's NPL more than doubled from 7.01% to 15.55% in a single year, and the bank swung to a Rs 241 million net loss; its distributable profit is negative by more than Rs 6 arba. Himalayan is the slower version of the same story. It acquired Civil Bank in 2023, a bank carrying roughly 27% NPL, and three years later its Rs 1.48 arba profit sits on top of a Rs 9.84 arba hole.

If you want to know whether a bank will pay next year, its NPL trend is a better predictor than its profit growth. How to read that and the other numbers in a quarterly report is covered in EPS, P/E and book value on NEPSE.

Mergers made the denominator bigger

The second structural change is the denominator. A dividend percentage is a share of paid-up capital, and paid-up capital in Nepali banking has grown far faster than the economy underneath it.

NRB ordered banks to quadruple paid-up capital from Rs 2 arba to Rs 8 arba in FY 2015/16, and most met it with bonus shares and rights issues rather than fresh business. Then came the merger wave of FY 2079/80, when twelve commercial banks became six: Kumari with NCC, Global IME with Bank of Kathmandu, Prabhu with Century, Nepal Investment with Mega, Himalayan with Civil. Global IME now carries Rs 38.11 arba of paid-up capital. Paying it a 10% dividend costs Rs 3.8 arba; paying the pre-merger Global IME the same 10% would have cost less than half that.

Three of the four banks with negative distributable profit this year are products of that merger round. Merger accounting created goodwill and bargain-purchase gains, both of which Directive 4 strips out of distributable profit, and the acquired loan books brought the NPLs that fill the reserve. The Everest CEO put the bonus-share side of the problem plainly to the Annapurna Express: bonus share distributions "automatically reduce dividend capacity for subsequent years." Every 10% bonus you received in 2075 raised the capital base that every later dividend is measured against. The trade-off between bonus and cash, and why the market prices them differently, is in bonus share vs cash dividend.

The long-run record shows the two forces compounding. The average commercial-bank dividend was 22.25% in FY 2012/13, 13.51% in FY 2021/22 and 11.51% in FY 2022/23. Nabil, the steadiest payer in the system, went from 48% in FY 2073/74 to 11% in FY 2079/80, then 10%, 12.5% and now 15.8%. Its profit over the same period roughly doubled.

This season's scoreboard

Eight of the twenty commercial banks had announced by 18 September 2026. All are proposals, subject to NRB approval and the AGM.

Look at the shape of the cash column. Citizens' 0.20% cash and Everest's note that its 10% cash "includes the tax amount required for the bonus shares" are the same device: the cash slice exists to pay the 5% tax due on the bonus, so shareholders are not left owing tax on paper. How that tax works is its own post.

Twelve banks were still silent at the time of writing, and silence in September has a pattern. Last year only eleven banks had declared by early December; Kumari, Nepal Bank, Nepal Investment Mega, Himalayan, NIC Asia and Prabhu paid nothing for FY 2081/82, and the formal "no dividend" notices came in Mangsir and Poush. Expect the same rhythm.

The Shrawan 15 rule

One thing did change in the banks' favour this year. On 15 July 2026 NRB issued Circular 19/082/83, amending Directive 4 so that accrued interest a bank actually collects within 15 days after the fiscal year ends, by Shrawan 15, no longer has to be parked in the regulatory reserve. Only the interest still uncollected after that window goes to the reserve, and it comes back to retained earnings once received.

Bankers quoted by Peoples' Review estimated the window would release around Rs 10 billion of distributable profit across the twenty banks, with Global IME alone expecting Rs 400 to 450 million. Treat that as a banker's estimate; the circular itself carries no rupee figure. It is a timing fix for the borrower who pays on Shrawan 5 instead of Asar 30, and a sensible one. It does not touch a loan that has not paid since last Dashain, which is the loan filling Himalayan's and Prabhu's reserves.

What a 10% dividend is actually worth to you

The percentage is quoted on the Rs 100 par value, and nobody buys a bank share at par. Global IME traded around Rs 240 when it announced 10%. That is a gross yield of about 4.2%; after the 5% final withholding under Section 88 of the Income Tax Act, roughly 4.0% lands in your account. (Calculation: 10 ÷ 240 × 0.95.) An individual fixed deposit at a commercial bank pays around 4.3% before the 6% TDS, so about 4.0% net. On yield alone, the two are a coin toss, and only one of them can fall 20% in a bad quarter.

Bank shares traded at a P/E near 16 against a NEPSE average near 38 in late 2025, which is the argument for them; whether the market as a whole is expensive is a different question from whether a given bank can pay.

The timing matters too. The Companies Act 2063 requires the AGM within six months of the fiscal year end, which puts the outer deadline around mid-January. Book closure falls just before the AGM, and who gets the dividend depends on holding the share on the last trading day before it. Everest's AGM is on 30 September, Sanima's on 28 September; the last day to buy for either was 16 September. A September announcement from a bank that has not yet set an AGM date is an intention, not a payment.

Reading the next announcement

Three questions turn a dividend headline into a decision.

1. What is the distributable profit per share, not the EPS? It is in the Q4 report, in the statement NRB prescribes. If distributable profit is negative, no dividend is possible regardless of the profit line; if it is a fraction of EPS, the reserve is absorbing the difference.
2. Which way is the NPL moving? A rising NPL in a year of rising profit means the profit is partly accrued interest that will be reclassified. Prabhu's 7% to 15.5% in a year was the loudest possible warning.
3. How much of the payout is bonus? Bonus shares enlarge the capital base that next year's percentage is measured on. A 15% dividend that is 10% bonus commits the bank to earning more just to stand still, as Everest's CEO said outright.

A bank paying 2% this year is not necessarily a bad bank; it may be a bank that provisioned honestly. A bank paying 20% is not necessarily safe; check that the number is cash from collected interest, not a reserve reversal. And a portfolio built for dividend income needs to be rebuilt around the banks whose reserve is shrinking, not the ones whose headline was highest in 2075.

What you actually need to know

1. Net profit is not what you are paid from. Statutory reserves and NRB's Directive 4 regulatory reserve stand between the Rs 69.78 arba the banks earned and the far smaller amount they may distribute. Kumari turned Rs 28 of EPS into Rs 3 of distributable profit and a 2.1% dividend; four banks ended the year below zero.
2. Bad loans and merger capital did the damage. Commercial-bank NPL went from 1.81% to 5.41% in a decade, and the 2023 mergers inflated paid-up capital while importing troubled loan books. The healthy banks still pay 15 to 20%; the merged and NPL-heavy ones cannot pay at all.
3. Judge the yield, the reserve and the timeline, not the percentage. A 10% dividend at Rs 240 is about 4% after tax, roughly an FD. Read the distributable-profit line and the NPL trend before the headline, and expect NRB approval and an AGM before any cash, which last year meant December or later.

Holding a bank that has gone quiet this season and wondering whether to wait it out? Email parjanya57@gmail.com with the ticker and I'll walk through its Q4 statement with you.

This post is part of the Nepal Money Basics guide — the investing section.