Valuation Ratios in the NEPSE Context

First published 23 Aug 2026 · Last verified 29 Aug 2026

Lesson 39.1 — The P/E Ratio: How NEPSE Actually Reports It

Open any counter's page on Sharesansar or Merolagani and the first number analysts quote is the P/E ratio — price divided by earnings per share. The formula never changes, but the way Nepal's market computes and displays the two halves of that fraction has enough local texture that a mechanical import of a foreign textbook definition will mislead you.

The price half is straightforward: the last traded price (LTP) on NEPSE at market close. The earnings half is where the local nuance sits. Nepali brokerage platforms typically quote P/E against the most recently reported annual EPS from audited financials, but during the fiscal year — between quarterly unaudited reports — many platforms switch to an "annualized EPS," taking the trailing four quarters or, for younger companies, multiplying the latest quarter's EPS by four. This second method is convenient but dangerous for any business with seasonal earnings, a point we return to in Lesson 39.2 when hydropower enters the discussion.

A worked example grounds this. Consider the second-quarter (Poush-end) results of fiscal year 2082/83 — the most recent full reporting cycle available at the time of writing — across a cross-section of Nepal's commercial banks.

("Implied price" is simply P/E × EPS, back-solved from the same source table, and is shown here to make the arithmetic transparent — treat it as an illustration of method, not a live quote.)

Two things should jump out. First, within a single sector — commercial banking, the most homogeneous, most heavily regulated corner of NEPSE — P/E multiples still range from under 8x to over 300x. Second, the outlier is not a growth story; it is an accounting artefact. NIC Asia's EPS of Rs 1.76 was compressed by a large capital base relative to a temporarily subdued profit quarter, and dividing a perfectly ordinary share price by a near-zero EPS manufactures a P/E of 343x that tells you nothing about value and everything about the denominator problem. A bank at Rs 604 is not "343 times more expensive" than Nepal Bank at Rs 136 trading on 7.67x — it is a warning that the ratio has broken down.

A single-digit P/E does not automatically mean "cheap," and a triple-digit P/E does not automatically mean "expensive" — it means look at the denominator first.

Lesson 39.2 — Why Sector P/E Differs Structurally

It would be a mistake to treat NEPSE as one market with one "fair" P/E. NEPSE is really eleven-odd sub-markets bolted together under a single index, and each sub-market prices earnings differently for structural reasons that have nothing to do with mispricing.

The clearest illustration comes from a full sector-by-sector P/E survey Sharesansar ran across all listed NEPSE companies. Though the underlying data point is from an earlier period, the ranking of sectors it revealed — which sectors trade rich and which trade cheap, and why — has held up remarkably consistently in the years since, including in more recent data.

Absolute levels drift year to year — commercial banking, for instance, was averaging closer to 15.8x with a P/B near 1.51x in a late-2025 sector survey, still the "most rationally valued" segment of the market even as NEPSE's headline P/E pushed toward record territory. But the ordering — insurance and microfinance rich, hydropower volatile, banking and manufacturing comparatively grounded — has been remarkably stable across cycles. That ordering is the useful lesson, more than any single year's absolute number.

Why do banks anchor the cheap end? Nepal Rastra Bank's capital adequacy framework, single-obligor limits, and CD-ratio ceiling effectively cap how fast a bank's balance sheet — and therefore its earnings — can grow in any given year. Investors know this, so they refuse to pay insurance-sector multiples for bank-sector growth. Why does hydropower sprawl from 4x to over 100x? Because a run-of-river plant's output swings with monsoon rainfall, and a newly commissioned plant reports a tiny, almost accidental profit in its first full year before ramping toward normalised capacity utilisation — dividing a normal share price by that accidental first-year profit produces an enormous, meaningless P/E.

Subhead: A live example of the hydropower trap

Upper Tamakoshi Hydropower, one of the country's flagship run-of-river plants, has traded at a P/E near 114x in recent periods — not because the market expects the company to grow earnings 114-fold, but because a specific reporting window caught unusually thin recent earnings relative to its installed capacity and asset base. Compare that to Chilime Hydropower, which has built a dividend track record stretching back to fiscal year 2060/61, with an average payout around 29% and a peak near 70% — a company old enough that its earnings have normalised into a steadier, more interpretable pattern. Same sector, wildly different reliability of the P/E signal, purely as a function of the company's age and monsoon-cycle position.

Lesson 39.3 — P/B Ratio and Why It Matters Most for Banks

Price-to-book compares the market price to the company's book value per share — shareholders' equity divided by shares outstanding. For most industrial or trading companies, book value is a poor proxy for what the business is actually worth, because plant, inventory, and goodwill are recorded at historical cost and say little about earning power. Banks are the exception, and this is worth understanding precisely rather than by rule of thumb.

A bank's balance sheet is, in an accounting sense, close to what it appears to be: loans, investments, and deposits are financial instruments carried at values Nepal Rastra Bank's prudential and disclosure regime forces the bank to mark reasonably close to reality, through loan-loss provisioning rules, capital adequacy reporting, and NRB-mandated disclosure formats. That regulatory transparency is precisely why P/B carries more information for a bank than it does for a hydropower company or a trading house — you can trust the denominator.

A recent cross-section of Nepali commercial banks makes the point:

The pattern is not random. A bank with a non-performing loan ratio under 1% (EBL) commands a materially higher price-to-book than one carrying an NPL ratio nearly seven times as high (KBL), because the market is, correctly, pricing in the probability that KBL's book value overstates the true collectible value of its loan portfolio. This is the single most useful application of P/B in the NEPSE context: use it to cross-check whether a bank's reported book value should be trusted at face value, by reading it alongside the NPL ratio, capital adequacy ratio, and provisioning coverage disclosed in the same quarterly report.

Subhead: The mechanics of book value per share

Suppose a commercial bank reports total shareholders' equity — paid-up capital plus reserves and retained earnings — of Rs 18 billion, against 120 million shares outstanding. Book value per share is Rs 18,000,000,000 ÷ 120,000,000 = Rs 150. If the share trades at Rs 270, the P/B is 1.8x. Compare that instantly to the bank's return on equity: a bank earning an ROE of 16-18% arguably deserves to trade above book, because it is compounding shareholder capital faster than a bank earning 8-10% ROE, all else equal. P/B divorced from ROE is an incomplete comparison — the two ratios are meant to be read together, not separately.

Lesson 39.4 — Dividend Yield, Nepal-Style: The Face-Value Trap

Dividend yield, everywhere else in the world, means one thing: annual dividend per share divided by current market price. In Nepal's retail investment culture, however, the headline number that circulates — in AGM notices, in newspaper tables, in WhatsApp groups — is something else entirely: the dividend percentage declared on face value (par value), which for the overwhelming majority of NEPSE-listed companies is Rs 100 per share.

This distinction is not a technicality; it is the single most common source of confusion for a new NEPSE investor, and getting it wrong leads directly to overpaying for "high-yield" stocks that are nothing of the sort.

A recent ranking of NEPSE's highest dividend-declaring companies illustrates the headline numbers investors actually see:

Read literally, Unilever Nepal's "1,842%" looks absurd — and it is, until you remember it is 1,842% of Rs 100 face value, i.e., Rs 1,842 per share in cash, on a stock that trades in the tens of thousands of rupees because of its extremely small share count and near-total absence of bonus dilution over decades. The percentage on face value and the actual yield on market price can differ by an order of magnitude or more.

Subhead: Converting the headline into a real yield

Take Nabil Bank. It declared a 12.5% cash dividend, meaning Rs 12.50 per share on face value. From the same reporting period used in Lesson 39.1, Nabil's price was approximately Rs 546. The actual dividend yield an investor buying at that price would have earned is:

Rs 12.50 ÷ Rs 546 = 2.29%

Now take Standard Chartered Nepal, which declared 19% cash — Rs 19 per share — against an implied price of roughly Rs 628:

Rs 19 ÷ Rs 628 = 3.02%

Both look unremarkable once converted — squarely in the 2-3% range that characterises yield on Nepal's larger, more mature bank counters, broadly consistent with NEPSE's market-wide dividend yield, which has hovered around just 1-2% during periods when the index itself has been expensive. That is a world away from the "12.5%" or "19%" a reader skims off a newspaper table.

The distortion cuts both ways. A thinly-traded, high-priced counter can declare a modest-looking face-value percentage and still deliver a respectable true yield, while a low-priced, heavily bonus-diluted counter can declare a large-looking face-value percentage and deliver a true yield under 2%. Bonus components compound the confusion further: when ADBL declares "13% (3.25% bonus + 9.75% cash)," the cash component alone is the income return; the bonus component is not income at all — it is additional shares, which dilutes future EPS and, mechanically, tends to pull the share price down roughly in proportion on the ex-dividend date. Treating a bonus percentage as if it were cash yield is a second, related error worth guarding against.

Lesson 39.5 — EPS Mechanics: Weighted Average Shares and the Bonus Share Adjustment

Every ratio covered so far in this chapter has EPS sitting in its denominator or, in P/B's case, a close cousin of it — shares outstanding. Get the share count wrong, and every ratio built on top of it is wrong. Nepal's heavy and recurring use of bonus shares makes this the single most important mechanical detail in the entire chapter.

Under the Nepal Financial Reporting Standards that NEPSE-listed companies follow (mirroring international EPS accounting), earnings per share must be computed on a weighted average number of shares outstanding during the period — not simply the share count at the balance sheet date. This matters because share capital does not sit still during a Nepali fiscal year: rights issues, bonus issues, and mergers all change the denominator mid-year.

The two most common events — rights shares and bonus shares — are treated completely differently, and conflating them is the most frequent analytical error retail investors make.

Subhead: Rights shares are time-weighted; bonus shares are retroactive

A rights issue brings new cash into the company in exchange for new shares, at a specific date. Because real resources entered the business only from that date forward, the new shares are time-weighted into the denominator — a rights issue completed with three months left in the fiscal year adds only a quarter's worth of dilution to that year's weighted average share count.

A bonus issue, by contrast, capitalises existing reserves into new shares — no new cash comes in at all; shareholders simply receive more paper representing the same underlying company. Because nothing real changed on the date of issue, accounting standards require that bonus shares be treated as if they had always been outstanding — not just for the current year, but retroactively, restating the prior year's comparative EPS on the same enlarged share base, so that year-on-year EPS growth comparisons remain meaningful.

A worked numeric example: A company reports profit of Rs 400 million for the year and has 10,000,000 shares outstanding at the start of the year. Partway through the year it issues a 10% bonus (1,000,000 new shares). Two ways of computing EPS are possible, and only one is correct:

Incorrect (year-end share count only): Rs 400,000,000 ÷ 11,000,000 shares — but only applying this to the current year while leaving last year's comparative EPS on the old 10,000,000 share base overstates apparent earnings growth.

Correct (NFRS-compliant): Both this year's and last year's EPS are calculated on the post-bonus 11,000,000 shares. If last year's profit was Rs 350 million, restated prior-year EPS becomes Rs 350,000,000 ÷ 11,000,000 = Rs 31.82, and current-year EPS becomes Rs 400,000,000 ÷ 11,000,000 = Rs 36.36 — a genuine, comparable growth rate of roughly 14.3%, rather than a growth rate artificially inflated by comparing an unadjusted small prior-year denominator to a larger current one.

The NIC Asia case from Lesson 39.1 — a P/E of 343x built on an EPS of just Rs 1.76 — is worth revisiting through this lens. Whenever you encounter an EPS that looks anomalously low relative to a company's history and peer set, the first three questions to ask are: was there a recent bonus or rights issue that enlarged the share base faster than profit grew; was there a merger or acquisition that reset the share count; and is the reporting period annualized correctly. Only once those three questions are answered should you conclude the low EPS reflects a genuine deterioration in the underlying business.

Lesson 39.6 — NEPSE's Market-Wide P/E as a Sentiment Barometer

Every individual counter's P/E tells you something about that company. NEPSE's aggregate, market-wide P/E — the weighted average across all listed companies, published by NEPSE itself and tracked continuously by Sharesansar — tells you something different: where collective investor sentiment sits in the market's own historical range, and how Nepal compares to its regional peers.

The number has swung dramatically across cycles. During the speculative peak of 2021, NEPSE's market P/E reached roughly 42.28x — a level that, in hindsight, marked an unsustainable valuation extreme, followed by a multi-year correction that pulled the ratio down substantially before a subsequent recovery. As of a recent reading in early 2026, the market P/E had climbed back to approximately 38x, with the market's aggregate price-to-book near 2.78x, dividend yield compressed to just 1-2%, and market capitalisation equal to roughly 72.6% of GDP — all readings clustered near the upper end of NEPSE's own historical range.

What makes this figure genuinely useful as a sentiment gauge is not its absolute level in isolation, but how it compares to other frontier and emerging markets facing broadly similar macro constraints:

The pattern is striking and worth sitting with. NEPSE, at roughly 38x, trades at nearly double the multiple of the MSCI Frontier Markets basket it structurally belongs to, and at two to five times the multiples of comparable South Asian peers — Bangladesh, Sri Lanka, and Pakistan — that share Nepal's broad emerging-market growth profile. Meanwhile NEPSE's dividend yield, at 1-2%, sits well below every one of those peer markets, several of which pay 3-7%. A market can be simultaneously expensive on earnings and stingy on income, and that combination is exactly what elevated valuation with compressed yield signals: a market being carried more by capital appreciation expectations and liquidity than by the cash generation investors are actually being paid today.

Why does NEPSE persistently command such a premium multiple relative to peers with arguably better growth and profitability fundamentals? Structural, not fundamental, reasons dominate the explanation: a chronically small free float relative to demand from a large domestic retail base with few alternative investment channels (real estate transaction friction, capital controls limiting outbound investment, low fixed-deposit rates during liquidity-surplus periods), combined with concentrated retail participation that trades on price momentum and rumour as much as on earnings. None of this means the multiple is "wrong" in a way that forces immediate correction — Nepal's market has sustained rich multiples for extended stretches before — but it does mean a NEPSE-wide P/E near 38x, at nearly the same level that preceded the 2021 correction, is information a disciplined investor should not ignore, even while continuing to buy selectively into individual counters that screen cheap on the sector-relative and company-specific metrics covered earlier in this chapter.

The practical takeaway for a retail investor is not to sell everything the moment the market-wide P/E crosses some magic threshold — no such precise trigger exists — but to treat a rising aggregate multiple as a reason to demand a higher margin of safety on new purchases, to lean more heavily on the sector-relative and company-specific tools from Lessons 39.1 through 39.5 rather than on market momentum, and to remember that the last time NEPSE's P/E sat this high, it was followed by years of underperformance for anyone who bought at the peak on faith in continued re-rating rather than on earnings.