Profitability Ratios — Sector by Sector

First published 23 Aug 2026 · Last verified 29 Aug 2026

Lesson 40.1 — ROE and ROA: The Universal Starting Point (and Why It Misleads Fast)

Every analyst who opens a NEPSE annual report ends up circling the same two lines: net profit and shareholders' equity. Divide one by the other and you have Return on Equity — the single number that Nepali investors quote most often to decide whether a company is "good." Divide net profit by total assets instead, and you get Return on Assets, the quieter cousin that tells you how efficiently a company turns its balance sheet into profit, independent of how that balance sheet was financed.

ROE = Net Profit / Average Shareholders' Equity ROA = Net Profit / Average Total Assets

Both ratios are expressed as annualized percentages, and both are backward-looking — they describe what already happened, not what will happen. That is fine for judging management's stewardship of capital already entrusted to them; it is dangerous when used, unadjusted, to rank a commercial bank against a hydropower company against an insurer against a noodle manufacturer, which is exactly what happens every day in NEPSE Facebook groups and brokerage WhatsApp threads.

The reason cross-sector ROE comparison misleads is structural, not incidental. A commercial bank in Nepal typically carries assets nine to eleven times the size of its equity — deposits fund most of the balance sheet, and a thin equity sliver absorbs both the risk and, when things go well, an outsized share of the return. A hydropower company under construction may be sitting on IPO proceeds parked in fixed deposits, generating no operating revenue at all, so its ROE looks near zero — not because the business is bad, but because the business has not started yet. A life insurance company's profit is driven as much by where its investment department parks policyholder reserves as by how well its actuaries priced the policies. A manufacturing company with a decades-old, barely-diluted paid-up capital base can show a triple-digit ROE off a genuinely modest operating margin, simply because the equity denominator is so small.

None of these numbers are wrong. They are all correctly calculated. What is wrong is reading them as if they measured the same thing.

Why ROA is the more honest cross-sector yardstick

Because ROA divides profit by total assets rather than by equity, it strips out the leverage effect and gets closer to a pure measure of operating efficiency — how much profit a company extracts from every rupee it deploys, regardless of whether that rupee came from shareholders or depositors or bondholders. This is precisely why bank ROA figures look deceptively small (often under 2%) next to bank ROE figures that can reach the mid-teens: banks are leverage machines by design, regulated to run on thin equity cushions, and ROA reflects that reality honestly while ROE amplifies it.

The practical implication for a Nepali retail investor is this: never let ROE alone decide a buy or sell. Pair it with ROA, and — as this chapter will show in Lesson 40.2 — decompose it further with DuPont analysis before drawing conclusions about quality of earnings. A hydropower stock at 4% ROE two years after commissioning may be a better long-term hold than a finance company at 14% ROE built on aggressive leverage into a rate-sensitive book. The number alone cannot tell you that. The sector context, and the components underneath the number, can.

This chapter walks through profitability the way a working analyst actually has to: sector by sector, with the specific mechanics that make each sector's ROE and ROA behave differently, and then a method for putting them back on a common footing when you genuinely need to compare across sectors — say, when deciding whether your next lakh goes into a bank or a hydropower IPO.

Lesson 40.2 — Banks: Leverage, Net Interest Margin, and the Regulatory Ceiling on Returns

Banking is the sector where ROE and ROA diverge most dramatically, and understanding why requires understanding the bank balance sheet itself. A commercial bank does not manufacture anything or generate electricity; it borrows short (deposits) and lends long (loans and advances), pocketing the spread, and it does this at a scale many multiples of its own equity. Nepal Rastra Bank's minimum paid-up capital requirement for commercial banks — NPR 8 billion, in place since the 2015 merger-driven capital hike — sits underneath balance sheets that, for the larger banks, now exceed NPR 150-250 billion in assets. That is an equity multiplier of roughly nine to twelve times, and it is precisely why bank ROE numbers look so much larger than bank ROA numbers.

The sector-wide numbers, and why they moved

Nepal's banking sector has been through a difficult multi-year stretch of tight liquidity, high provisioning for non-performing loans, and margin compression, and the profitability data reflects it plainly. Nepal Rastra Bank's own Financial Stability Report data and sector coverage through FY2024/25 show the aggregate Return on Equity for commercial banks falling into the high single digits — reported as low as roughly 7.7% in one widely cited aggregate reading — down from the low-to-mid teens the sector used to post in the years before the 2022-2023 liquidity crunch. A separate reading of the top ten banks by size for the second quarter of FY2081/82 put aggregate ROE at around 9.4%. Individual banks still vary widely around that average: well-capitalised, well-managed banks with strong CASA (current and savings account) deposit franchises can post ROE in the mid-teens even in a soft year, while smaller or asset-quality-challenged banks can sit in the low single digits or, in a bad year, near zero. ROA across the sector, consistent with the leverage story above, typically sits in a much narrower band of roughly 1.0% to 1.8%, rarely straying far from that regardless of which individual bank you look at — because ROA is a function of margin and efficiency, which move slowly, while ROE amplifies whatever leverage a bank happens to be running.

Net Interest Margin — the bank-specific profitability metric

No other NEPSE sector has an equivalent to Net Interest Margin, because no other sector's core business is borrowing at one rate and lending at another. NIM is calculated as:

NIM = Net Interest Income / Average Interest-Earning Assets

where Net Interest Income is interest income (from loans, investments in government securities, interbank placements) minus interest expense (on deposits, borrowings, debentures). NIM is the engine room of bank profitability — it drives the bulk of operating income, before fee income, and ultimately drives both ROA and ROE. Nepali commercial banks have typically operated with NIM in a range of roughly 3.0% to 4.5%, though the exact figure has compressed over the past two years as deposit rates rose faster than banks could reprice their loan books, and as intense competition for deposits during liquidity-tight periods pushed up the cost of funds. A bank's NIM trend, more than any single-period ROE snapshot, is the earliest and cleanest signal of whether its core lending business is getting healthier or getting squeezed — it moves before ROE does, because ROE also absorbs one-off items like loan loss provisioning reversals, tax adjustments, and trading gains that can mask or exaggerate the underlying trend for a quarter or two.

DuPont analysis — decomposing a bank's ROE

DuPont analysis breaks ROE into three multiplicative components, so that instead of one opaque number you get three transparent ones:

ROE = Net Profit Margin x Asset Turnover x Equity Multiplier

= (Net Profit / Total Operating Income) x (Total Operating Income / Average Total Assets) x (Average Total Assets / Average Equity)

For a bank, "Total Operating Income" is the appropriate revenue proxy — net interest income plus fee and commission income plus other operating income — rather than gross interest income, because gross interest income overstates the bank's actual economic throughput (a large chunk of it simply gets paid back out as interest expense to depositors).

Consider a worked, illustrative example built to match the disclosure format of a typical mid-to-large Nepali commercial bank (figures below are a stylized composite for teaching purposes, not a specific bank's actual reported numbers):

From this:

Notice what the decomposition reveals that the headline 12.9% ROE alone does not: this bank's operating margin (25.3%) is healthy and its asset efficiency (5.1% — essentially its "yield" on the whole balance sheet) is unremarkable but normal for the sector, and the entire ROE outcome depends heavily on the 10.0x equity multiplier. If Nepal Rastra Bank tightened capital adequacy requirements and forced this bank to raise fresh equity, pushing the multiplier down to 8.0x, ROE would fall to roughly 10.3% even with identical operating performance — nothing about the underlying business changed, only its capital structure did. This is exactly the kind of shift NEPSE investors lived through during the 2022-2023 period, when several banks issued rights shares or FPOs to meet capital requirements, diluting equity and mechanically compressing ROE even as NIM and asset quality were, in some cases, stable or improving.

The DuPont framework is also the cleanest way to compare two banks that post similar ROE through very different means — one bank might get there through a fat margin and modest leverage, another through a thin margin and heavy leverage. The first is the more resilient business in a downturn; the second is the one that gets hurt fastest when asset quality deteriorates, because its equity cushion is thinner relative to its asset book.

Lesson 40.3 — Hydropower: The Pre-COD Mirage and the Post-COD Reality

Hydropower is the sector where a single date on the calendar splits a company's entire profitability profile in two. That date is the Commercial Operation Date (COD) — the day the plant is commissioned, synchronized to the grid, and begins actually selling electricity to the Nepal Electricity Authority under its Power Purchase Agreement (PPA). Before COD, a hydropower company earns essentially no operating revenue. After COD, it becomes, in effect, an annuity-like cash generator with a fixed offtake price and a fixed offtake buyer (NEA) for the life of the PPA, typically 25 to 35 years, with generation subject mainly to hydrology.

Why pre-COD ROE is close to meaningless

Under Nepal Financial Reporting Standards (aligned with NAS 23 on borrowing costs), interest paid on construction-period loans is capitalised into the cost of the project rather than expensed through the profit and loss statement while the plant is under construction. This is the correct accounting treatment, but it has a side effect that trips up unwary investors: a hydropower company under construction shows little to no interest expense and little to no revenue, so its reported net profit during this phase is often a small positive number — arising almost entirely from bank interest earned on IPO proceeds and promoter capital sitting in fixed deposits while construction is underway — rather than from anything resembling the eventual power-generation business. A 1-2% ROE on a pre-COD hydropower company tells you almost nothing about what its ROE will look like once turbines are spinning. Investors who screen hydropower IPOs by pre-COD profitability metrics are, without realising it, screening for which promoters parked their construction float most efficiently in fixed deposits — not for which project will generate the best returns.

The post-COD ramp-up, and why it is lumpy

Commissioning does not flip a switch to full profitability overnight. In the first one to two years after COD, most Nepali hydropower companies go through a ramp-up period: initial capacity utilisation is often below design levels as operational teams learn the plant's actual hydrology response, minor defects liability issues get resolved with the EPC contractor, and — critically for run-of-river plants, which make up the overwhelming majority of NEPSE-listed hydropower — generation is highly seasonal. Wet-season months (roughly Ashadh through Ashoj, June through October) can generate several times the electricity of dry-season months (Poush through Chaitra, December through March), when river flows fall and some run-of-river plants generate at a fraction of installed capacity. A quarterly ROA or ROE reading taken in isolation, without knowing which season it covers, is close to uninterpretable for a run-of-river hydropower stock — a strong wet-season quarter can flatter annualized ROE readings, and a dry-season quarter can make a perfectly healthy company look unprofitable. NEPSE Trading's coverage of names such as Khanikhola Hydropower illustrates the pattern directly: stable or growing revenue on a full-year basis while the company continued to post losses in specific reporting periods, a combination that only makes sense once you separate seasonal generation swings and ramp-up effects from a genuine deterioration in the underlying business.

Once a hydropower company has run through two to three full hydrological cycles post-COD, its profitability profile stabilises and becomes genuinely comparable across companies and across years. Mature, well-established plants — the classic examples on NEPSE being names like Chilime Hydropower Company and the state-linked Upper Tamakoshi Hydroelectric project — have built long dividend-paying track records once past this maturation phase, with Chilime in particular known for consistent double-digit cash dividend payouts in mature years, reflecting a business that, once ramped up, throws off free cash flow well in excess of what it needs to reinvest.

The other structural driver of hydropower ROE, once mature, is leverage, and here hydropower resembles banking more than it resembles manufacturing: Nepali hydropower projects are typically financed at debt-to-equity ratios of roughly 70:30 or even 80:20 during construction, funded by a syndicate of commercial banks and development banks under NRB's hydropower lending exposure norms. Once operational and cash-generative, this high leverage means that even a modest return on the total project asset base translates into a much larger return on the comparatively thin equity slice — the same equity-multiplier mechanic that drives bank ROE, just applied to a physical, single-asset business instead of a loan book. This is precisely why mature hydropower ROE can look surprisingly close to bank ROE despite the two sectors having almost nothing else in common operationally.

Lesson 40.4 — Insurance: When Investment Income, Not Underwriting, Drives ROE

Insurance is the sector where the profitability story most often gets told backwards by casual NEPSE commentary. The instinctive assumption is that an insurer's profit comes from underwriting — collecting more in premiums than it pays out in claims and expenses. For Nepali life insurers in particular, that assumption is largely wrong, and understanding why is essential to reading their ROE correctly.

The life insurance profit engine: the investment portfolio

A life insurance company sits on a large and growing pool of policyholder reserves — the life fund — built up because premiums are collected upfront (often annually or in lump sums) while claims and maturity benefits are paid out years or decades later. Nepal's insurance regulator, the Nepal Insurance Authority (the renamed successor to the former Beema Samiti), mandates how this pool must be invested: a large share in government securities and bank fixed deposits, a regulated ceiling on equity market exposure, and specific limits on real estate and other asset classes. The scale of this life fund, for an established insurer like Nepal Life Insurance Company — the largest life insurer listed on NEPSE — dwarfs the company's own paid-up equity capital many times over, and the investment income this fund throws off (interest on government bonds, fixed deposit interest, dividend income, and realised gains on the regulated equity sleeve) typically contributes the majority of reported net profit in a normal year, far outweighing the technical underwriting result.

This has a direct consequence for ROE: because the equity base of most Nepali insurers is comparatively small relative to the size of the life fund and total assets they manage, and because investment income is largely a function of prevailing government securities yields and fixed deposit rates rather than underwriting skill, insurer ROE tends to move with interest rate cycles almost as much as it moves with premium growth or claims experience. Established, well-capitalised life insurers have historically posted ROE in the wide range of roughly 12% to 25%+ in strong years, materially higher than the banking sector average, precisely because investment income compounds on a large fund base sitting atop a comparatively thin equity layer — the insurance-sector version of the same leverage-like effect seen in banking and hydropower, except here the "leverage" is policyholder reserves rather than deposits or project debt.

Non-life insurance: thinner margins, more claims volatility

Non-life insurers (motor, fire, marine, engineering, and miscellaneous classes) run a structurally different book: shorter-tail liabilities, faster premium-to-claim turnaround, and a combined ratio (claims plus expenses, as a share of earned premium) that investors should track alongside ROE. Non-life underwriting profitability in Nepal has been persistently pressured by high motor claims frequency and price competition among the many licensed non-life insurers, meaning that, for this sub-sector even more than for life insurance, investment income on the comparatively smaller technical reserves is what keeps ROE positive and respectable in years when the combined ratio deteriorates. Non-life insurer ROE in Nepal has typically run somewhat lower and more volatile than life insurer ROE — commonly in a rough 8% to 15% band — reflecting both thinner reserve bases to invest and more claims-driven earnings volatility.

The other feature specific to insurance is Nepal Reinsurance Company Limited, the country's sole domestic reinsurer, which sits one layer removed from retail policyholders and whose profitability is driven by the ceded-premium share it accepts from primary insurers and its own investment portfolio — a useful reminder that "insurance sector" on NEPSE actually spans several distinct business models (life, non-life, reinsurance) each with its own profitability drivers, not one homogeneous group.

Lesson 40.5 — Manufacturing and Consumer Names: High Margins, Thin Equity, and the Illusion of an Extraordinary ROE

Manufacturing and consumer-goods names are a small but closely watched corner of NEPSE, dominated by long-established, brand-backed companies — Unilever Nepal Limited being the most frequently cited example, alongside bottling and beverage names tied to global consumer brands operating in Nepal. These companies behave almost nothing like banks, hydropower plants, or insurers, and their ROE patterns require a third, entirely separate mental model.

Why some manufacturing ROE numbers look almost unbelievable

Unilever Nepal is the textbook case. It carries an unusually small paid-up capital base relative to the scale of its operating profit — a legacy of decades of retained-earnings accumulation without proportionate equity dilution, combined with genuinely strong operating margins from a portfolio of established branded consumer products (soaps, detergents, personal care, tea) with real pricing power and distribution reach across Nepal. The arithmetic consequence is an ROE that can run into the tens of percent, and in some fiscal years has been reported far above what almost any bank, insurer, or hydropower company on NEPSE could plausibly post — not because the operating business is dozens of times better run than a well-managed bank, but because the equity denominator is so small relative to the profit numerator. A company can have a perfectly ordinary net profit margin and a perfectly ordinary asset turnover and still produce an extraordinary ROE if its equity multiplier (assets divided by equity, or in a simpler single-business context, the inverse of how much of the balance sheet is actually funded by shareholders' paid-up capital and reserves) is unusually high because of a thin paid-up capital history.

The more representative picture: ROA and operating margin

Because manufacturing and consumer companies do not run on financial leverage the way banks and (mature) hydropower companies do, their ROA is a far more representative and comparable figure — typically in a healthy mid-to-high single digit to low double-digit range for average NEPSE-listed manufacturers, and distinctly higher, sometimes reaching into the high teens or twenties, for the strongest branded consumer names with genuine pricing power and asset-light operating models (contract bottling and manufacturing arrangements, for instance, keep the fixed-asset base lean relative to revenue). The gap between a manufacturing company's very high ROE and its far more modest — though still strong — ROA is the clearest possible illustration of why ROE alone, unadjusted for capital structure, is the wrong tool for cross-sector comparison.

This is also the sector where dividend policy interacts most visibly with ROE: because these are often mature, low-capex-growth businesses that do not need to retain much capital for expansion, they tend to pay out a high proportion of profit as cash dividends, which keeps the equity base from growing much year to year — reinforcing, rather than correcting, the thin-equity, high-ROE pattern over time. A growing manufacturer that is reinvesting heavily (building new capacity, expanding distribution) will show equity growing faster than a mature payout-focused one, and its ROE trend should be read in that light — a declining ROE at a growth-phase manufacturer can simply reflect a larger equity base doing more (as-yet-unrealized) work, not a deteriorating business.

Lesson 40.6 — Building Your Own Cross-Sector Benchmark Table

Having walked through banking, hydropower, insurance, and manufacturing separately, the practical task for a Nepali investor holding — or considering — positions across more than one of these sectors is to build a mental (or literal, in a spreadsheet) benchmark table that lets you judge each stock against its own sector's realistic range, not against a single NEPSE-wide number that does not actually exist for any of them.

Three rules for using this table honestly

First, always locate a company within its own sector's row before judging the number — a 12% ROE is disappointing for a strong life insurer in a normal rate year, roughly average for a bank, and quite good for a two-year-old hydropower plant still working through its ramp-up. Second, always look at the ROE trend and the ROA alongside it, not the ROE level in isolation for a single period — a bank's ROE dip after a rights issue, a hydropower plant's ROE swing between wet and dry season quarters, and an insurer's ROE bump from a good year for government securities yields are all explainable by mechanics this chapter has covered, and none of them, by themselves, should trigger a buy or sell decision. Third, when you do need to compare across sectors — for instance, deciding whether fresh investable funds should go into a bank stock or a hydropower IPO — use ROA and the DuPont components, not headline ROE, as your primary cross-sector yardstick, and treat sector-specific metrics (NIM for banks, capacity utilisation and PPA tariff for hydropower, combined ratio and investment yield for insurers) as the qualitative context that explains why the ROA is what it is.

Applying this discipline does not mean profitability ratios stop being useful for cross-sector decisions — it means using them correctly. A retail investor who understands that a hydropower company's low ROE in year one post-COD is structural and temporary, that a bank's ROE compression after a rights issue is mechanical and often temporary, that an insurer's outstanding ROE in a high-yield year partly reflects the interest rate cycle rather than management skill, and that a manufacturer's spectacular ROE partly reflects decades-old capital structure rather than current brilliance — that investor is reading the same numbers everyone else on NEPSE sees, but reading them correctly, sector context and all.