Hydropower Project Accounting Logic

First published 22 Aug 2026 · Last verified 29 Aug 2026

Lesson 32.1 — The Two Lives of a Hydropower Balance Sheet

Open the annual report of almost any hydropower company on NEPSE and you will notice something odd if you are used to reading manufacturing or banking accounts: for years, sometimes a full decade, the income statement is nearly empty while the balance sheet grows enormous. A company can report zero revenue, zero profit, and yet carry assets worth tens of billions of rupees. This is not a company in distress. It is a company under construction, and construction-phase hydropower accounting follows a logic entirely different from operation-phase accounting. Understanding the seam between these two phases — and knowing which phase a company you are evaluating actually sits in — is the single most important skill in reading a Nepali hydropower stock.

During construction, nearly every cost a project incurs is capitalised rather than expensed. Land acquisition and resettlement, tunnel excavation, penstock and powerhouse civil works, turbines and generators, transmission line interconnection, consultancy and supervision fees, and — critically — interest on the loans used to fund all of the above, all accumulate on the balance sheet under "Capital Work in Progress" (CWIP). Nothing hits the profit and loss account except perhaps small amounts of administrative overhead that auditors judge unrelated to bringing the asset to its intended use. The company can be burning through its entire equity base and drawing down its full debt facility, and its income statement will still show a thin trickle of bank interest income on unspent share proceeds sitting in a construction account.

The moment the plant reaches Commercial Operation Date (COD) — sometimes called Rated Commercial Operation Date, RCOD, in the tax rebate schedules — everything changes. CWIP is reclassified into Property, Plant and Equipment (PPE) across specific asset categories: civil structures, hydro-mechanical equipment, electro-mechanical equipment, transmission assets. Depreciation begins. Interest capitalisation stops, because Nepal Accounting Standard 23 (Borrowing Costs) — the local equivalent of IAS 23 — only permits capitalising borrowing costs directly attributable to acquiring, constructing, or producing a qualifying asset for the period until that asset is "substantially complete and ready for its intended use." From COD onward, interest expense flows straight through the income statement, and so does depreciation, and so does the entire PPA revenue stream. A company that showed almost no P&L activity for eight years can suddenly show a debt-servicing burden, a depreciation charge, and a revenue line all in the same annual report — and the market frequently misprices this transition in both directions, treating a pre-COD company as if it were already earning, or treating a freshly commissioned company's first full year of depreciation and interest as a permanent decline in profitability rather than the new steady state.

The practical skill for an investor is to identify, from the notes to accounts, exactly where a company sits on this timeline: percentage physical completion, expected COD, and — this is the number analysts often skip — the cumulative interest capitalised to date. That last figure tells you how much of the eventual asset base is not concrete and steel but financing cost, and it previews the depreciation and amortisation charge the company will carry for the rest of its operating life.

Lesson 32.2 — Interest During Construction: The Silent Multiplier

Nepali hydropower projects are financed on capital structures that would be considered dangerously leveraged in almost any other sector — commonly 70:30 or even 80:20 debt-to-equity, with the debt drawn from a syndicate of Nepali commercial banks (a "consortium loan," since single-bank lending limits under Nepal Rastra Bank's Single Obligor Limit rules make one bank funding a large project legally impossible). A run-of-river project with a five- or six-year construction period will typically capitalise interest for the entire period, and because construction delays in Nepal are the rule rather than the exception — landslide-damaged access roads, monsoon-halted tunneling, contractor disputes, transmission-line right-of-way litigation — the capitalised interest bill frequently ends up being a bigger surprise to shareholders than the base construction cost itself.

Interest During Construction (IDC) capitalisation

The mechanics are straightforward but consequential. NAS 23 requires a company to capitalise the actual borrowing costs incurred on funds specifically borrowed for the project, net of any investment income earned on temporarily parking undrawn loan proceeds. Where general corporate borrowing is used, a weighted average capitalisation rate applies to expenditure on the qualifying asset. Nepal Rastra Bank's unified directives to licensed banks go a step further, giving banks explicit guidance on how long-gestation infrastructure loans — hydropower chief among them — may treat accrued-but-unpaid interest during the construction period, which in turn shapes how promoters structure moratoriums and how the capitalised interest line grows on the borrower's books.

The scale this can reach is best illustrated by Nepal's largest domestically financed plant, the 456 MW Upper Tamakoshi Hydroelectric Project (NEPSE: UPPER), built and operated by Nepal Electricity Authority's subsidiary Upper Tamakoshi Hydropower Limited. The project's construction cost estimate rose from roughly NPR 49.29 billion to about NPR 52.29 billion as delays compounded — first from contractor underperformance (the original hydro-mechanical contractor, India's Texmaco Engineering and Railway Company, withdrew from the project, with Austria's Andritz Hydro brought in to complete the work), and further from the damage the 2015 Gorkha earthquake inflicted on under-construction structures. But the number that matters most for an accounting-literate investor is a different one: by the time the plant reached commercial operation, total project cost — construction cost plus capitalised interest — had climbed toward roughly NPR 76 billion, of which approximately NPR 24 billion was interest capitalised during construction. Financing for the project was split between about NPR 10.59 billion of share capital (with NEA, Nepal Telecom, Citizen Investment Trust, and Rastriya Beema Sansthan among the major institutional shareholders) and roughly NPR 41.70 billion in bank debt.

Here is the reported financing structure for Upper Tamakoshi at the point construction cost estimates were last revised, alongside its post-COD scale, to show how the pieces fit together:

A general rule follows from this: the longer a project takes to build, the larger the share of its eventual asset base that is financing cost rather than construction cost — and the higher its post-COD depreciation and interest expense will be relative to what the original project feasibility study assumed. When you read a hydropower prospectus or annual report projecting future profitability, always ask what construction timeline the projection assumes, and compare it against the company's actual physical progress percentage disclosed in the same report.

Lesson 32.3 — Reading the PPA: Wet Season, Dry Season, and the Escalation Clock

A hydropower company's revenue is not a single number multiplied by units sold. It is the output of a contract — the Power Purchase Agreement with Nepal Electricity Authority — whose structure was negotiated at a specific point in time and which determines, almost mechanically, what the company can report as revenue for the next 20 to 30 years. Reading a hydropower income statement without reading its PPA is like reading a bank's income statement without knowing its interest rate on deposits.

Nepal's rivers are monsoon-fed, so a run-of-river plant (no storage reservoir) generates several times more energy in the wet season (roughly mid-June to mid-October) than in the dry season, when flows can drop to a fraction of wet-season levels. NEA's standard PPA template for small and medium run-of-river IPPs has, for years, split the tariff itself along the same seasonal line, paying a lower per-unit rate for wet-season energy (when supply is abundant and hydrologically "cheap" to the buyer) and a materially higher rate for dry-season energy (when the same unit is scarcer and more valuable to the grid). Balephi Hydropower Limited's PPA with NEA, signed in December 2015, is a representative example of this template: NPR 4.80 per kWh for wet-season energy against NPR 8.40 per kWh for dry-season energy — a dry-season premium of 75 percent over the wet-season rate — with the contract further specifying escalation of 3 percent per year, applied seven times, starting from the thirteenth month after commercial operation (Balephi's own escalation schedule shows one escalation forfeited because commercial operation started later than planned, which is itself a reminder that construction delay does not just raise cost — it can also compress the revenue-escalation window a company was counting on).

Because tariffs escalate annually for a defined number of years and then flatten, a company's revenue-per-unit trajectory is knowable years in advance if you have the PPA schedule — which is exactly why serious hydropower analysis in Nepal starts from the PPA annexures, not from trailing income statement growth rates.

Storage projects change the calculus further. A storage or peaking-storage plant can shift generation from wet season to dry season by holding back water, which is precisely why Nepal Electricity Authority and the Electricity Regulatory Commission (ERC) have spent the past several years designing a differentiated tariff framework specifically to make storage projects financeable — since a storage plant's entire value proposition to the grid is dry-season and peak-hour firm capacity, which is worth substantially more than the same energy delivered in the monsoon glut. An ERC discussion paper on storage-hydro tariff design illustrates how wide this gap is expected to be, and how individual large storage projects have priced their own asks even wider once actual Nepali financing costs (materially higher debt costs and shorter tenors than the concessional assumptions used in illustrative models) are factored in.

A further structural feature to watch is whether a PPA is take-or-pay or take-and-pay. Under take-or-pay, NEA is contractually obligated to pay for a defined quantity of deliverable energy whether or not it actually draws that energy off the grid (subject to force majeure and grid-availability carve-outs), which effectively transfers demand risk to the buyer and gives the IPP a revenue stream closer to a fixed annuity. Under take-and-pay, NEA pays only for energy it actually takes, leaving the IPP exposed to curtailment risk — a real issue in a system that has, at various points, faced transmission bottlenecks preventing it from absorbing all available wet-season generation. NEA has moved smaller hydropower PPAs from take-and-pay toward take-or-pay terms in recent years, which is a meaningfully positive development for revenue predictability, but the legacy PPA a given listed company signed years ago may still carry the older, weaker structure — so the take-or-pay/take-and-pay distinction belongs on your checklist for every hydropower stock, not just new listings.

What the escalation clock means for your model

Because escalation is time-bound (commonly capped at seven or eight annual steps before the tariff plateaus), a hydropower company's per-unit revenue growth is front-loaded and mechanical, not driven by operating performance. An investor modelling five years forward should pull the exact escalation schedule from the PPA rather than assuming a flat growth rate — and should separately flag the year the escalation stops, because that year marks the point after which revenue growth can only come from higher plant availability, not from contractual tariff increases.

Lesson 32.4 — Depreciation, License Life, and the Financial Asset Question

Once a hydropower asset is capitalised, how should it be depreciated — and, more fundamentally, is "depreciation" even the right accounting model for an asset built under a Build-Own-Operate-Transfer (BOOT) survey license that must eventually revert to the Government of Nepal? This is a live technical debate in Nepali accounting practice, and it matters to investors because the classification choice changes reported profit, tax timing, and the comparability of hydropower companies against one another.

Most listed Nepali hydropower companies depreciate plant and equipment as ordinary Property, Plant and Equipment under NAS 16, generally on a straight-line basis over useful lives set to align with (or fall within) the survey/generation license period — typically structured so the asset is substantially depreciated by the time the BOOT transfer obligation to the government falls due. This treatment produces a familiar depreciation-and-interest income statement, much like any other capital-intensive company, and it is the treatment nearly all NEPSE-listed hydropower issuers currently use.

But a technical argument exists — grounded in IFRIC 12, Service Concession Arrangements, the international standard governing exactly this kind of infrastructure-under-license arrangement — that many Nepali hydropower BOOT projects should instead recognise a financial asset (a receivable) rather than PPE, whenever two conditions both hold: the grantor (NEA, acting for the state) controls what services the operator must provide and at what price, and the grantor retains a significant residual interest in the infrastructure at the end of the concession. A run-of-river plant with a fixed, NEA-dictated PPA tariff, a take-or-pay payment mechanism tied to availability rather than market-negotiated pricing, and a mandatory transfer of the plant to the Government of Nepal at the end of a fixed license term arguably satisfies both conditions — which would mean the "right" accounting is to recognise a financial asset that unwinds through an effective-interest calculation over the concession life, not a depreciating fixed asset.

The clearest illustration in the Nepali market of a project fitting the financial-asset description is Tamakoshi-V, a 99.8 MW run-of-river plant in Dolakha district operating under a 30-year concession with a fixed tariff (subject to the standard 3 percent annual escalation) and a mandatory transfer of the asset to the Government of Nepal at the end of the concession — textbook IFRIC 12 conditions on paper. By contrast, Butwal Power Company (BPC), which operates the Andhikhola (9.4 MW) and Jhimruk (12 MW) plants, currently carries its concession rights as an intangible asset under NAS 38 rather than as PPE or a financial asset, on the argument that BPC operates its own distribution network and therefore bears demand risk directly rather than having NEA guarantee its offtake — a genuinely different risk allocation that plausibly does justify a different accounting model. Even within that intangible-asset treatment, though, the amortisation method matters: straight-line amortisation over the license term produces a very different year-by-year profit profile than usage-based (units-of-production) amortisation tied to actual generation, and the two are not always easy to distinguish from the face of the financial statements alone.

None of this means an investor needs to resolve the IFRIC 12 debate independently — that is genuinely contested technical ground even among Nepali chartered accountants. The practical takeaway is narrower: because Nepali hydropower companies are not fully uniform in how they classify and depreciate their core asset, you cannot compare depreciation-to-revenue ratios across two hydropower companies and assume you are comparing like with like. Always check the accounting policy note before drawing a cross-company conclusion from depreciation figures alone.

Lesson 32.5 — Royalty, Tax Holidays, and the Government's Take

A hydropower company's relationship with the state runs deeper than its PPA counterparty. Every licensed hydropower generator in Nepal pays royalty to the Government of Nepal under the Electricity Act framework, structured in two components — a capacity royalty (a fixed annual charge per kW of installed capacity) and an energy royalty (a percentage of the value of energy generated or sold) — and both components step up sharply once a project passes its fifteenth year of commercial operation.

The royalty step-up at year 16

For the first fifteen years of operation, a project typically pays a capacity royalty of roughly NPR 100 per kW per year and an energy royalty of about 2 percent of revenue. From the sixteenth through the thirtieth year, both figures rise steeply: capacity royalty to roughly NPR 1,000 per kW per year (a tenfold increase) and energy royalty to about 10 percent of revenue (a fivefold increase). For a large plant, this is not a rounding adjustment — it is a structural shift in the cost base that occurs on a fixed calendar regardless of the company's operating performance, and it is a shift many retail investors modelling a hydropower stock's "steady state" margin never account for, because they anchor their expectations to whatever margin the company is currently reporting in, say, year 6 or year 8 of operation.

This royalty schedule runs alongside a separate, and currently more investor-visible, tax incentive schedule administered under the Income Tax Act: electricity generation enterprises are taxed at a 20 percent corporate rate, but qualifying companies (generally those reaching Rated Commercial Operation Date within a government-specified incentive window) receive a 100 percent tax rebate for their first ten years of operation and a 50 percent rebate for years eleven through fifteen — meaning a project can run essentially tax-free for a decade and then pay half the standard rate for five more years before settling into full taxation. Dividend distributions carry a further 5 percent dividend tax at the shareholder level, and capital gains on listed hydropower shares are taxed at 7.5 percent long-term or 10 percent short-term, same as other NEPSE-listed equity.

There is also a construction-phase tax and duty regime worth knowing, because it explains part of why the capital cost structure looks the way it does: import duty on plant, machinery, and steel penstock/pipe is set at a nominal 1 percent with full exemption from the standard 13 percent VAT, while domestic engineering, civil, and transportation costs remain subject to the ordinary 13 percent VAT — with the government separately offering a partial VAT refund (historically around NPR 0.5 crore, i.e., roughly NPR 5 million, per MW of capacity) against engineering, transportation, and civil-construction VAT paid. These construction-phase concessions are precisely the sort of detail that shows up as "other income" or reduced CWIP additions in a company's financial statements without being separately labelled, so a careful reader reconciling projected versus actual project cost should check the notes for VAT refund receivables before assuming a cost overrun is larger than it actually is.

Lesson 32.6 — Debt, Currency, and Governance Risk: Reading Between the Lines

The debt-heavy capital structure that makes IDC so consequential (Lesson 32.2) does not disappear once a plant is commissioned — it becomes the central determinant of whether operating cash flow actually reaches shareholders as dividends, or is absorbed by debt service for years after COD. Most Nepali hydropower debt is rupee-denominated, drawn from domestic bank consortiums, which removes the currency-mismatch risk that plagues hydropower financing in many other developing markets — but it does not remove interest-rate risk, since Nepali bank lending rates float with the base-rate cycle, and a plant financed at 70:30 or 80:20 debt-to-equity carries a debt-service coverage ratio that can swing meaningfully across a single interest-rate cycle even with no change in physical output. Where foreign-currency debt or foreign-currency PPA elements do appear — typically on larger cross-border or donor/multilateral-financed projects — the exposure is real and needs to be checked explicitly in the borrowings note: a hydropower company servicing a foreign-currency loan against rupee-denominated PPA revenue is carrying open currency risk that a purely domestically financed peer does not have.

Two live cases illustrate how debt structure, delay, and governance combine into the risks that matter most in practice.

Upper Tamakoshi's post-commissioning governance strain is the more urgent of the two as of this writing. Despite being Nepal's largest domestically financed plant and a long-time dividend-paying blue chip, UPPER had — as of mid-2026 — gone three consecutive years without holding an Annual General Meeting, with the last AGM held in Shrawan 2080 for fiscal year 2079/80, leaving shareholders without a formal channel to review audited results or elect independent directors for an extended period. Rating agency ICRA Nepal downgraded the company into its 'D' (default) category following debt-servicing delays that exceeded thirty days, even as the company's long-term liabilities stood at roughly NPR 45.36 billion and it carried negative retained earnings of about NPR 10.77 billion — a stark reminder that "largest plant in the country" and "high dividend history" do not immunize a company against a debt-service or governance crisis. Analysts have also flagged a structural governance concern: NEA's Managing Director has simultaneously served as UPPER's board chairperson, concentrating institutional control at a company where NEA and other public institutions together hold roughly 51 percent of shares, limiting the practical influence of minority shareholders precisely when governance oversight matters most.

The second case, GMR Upper Karnali, illustrates delay risk at the extreme end and the added complexity of cross-border, export-oriented hydropower. The 900 MW project, first taken up by India's GMR Group under an initial understanding roughly two decades ago, spent some eighteen years in survey, PPA negotiation (including a prolonged tariff price standoff with NEA), and financing-arrangement limbo before construction work finally began in mid-2025 — with the project structured to export power to India and, under a trilateral arrangement, Bangladesh. For a project of this scale and export orientation, the accounting questions of Lessons 32.1 through 32.5 all apply, but with an added layer: cross-border PPA pricing, multi-jurisdictional regulatory approval, and financing that may combine foreign lenders with foreign-currency-denominated debt in a way most domestically consumed RoR plants do not carry.

Finally, the accounting and disclosure issues covered in this chapter feed directly into a live listing-market controversy: Nepal's Securities Board (SEBON) has, in recent years, applied a "real net worth" threshold of roughly NPR 90 per share — a figure that appears nowhere in the Securities Act and was set administratively following a Public Accounts Committee directive in late 2023 — as an informal bar for IPO approval. Because construction-phase hydropower companies structurally carry lower reported net worth per share than operating companies (their equity has not yet been "proven" by revenue, even where their underlying project economics are sound), this threshold has disproportionately stalled hydropower IPOs specifically: roughly NPR 66.23 billion in intended capital raises across 98 companies were reported frozen in the approval queue, with fourteen hydropower companies — including names such as Laughing Buddha Power Nepal, Yambaling Hydropower, Unique Hydropower, Beni Hydro, and Puwa Khola One Hydro — formally removed from the queue altogether. One documented side effect has been the growth of an unregulated pre-IPO share market in which promoter-stage shares change hands well above the NPR 100 face value, with buyers told to expect post-listing prices in the NPR 1,500-2,500 range — precisely the kind of informal, disclosure-free trading environment that hydropower's construction-phase accounting opacity tends to invite.