Case Study 9 — A Hydropower Overrun and Delay

First published 26 Aug 2026 · Last verified 29 Aug 2026

Case Study 9 — A Hydropower Overrun and Delay

Suman Ghimire had done everything the earlier chapters of this Canon told him to do. He read the prospectus twice. He looked up the river basin on a map before buying a single share. He even asked his cousin, who worked for a trekking company in Dolakha, to send him photographs of the construction site. By the time the Upper Tamakoshi Hydropower Limited shares reached the hands of ordinary retail investors, Suman felt he understood the company better than most people at his brokerage counter in Putalisadan.

Upper Tamakoshi Hydropower Limited, commonly traded on NEPSE under the symbol UPPER, is a 456 megawatt (MW) run-of-river hydropower project on the Tamakoshi River in Dolakha district. A megawatt is simply a unit of electrical power — a way of measuring how much electricity a plant can produce at any given instant, the same way a car's engine is rated in horsepower. A run-of-river plant, unlike a storage or reservoir plant, does not hold back a large lake of water behind a tall dam. It diverts a portion of the river's flow through a tunnel and a set of pipes to spin turbines, then returns the water to the river downstream. This matters for our story because run-of-river plants are hostages to two things: the river's natural rhythm through the year, and whatever man-made structure — a tunnel, a headrace canal, an intake — carries the water from the river to the turbines. If that structure is damaged, the plant does not merely produce less power. It can stop producing power entirely, sometimes for years.

Upper Tamakoshi was owned and built by a subsidiary of Nepal Electricity Authority (NEA), Nepal's state-owned power utility, but it was financed and structured in a way that involved a great deal of Nepali capital beyond the state itself — loans from Nepali institutions like the Employees Provident Fund and Citizens Investment Trust, plus share ownership eventually opened to project-affected locals and then to the wider investing public. This made it, at the time, the largest hydropower project built substantially with domestic Nepali financing and expertise, a matter of considerable national pride. It is exactly the kind of project a patriotic, income-hungry Nepali retail investor would fall in love with — and exactly the kind of project where that love needs to be checked against arithmetic.

This chapter follows Suman's journey with Upper Tamakoshi from the original construction timeline, through the 2015 earthquake and the 2016–17 monsoon floods that damaged the project's tunnel and access infrastructure, to the eventual commissioning years later than planned and at a cost far above the original budget — and, as we will see, through several more difficult years after that. Along the way we will apply Chapter 64's full seven-dimension Canon Score to Upper Tamakoshi as it actually stands today, and revisit the hydropower financial modelling logic from Chapters 42 through 44 to show exactly how a disciplined investor could have priced delay risk into the original purchase decision, and how to correctly re-underwrite the position once a delay actually landed. As you will see, a construction delay does not touch all seven of Chapter 64's dimensions equally — some move sharply, some barely move at all, and the discipline of scoring them separately is exactly what keeps an investor from either panicking or staying blindly loyal to a story that has changed.

Lesson 91.1 — Reading a Hydropower Project Before the Turbines Turn

Before we get to what went wrong, we need to understand what Suman was actually buying, because a share in a hydropower company is not really a share in "electricity." It is a claim on a very specific, very long-dated cash flow stream that depends on a chain of assumptions holding true, one after another, like a line of dominoes.

The first domino is the Power Purchase Agreement, or PPA. A PPA is a long-term contract, typically running 20 to 35 years, in which the hydropower company agrees to sell its electricity to a single buyer — in Nepal's case, almost always NEA — at a pre-agreed tariff, or price per unit of electricity. Think of a PPA the way you might think of a fixed-rent lease: it tells you, in advance, what money is coming in and for how long, provided the tenant (NEA) pays and the building (the power plant) is standing and functional. Upper Tamakoshi's PPA set differentiated tariffs for the wet season, when Nepal's rivers run high and electricity is abundant, and the dry season, when river flows shrink and electricity becomes scarce and valuable. This wet-dry tariff differential is a recurring theme across nearly every Nepali hydropower company's income statement, and it means a plant's revenue is never a flat, smooth number — it swings meaningfully across the fiscal year.

The second domino is capacity factor — the percentage of a plant's theoretical maximum output that it actually produces over a year, once you account for the river running low in winter, scheduled maintenance, and unplanned outages. A run-of-river plant on a glacier-fed Himalayan river might have a capacity factor anywhere from 45 to 60 percent, because winter flows are a fraction of monsoon flows. Chapter 42 walked through how to build a simple annual generation estimate: multiply installed capacity by 8,760 hours in a year, then multiply by your estimated capacity factor, to get expected annual units of electricity generated. That number, multiplied by the blended tariff across wet and dry seasons, gives you top-line revenue — before you have even opened the cost side of the model.

The third domino, and the one this case study is really about, is construction timeline and capital cost. Nearly every hydropower project is financed with a mix of equity (money from shareholders, including NEA itself and public investors like Suman) and debt (loans from banks, provident funds, and sometimes multilateral lenders). During construction, before the plant generates a single unit of revenue, the debt still accrues interest. This accrued interest during the construction period is added to the total project cost rather than expensed immediately, through an accounting treatment called Interest During Construction, or IDC.

This is the single most important piece of arithmetic in this entire case study, so it is worth sitting with before we move to what actually happened to Upper Tamakoshi. If a project was financed 70 percent by debt at, say, a 10 percent annual interest rate, and it is delayed by two extra years, the IDC alone on that debt — compounding, since unpaid interest is often itself capitalised and grows in subsequent periods — can add a cost equivalent to 14 to 20 percent or more of the original loan amount, before you even count the extra cost of remobilizing contractors, replacing damaged equipment, or repairing washed-out access roads. A hydropower delay is therefore never a "linear" problem. It is a compounding problem.

Suman's original 2016-era thesis on Upper Tamakoshi, before he understood any of this fully, went roughly like this: a 456 MW plant, one of Nepal's largest, backed by the state utility, with financing already largely secured, targeting commissioning within a few years, selling into a power-hungry national grid that was still running daily load-shedding at the time — surely that is close to a sure thing. That thesis was not wrong about the demand side. It was dangerously incomplete about the construction-risk side.

Lesson 91.2 — When the River Fights Back: The Chain of Delays

Construction on Upper Tamakoshi began in the early 2010s, with an original commissioning target that project documents and NEA statements at the time placed around the middle of the decade — roughly 2015 to 2016. This was already an ambitious target for a project of its scale and terrain: a headrace tunnel several kilometers long bored through the Himalayan foothills, a steep-mountain access road, and a powerhouse cavern, all in a district with difficult logistics.

Then, in April 2015, the Gorkha earthquake struck. Dolakha district, where Upper Tamakoshi sits, was among the hardest-hit districts in the entire country. The earthquake and its aftershocks damaged the project's tunnel works, access roads, and camp infrastructure, and — just as importantly — diverted skilled labor, engineers, and construction materials across the country toward emergency reconstruction. A hydropower project does not exist in a bubble; it competes for cement, steel, machinery, and manpower with every other reconstruction effort happening around it. The earthquake alone pushed the project's timeline back by roughly a year or more, by most contemporary accounts.

Just as the project was recovering and remobilizing, the monsoon seasons of 2016 and 2017 brought further damage. Flash floods and landslides along the Tamakoshi valley — a steep, narrow, geologically active corridor — damaged the headrace tunnel intake area and the access roads and bridges the project depended on to move equipment and materials in and out of the site. Publicly reported accounts from the period describe the project's own engineers estimating additional years of delay from the flood damage alone, on top of the earthquake-related delay already absorbed.

By the time Upper Tamakoshi was finally commissioned — with grid connection and testing completed around 2021 — the project had slipped from an original target of roughly 2015–2016 to an actual commissioning roughly five to six years later. This is not a footnote-level delay. It is a delay on the same order of magnitude as the original construction period itself. A project planned to take about four to five years from groundbreaking to commissioning instead took nearly a decade.

Delays of this scale are not unique to Upper Tamakoshi in Nepal's hydropower history — projects have been repeatedly delayed by monsoon damage, contractor disputes, transmission-line bottlenecks (where the power plant is ready but the high-voltage line needed to evacuate its electricity to the national grid is not), and land acquisition disputes with local communities. Upper Tamakoshi is simply one of the most thoroughly documented and highest-profile examples, precisely because it was meant to be a flagship of domestically financed Nepali hydropower.

Lesson 91.3 — Running the Numbers: What a Multi-Year Delay Actually Costs

Let's now do what Chapters 42 through 44 taught us to do with any hydropower position: build a simple model, then stress it.

Upper Tamakoshi's original approved project cost was NPR 35 billion, financed with a debt-to-equity structure weighted toward debt — a common pattern in Nepali hydropower, where debt makes up 70 percent or more of total financing, with the remainder coming from equity contributed by NEA and, eventually, public shareholders including project-affected locals.

Under the original timeline, debt service — the repayment of loan principal and interest — was expected to begin only after commissioning, once the plant started earning tariff revenue. During the construction years themselves, interest on the debt already drawn down was capitalised as IDC, added to the project's total cost rather than paid out of pocket. This is normal and expected for the planned construction period. The problem is what happens when the planned construction period roughly doubles.

Contemporaneous reporting on Upper Tamakoshi lets us watch the IDC mechanism bite in real time rather than merely describe it. Loans were carrying an interest rate of roughly 11 percent, and by mid-2019 — with the project still years from commissioning — cumulative interest charges on the project's debt had climbed from about NPR 6.7 billion in 2016 to roughly NPR 14.4 billion. By that point, the project's cost excluding capitalised interest stood near NPR 49 billion, but including that capitalised interest the figure was already closer to NPR 69 billion, on a project that had not yet produced a single billable unit of electricity. The rupee's roughly 25 percent depreciation against the dollar and euro over the same period made imported equipment and contractor payments more expensive still, and a crane failure during penstock installation in May 2019 added yet another few months of delay on top of everything else. By the time the project was finally reconciled after commissioning, the per-megawatt cost had risen from an original NPR 77 million per MW to roughly NPR 196 million per MW — a 154 percent increase, putting total project cost at approximately NPR 89 billion against the original NPR 35 billion budget, roughly two and a half times the original estimate. A meaningful share of that increase came from direct repair and reconstruction costs after the earthquake and floods. A larger share of it came from exactly the IDC compounding mechanism described above: money that shareholders never "spent" in any visible sense, but that was quietly added to the capital base they would now need to earn a return on.

This connects directly to a second concept from Chapter 43: the Debt Service Coverage Ratio, or DSCR. DSCR measures how many times over a project's annual operating cash flow can cover its annual debt repayment obligations — calculated as operating cash flow divided by scheduled principal and interest payments for the year. A DSCR of 1.0 means the project generates exactly enough cash to service its debt with nothing left over; lenders typically require a minimum DSCR comfortably above 1.0 (often 1.2 to 1.3 or higher) as a safety cushion.

A larger total project cost, financed with a similar debt proportion, means a larger absolute debt balance to service once revenue finally starts flowing. Even if Upper Tamakoshi generates precisely the electricity its engineers originally projected, its shareholders are now servicing a debt load roughly proportional to a NPR 89 billion project rather than a NPR 35 billion one. The plant's physical capacity to generate electricity did not change. The cost of the capital sitting behind that capacity did — and as Lesson 91.5 will show, this is exactly why the company kept losing money for years after the turbines finally started turning.

This is the crucial, underappreciated lesson of every hydropower overrun: delay risk and cost-overrun risk are really one and the same risk wearing two names, connected by the mechanism of capitalised interest. An investor who only asks "will the plant get built?" is asking half the right question. The full question is "will the plant get built, on what timeline, financed by how much additional debt, and will the resulting per-unit cost of electricity still clear a healthy return once tariffs are applied against a larger capital base?"

Lesson 91.4 — Hemisphere 1: Liquidity, Governance, and Durability

This Canon has, since Chapter 64, asked you to score any prospective holding across seven real dimensions rather than fall in love with a single narrative: Financial Strength & Profitability (20 points), Governance & Promoter Behaviour (15 points), Liquidity & Tradability (10 points), Valuation Reasonableness (15 points), Sector & Business Model Durability (15 points), Growth Trajectory (15 points), and Dividend & Capital Return Discipline (10 points), summing to 100 — with a governance override that caps the whole score in the Weak/Avoid band if the Governance sub-score falls below 5 out of 15. Let's run Upper Tamakoshi through all seven, honestly, using where it actually stands today rather than the story Suman told himself in 2016.

Three of the seven dimensions are best thought of together, because a multi-year construction delay barely moves them one way or the other: Liquidity & Tradability, Governance & Promoter Behaviour, and Sector & Business Model Durability. These describe what kind of company Upper Tamakoshi structurally is, not how its particular construction history unfolded.

Liquidity & Tradability (out of 10). Upper Tamakoshi is one of the largest hydropower listings on NEPSE by market value — roughly NPR 40 billion in market capitalisation on 211.8 million shares, all of it publicly held with no separate promoter block locking up supply. That size and float support real day-to-day tradability. Against that, the share has been in a clear downtrend for over a year, sliding from a 52-week high near NPR 243 to the mid-NPR 180s, which tends to thin out willing buyers even in a large-cap name. We score this 7 out of 10 — a genuinely tradable large-cap hydropower name, but not a top mark, given the cooling momentum.

Governance & Promoter Behaviour (out of 15). Upper Tamakoshi is a subsidiary of Nepal Electricity Authority, the state utility, financed substantially by Nepali institutions including the Employees Provident Fund and Citizens Investment Trust. There is no evidence here of the kind of self-dealing, related-party lending, or leadership fraud this Canon has flagged in other case studies — the record is one of execution difficulty, not misconduct. That said, execution difficulty is itself a governance fact: the project missed its commissioning deadline five separate times, and its per-megawatt cost rose 154 percent before the project was finally reconciled. On the positive side of the ledger, management has shown real capital discipline once the plant was operating — most notably renegotiating its long-term loan rate down from 8.25 percent to 6.75 percent for three years, a concrete, provable action rather than a promise. We score this 11 out of 15: comfortably above Chapter 64's 5-point override floor, reflecting a credible, non-fraudulent sponsor with a genuinely weak execution record that it has since worked to repair.

Sector & Business Model Durability (out of 15). Chapter 65 treats a PPA-backed hydropower project's basic moat as its textbook full-marks case: a 20-to-35-year contracted buyer in NEA, a tariff set in advance, and no exposure to a competitive spot market. That structural moat does not weaken just because a specific project ran years behind schedule, so this sub-component scores a full 8 out of 8. The other half of Durability is concentration risk — and here, like every hydropower company examined in this Canon, Upper Tamakoshi sells effectively all of its output to a single buyer, NEA. Chapter 65 flags this near-universal single-offtaker dependency as a real but only partially quantifiable risk across the entire sector, and consistent with Chapters 84, 87, 88, and 90, we score this sub-component 4 out of 7. Durability totals 12 out of 15.

Hemisphere 1 — the three dimensions least disturbed by a construction delay — comes to 7 + 11 + 12 = 30 out of 40.

Lesson 91.5 — Hemisphere 2: The Four Dimensions a Delay Actually Moves

The remaining four dimensions — Financial Strength & Profitability, Valuation Reasonableness, Growth Trajectory, and Dividend & Capital Return Discipline — are exactly where a multi-year cost overrun shows up, because they depend on real financial statements, and Upper Tamakoshi did not produce any until the plant was finally generating revenue in 2021. What those statements have shown since is sobering: the story does not end at commissioning.

Financial Strength & Profitability (out of 20). Once Upper Tamakoshi finally started billing NEA for electricity, it did not become profitable. It posted losses for four consecutive fiscal years, driven by the interest burden on its now-roughly-doubled-and-a-half debt load and by "hydrology penalties" — contractual charges NEA can levy when a plant's actual generation falls short of its contracted design energy in a given season — which cost the company close to NPR 1 billion in total across those four years, rising from a token NPR 3.3 million in 2021/22 to NPR 220 million in 2024/25. In fiscal year 2024/25 alone, revenue was a healthy NPR 6.92 billion, up 19.4 percent year over year, yet the company still posted a net loss of NPR 2.57 billion — a net margin of negative 37 percent — on earnings per share of negative NPR 12.15 and a return on equity of negative 22.8 percent. Only in the most recent fiscal year did the picture turn, after management renegotiated its loan rate down and posted a rare penalty-free year, with the company projecting roughly NPR 1 billion in profit and trailing EPS turning positive at NPR 4.37. We score the sub-components as follows: return on equity 2 out of 8 (a single quarter's worth of positive earnings after four straight loss years does not yet make a trend), leverage 2 out of 7 (the balance sheet is still carrying the debt built up by the overrun, now serviced at a lower but still meaningful rate), and earnings quality 1 out of 5 (revenue is genuinely strong and growing, but the swing from a NPR 2.57 billion loss to a projected NPR 1 billion profit rests heavily on a one-time rate renegotiation and a lucky hydrology year rather than a structural improvement in the underlying unit economics). Financial Strength totals 5 out of 20.

Valuation Reasonableness (out of 15). At a recent price near NPR 185, Upper Tamakoshi trades at a trailing price-to-earnings ratio of roughly 43 times — more than double the hydropower sector's average of about 18 times, according to recent sector-wide NEPSE valuation data — despite a company that has just emerged from four consecutive loss-making years. Its price-to-book ratio of roughly 4.25 times is being applied to a book value per share, near NPR 44, that has itself already been eroded by those years of accumulated losses (book value per share was above NPR 56 before the loss years took their toll). Paying a premium multiple over an already-shrunken equity base, for a single quarter of profit, is not a conservative valuation. We score the P/E sub-component 2 out of 8 and the P/B sub-component 2 out of 7, for a Valuation total of 4 out of 15.

Growth Trajectory (out of 15). It is important to separate the sector-wide demand story from this company-specific dimension. Nepal's electricity demand genuinely has grown steadily, and Upper Tamakoshi's own revenue grew a healthy 19.4 percent in its most recent reported fiscal year — that part of Suman's original instinct was sound. But Chapter 64's Growth Trajectory dimension asks about the company's own demonstrated trajectory of revenue and earnings together, not the macro backdrop alone, and a single fixed-capacity plant with no announced expansion, whose earnings only turned positive after a debt renegotiation rather than organic growth, does not yet show a durable growth trajectory. We score this 5 out of 15 — real revenue growth, but earnings growth that is still unproven.

Dividend & Capital Return Discipline (out of 10). This is the starkest number in the whole score. Since Upper Tamakoshi's shares were opened to public investors, the company has never declared a dividend — its public trading record shows no dividend distribution at all. As of August 2026 that remains true: NEPSE data providers return an empty dividend history for UPPER, and the only capital action since listing has been the 1:1 rights issue of FY2080/81 — a call for more shareholder money rather than a return of it. A company financed for years on debt that outgrew its budget, that spent four straight years in the red, and that only recently returned to modest profitability, has simply had nothing to distribute. We score this 1 out of 10, allowing a single point for the credible near-term possibility of an eventual dividend now that the company projects a real profit, against a flat zero track record to date.

Hemisphere 2 comes to 5 + 4 + 5 + 1 = 15 out of 40.

Lesson 91.6 — The Full Worked Canon Score

Putting both hemispheres together gives us Upper Tamakoshi's real, current Canon Score.

Forty-five out of one hundred places Upper Tamakoshi in Chapter 64's Weak/Avoid band, well below the 55-point Adequate threshold. For a direct dimension-by-dimension contrast against a plant one twentieth its size, see Chapter 84, where Chilime scores 62 on the same rubric. Note that the governance override does not fire here — the Governance sub-score of 11 out of 15 clears the 5-point floor comfortably, because this is a story of expensive execution difficulty, not fraud or self-dealing, and that distinction matters. Compare this to Chapter 89's Karnali Development Bank, where the override did fire: two very different companies can land in the same Weak/Avoid band for very different reasons, and the Canon Score's dimension-by-dimension breakdown is what lets you tell those reasons apart rather than treating "Weak/Avoid" as a single undifferentiated verdict.

This is precisely why the Canon Score framework insists on scoring dimensions separately rather than blending them into a single fuzzy impression of "good company" or "bad company." Upper Tamakoshi's structural business — the PPA, the state sponsor, the size and tradability of the listing — remains genuinely solid, worth 30 of the available 40 points in Hemisphere 1. What the delay and its aftermath did was devastate the financial dimensions that depend on actual results: profitability, valuation support, demonstrated growth, and dividends. An investor who only tracked the demand story, as Suman originally did, would have missed the entire re-rating that a careful, dimension-by-dimension read would have caught.

Lesson 91.7 — Reassessing the Position Once the Delay Was Confirmed

Suppose Suman had bought his shares at the original IPO price, built on the original NPR 35 billion cost assumption and the original commissioning timeline. What should he have done once the earthquake struck in 2015, and then again once the 2016–17 flood damage was reported?

The discipline the Canon has taught in earlier chapters applies directly here: a confirmed delay is not a reason to panic-sell, nor a reason to shrug and do nothing. It is a trigger to re-run the model with updated assumptions and make a fresh decision, exactly as you would if a company you held announced a major factory fire or a regulatory tariff change.

The re-underwriting process has three steps. First, update the cost assumption using the best available public information — in this case, contemporaneous reporting showing cost escalating from NPR 49 billion excluding capitalised interest in 2019 (already NPR 69 billion including it) toward an eventual reconciled total near NPR 89 billion. Second, update the timeline assumption to the newly indicated commissioning date, which pushed out year after year as further monsoon damage was reported. Third, and most importantly, re-run the discounted cash flow — the technique from Chapter 44 for converting a future stream of expected dividends or earnings into a single present value using a discount rate that reflects the riskiness of the cash flows — using the new cost base, the new timeline, and, crucially, a higher discount rate than originally used, because a project that has already slipped once has demonstrated it is exposed to delay risk and may slip again. In Chapter 64 terms, this re-underwriting is really an early warning that Financial Strength and Valuation are about to deteriorate sharply, even while Governance, Durability, and Liquidity stay comparatively steady — exactly the split Lessons 91.4 and 91.5 quantified with real numbers.

A discount rate is simply the rate at which you shrink future money to reflect the fact that money later is worth less than money now — both because of the time value of money and because of the uncertainty attached to actually receiving it. A project that has already proven vulnerable to earthquake and flood damage deserves a somewhat higher discount rate than one with a clean, on-schedule track record, because the probability distribution of future cash flows has genuinely widened. Every additional year of delay pushes dividends further into the future, and the discounting mechanism means a dividend pushed back five years is worth meaningfully less today than the same dividend arriving on the original schedule — even before you touch the cost-overrun assumption at all.

Running this re-underwriting honestly, most careful analysts following Upper Tamakoshi through 2016 and 2017 would have concluded that the project's per-share intrinsic value estimate needed to come down from the original IPO-era assumption, even though the plant's physical output potential and its favourable demand backdrop were unchanged. This is not a reason to conclude the investment was a mistake — it is a reason to conclude that the price you would be willing to pay for additional shares, or the conviction with which you hold existing ones, needed to be recalibrated to the new facts.

This is also the moment to distinguish between a temporary paper loss and a permanent loss of capital. If Suman's shares fell in price on delay news, the question he needed to ask was not "has the price gone down," but "has my re-underwritten estimate of fair value gone down by more or less than the price has." If the market overreacted — pricing in a worse delay or a larger overrun than was actually likely — the position may have become more attractive at the lower price, not less. If the market underreacted, and the true re-underwritten value had fallen further than the price had, the position may have still been overvalued even after the drop. Only a fresh model run, not the direction of the price chart, can answer that question.

For a company like Upper Tamakoshi specifically, one further consideration mattered: its sponsor. NEA is the national utility, backed ultimately by the state's interest in seeing the country's flagship domestically financed hydropower project succeed. Projects with a strong, well-capitalised sponsor behind them are less likely to be abandoned mid-construction even after a severe cost overrun, because the sponsor has both the financial capacity and the strategic motivation to see the project through. A similarly sized delay and overrun at a smaller, thinly capitalised private developer with weaker banking relationships could plausibly have resulted in a stalled or even abandoned project, a much worse outcome for shareholders than "merely" a multi-year delay. Sponsor quality is therefore a legitimate input into how much benefit of the doubt a delayed project deserves — and it is why Upper Tamakoshi's Governance sub-score in Lesson 91.4 held up at 11 out of 15 even as its Financial Strength sub-score, in Lesson 91.5, collapsed to 5 out of 20.

Lesson 91.8 — What Actually Happened, and the Portfolio-Level Lesson

Upper Tamakoshi was eventually commissioned, with its units connected to the national grid in July 2021 — roughly five to six years after the original mid-decade target, and at a total cost that would eventually reconcile near NPR 89 billion, roughly two and a half times the original approved budget. But commissioning was not the happy ending Suman's original thesis had imagined. The plant began generating substantial revenue — NPR 6.92 billion in fiscal year 2024/25 alone, still growing — yet the company posted a net loss in each of its first four fiscal years of operation, driven by interest on the enlarged debt load and by hydrology penalties NEA charged when generation fell short of contracted design energy in a given season, together totalling close to NPR 1 billion over those four years. Book value per share fell from above NPR 56 to the mid-NPR 40s as those losses ate into shareholders' equity. No dividend was ever declared. Only in the most recent fiscal year, after management renegotiated its long-term loan rate down from 8.25 percent to 6.75 percent and the plant finally had a penalty-free year, did the company turn a projected profit of roughly NPR 1 billion — a genuine but fragile turnaround, arriving a full four years after commissioning, not at commissioning.

This is the sharpest correction this case study makes to the story Suman told himself in 2016: a construction delay is not the whole risk. The favourable demand backdrop he correctly identified was real, and the PPA did guarantee a buyer at a contracted tariff — but a project financed with debt that outgrew its budget can keep losing money for years after the turbines start turning, simply because the capital base behind those turbines has to be serviced before a single rupee reaches shareholders. Being broadly right about Nepal's electricity demand was not the same as being right about when, or whether, this particular capital structure would translate that demand into a return.

The final lesson of this case study is not "hydropower is a bad investment" — Suman would be the first to tell you that would be an overcorrection. Nepal's growth story and hydropower are deeply intertwined, and a well-run project with a solid PPA and a disciplined capital structure remains one of the more durable long-duration income ideas available on NEPSE. The lesson is narrower and more useful than that: construction-phase risk in hydropower is real, common, and mathematically compounding, and where a project's debt burden grew large enough during that construction phase, the pain can keep compounding well into the operating years too — a distinct and separate risk from the delay itself, and one this Canon Score's Financial Strength and Valuation dimensions are built specifically to catch.

Practically, this suggests a portfolio-construction habit worth adopting directly. Rather than concentrating hydropower exposure in a single under-construction project — however flagship, however patriotic the appeal — a disciplined investor following this Canon should think in terms of "vintages": holding a mix of already-operating hydropower companies, whose revenue and dividend patterns are known and can be modelled with much less uncertainty, alongside a smaller, deliberately sized allocation to under-construction projects where the higher potential return compensates for construction-phase risk that even the best due diligence cannot fully eliminate. An earthquake in Dolakha, a landslide in the Tamakoshi valley, a contractor dispute, a delayed transmission line elsewhere in the country — these are not signs that something unusual went wrong with any single company. They are recurring features of hydropower construction in Himalayan terrain, and a well-constructed portfolio should be sized so that no single delayed project can derail the investor's overall plan.

Suman, in the end, kept his Upper Tamakoshi shares through the entire delay period, having re-underwritten his position twice — once after the earthquake, once after the flood damage was confirmed — and having concluded both times that the demand-side thesis remained intact even as the financial dimensions had visibly weakened. He did not add significantly to his position during the delay years, judging that the compounding IDC and uncertain timeline made the risk-reward less attractive than simply waiting. When commissioning finally arrived in 2021, he resisted the temptation to treat it as the finish line, and re-ran his numbers a third time — this time using actual post-commissioning financial statements rather than a construction-era forecast. That third look told him the position still did not clear his bar: four years of real losses, an eroding book value, and a Canon Score stuck in the Weak/Avoid band were not a reason to sell a company he still believed in structurally, but they were a reason to stop adding and simply wait for the financial dimensions to catch up with the demand story — which, only very recently, they finally began to do. He considers the outcome a genuinely open question rather than a settled success: a reminder that being broadly right about a country's electricity demand is not the same as being right about exactly when, and at what cost, any single project will deliver on that demand.