Understanding Project Finance

First published 23 Aug 2026 · Last verified 29 Aug 2026

Lesson 42.1 — What Project Finance Actually Means

Picture a Kathmandu family that owns a house they live in, plus a small flat they bought purely to rent out. If the family wants to buy a second rental flat, they have two ways to borrow the money.

The first way is to walk into a bank and borrow against everything they own — their family home, their savings, their gold, their existing flat — pledging the whole household balance sheet as security. If the new flat turns out to be a bad investment and the rent doesn't cover the loan, the bank can still come after the family home. This is corporate finance, sometimes called balance-sheet lending: the borrower's entire asset base and income stream stand behind the loan, and the lender's recovery is not limited to the thing being financed.

The second way is fundamentally different. The family sets up a separate small entity — perhaps a private company — that owns only the new rental flat. That entity borrows money in its own name, using only the flat itself and the rent it will generate as security. If the flat fails to earn enough rent to service the debt, the bank's recourse stops at the flat. The family home, the parents' pension, the children's education fund — none of it is touched. The lender agreed, from day one, to be paid back only from what this one asset produces.

This second structure is project finance. The formal definition: project finance is the financing of a long-term infrastructure or industrial project based primarily on the projected cash flows of the project itself, rather than the balance sheets of its sponsors, with the project's assets, rights, and revenue-generating contracts serving as collateral. The entity that owns the project — the small company in our analogy — is called a Special Purpose Vehicle, or SPV. An SPV is a legally separate company created for one purpose only: to build, own, and operate a single project (or a small family of related projects) and nothing else. It has its own board, its own bank accounts, its own contracts, and critically, its own limited liability — its creditors cannot reach through it to the pockets of the shareholders who set it up.

Why would a lender ever agree to be paid back only from one asset's earnings, instead of demanding a claim on everything the borrower owns? Because in exchange, the lender gets something a corporate loan rarely offers: forensic-level control over the asset itself. In a project financing, the lender does not simply hand over money and hope for the best. The lender's lawyers and engineers pick apart every contract the project depends on — the construction contract, the agreement to sell electricity, the insurance policies, the land lease — before a single rupee is disbursed. The lender then wraps the project in a security package (a set of legal rights over the project's assets, contracts, and cash flows that the lender can enforce if things go wrong) so tight that, if the project stumbles, the lender can step in and run it, sell it, or hand it to someone else who can make it pay, all without ever touching the sponsor's other assets.

This is precisely why project finance is the dominant financing method for hydropower in Nepal — and around the world for large, capital-intensive infrastructure. A single 100 MW hydropower project can cost tens of billions of rupees to build. No promoter group, however wealthy, wants to stake its entire family or corporate fortune on one river's flow, one government's tariff policy, and one power purchaser's creditworthiness. And no bank, in turn, wants to lend that much money against a promoter's general promise to repay — it wants a contractual claim on the one thing that will actually generate the cash: the power plant and the revenue stream from selling its electricity.

Lesson 42.2 — The Anatomy of a Nepali Hydropower Deal

To understand how a Nepali hydropower financing is actually built, it helps to walk through the life of a project in order, because each stage adds a layer of structure that the next stage depends on.

It begins with the SPV. A group of promoters — sometimes a mix of local entrepreneurs, a construction house, and occasionally a foreign strategic investor — incorporates a company under Nepal's Companies Act, whose sole object clause is to develop, own, and operate one specific hydropower project on one specific river reach. This company then applies to the Department of Electricity Development (DoED) for a survey license and subsequently a generation license under the Electricity Act. Everything that follows — every contract, every loan, every regulatory approval — attaches to this one company.

The SPV then assembles what project finance practitioners call the "project contracts" — the web of agreements that together convert a hole in a riverbed into a bankable business:

First, the Power Purchase Agreement (PPA), a long-term contract (typically 20-35 years in Nepal, sometimes with a build-own-operate-transfer or "BOOT" tenor) under which Nepal Electricity Authority (NEA), the state-owned utility, agrees to buy all the electricity the project generates at a pre-agreed tariff structure (commonly a two-season, two-time-of-day rate reflecting Nepal's wet-season surplus and dry-season shortage). The PPA is the single most important document in the entire financing, because it is the contract that turns falling water into a predictable cash flow banks can lend against.

Second, the Engineering, Procurement and Construction (EPC) contract, under which a contractor — often a Chinese, Indian, or Nepali construction house — agrees to build the powerhouse, the dam or diversion weir, the headrace tunnel or canal, and the penstock for a fixed price and a fixed completion date, usually with liquidated damages (pre-agreed cash penalties) if it is late.

Third, land acquisition and government agreements — a survey license, a generation license, and often a Project Development Agreement or similar instrument with the Government of Nepal covering royalty payments, local benefit-sharing, and, for larger projects, tax and forex arrangements.

Fourth, insurance contracts covering construction-period risks (fire, flood, contractor default) and operational risks (business interruption, machinery breakdown), assigned to the lenders as part of their security.

Once these contracts exist in more or less final form, the SPV goes to the banks to arrange debt. This is where the financing structure — the actual split between how much of the project cost is borrowed versus contributed by shareholders — gets fixed.

Why do lenders tolerate such high leverage (a high proportion of debt relative to equity) for hydropower specifically? Because a well-structured, operating hydropower plant with a PPA in place has an unusually predictable and long-lived revenue stream — water keeps flowing, NEA keeps needing power, and the tariff is fixed by contract for decades. That predictability is what allows debt to be layered so heavily on top of a modest equity base — the opposite of, say, financing a restaurant or a trading business, where revenues are volatile and banks would rarely go beyond 50:50.

The table below summarises how a typical Nepali run-of-river hydropower project financing is commonly structured today, drawing on the terms disclosed in recent IPO prospectuses and credit rating rationales of NEPSE-listed hydropower companies.

Lesson 42.3 — Financing the Deal: Syndication and the Single-Obligor Constraint

Here is a number worth sitting with: as of recent industry estimates, the combined core capital of all commercial banks in Nepal is only around USD 900 million. Core capital is a bank's own equity cushion — its paid-up capital plus reserves — and it is the base figure regulators use to size how much any one bank can safely lend to any one borrower. Meanwhile, Nepal has issued generation licenses for well over 1,800 MW of hydropower capacity. A single large storage or peaking-run-of-river project — something in the 400-900 MW range — can cost well over NPR 100 billion. No single Nepali bank, and often not even a handful of them together, has a balance sheet anywhere near large enough to write that check alone.

This mismatch is precisely why syndication is not an optional financing technique in Nepal's hydropower sector — it is a structural necessity. Syndicated lending means multiple banks jointly provide a single loan to one borrower, sharing the loan amount, the collateral, and the risk in agreed proportions, under one common set of loan documents. One bank (or a small group) acts as the lead arranger, structuring the deal, negotiating terms with the SPV, and then inviting other banks to "participate" — to take a slice of the total loan on the same terms. A common form used in Nepal is the consortium financing arrangement, where several banks lend directly and severally to the borrower (each bank has its own direct claim on the SPV for its share), as opposed to a syndication proper where one bank lends and then sells participations — Nepali practice leans toward the consortium model, with a lead bank coordinating security documentation and cash-flow monitoring on behalf of all participants.

This regulatory picture changed materially in 2025 (fiscal year 2082/83 in the Nepali calendar), when NRB issued a directive removing the fixed Single Obligor Limit ceiling altogether, shifting responsibility for setting individual borrower exposure limits to each bank's own internal credit risk framework and board-approved policies, rather than a hard regulatory number. The stated rationale was to unlock financing for "mega projects" — large hydropower, transmission, and infrastructure schemes — that the old fixed ceiling was actively constraining, while leaving banks accountable to their own risk governance and to NRB's broader prudential supervision (capital adequacy, sector concentration limits, and loan classification rules remain firmly in place).

Even with the single-obligor ceiling gone, syndication remains the practical norm for genuinely large hydropower projects, for a simple reason that has nothing to do with regulation: risk diversification. No single bank's management or board wants 100% of a NPR 50-100 billion exposure sitting against one river, one EPC contractor, and one power purchaser, however creditworthy. Spreading the loan across six, eight, or a dozen banks means that if the project underperforms, no single institution's solvency is threatened, and the workout (the process of restructuring a troubled loan) becomes a shared, negotiated exercise among a consortium rather than a life-or-death event for one bank.

Two structural features distinguish Nepali hydropower syndication from a typical corporate syndicated loan. First, the lenders are almost always exclusively Nepali commercial banks and, increasingly, provident and pension funds such as the Employees Provident Fund (EPF) and Citizen Investment Trust (CIT), rather than international commercial banks — because NEA's PPA is denominated in Nepali Rupees, and foreign lenders are generally unwilling to take open Nepali Rupee currency risk on a 15-year loan without a hedge that simply does not exist in Nepal's shallow currency markets. Second, for genuinely large national-priority projects, the Government of Nepal and multilateral development banks (the World Bank, the Asian Development Bank, and sometimes bilateral lenders like India's Exim Bank or Indian public-sector lenders) step in as either direct project lenders or as guarantors/co-financiers alongside the domestic banking syndicate, precisely because the domestic system alone cannot absorb the exposure.

Lesson 42.4 — The Lender's Security Package

Once a syndicate of banks agrees to lend, the next question is: what exactly do they hold as collateral if the SPV cannot repay? In ordinary corporate lending the answer is often "a mortgage on the borrower's land and buildings, plus personal guarantees." In project finance the answer is far more elaborate, because the lenders are relying almost entirely on the project's own future cash flows rather than the sponsor's balance sheet — so they build layers of legal control designed to let them step into the project's shoes if things go wrong, well before the SPV is formally declared insolvent.

The core components of a Nepali hydropower lender's security package are:

Assignment of the PPA. Because NEA's payment obligation under the Power Purchase Agreement is the project's only real source of revenue, lenders require the SPV to assign its rights under the PPA to the lenders (or to a security agent acting on the syndicate's behalf) as collateral. This does not mean lenders take over selling electricity day to day — it means that if the SPV defaults on its loans, the lenders acquire the legal right to step into the PPA, receive the payments directly, or even take over operation of the plant (through a "step-in right") to keep the PPA alive rather than let NEA terminate it. NEA's consent to this assignment, formally recorded, is itself a standard closing condition for the loan.

Escrow account arrangements. An escrow account is a special bank account, typically held with the lead lender or security agent bank, into which all of the project's revenue (NEA's PPA payments) is deposited directly, and out of which cash flows are released only in a pre-agreed order, called a "cash flow waterfall." A simplified Nepali hydropower waterfall typically runs: (1) operating expenses, (2) scheduled debt service (principal and interest), (3) top-up of the debt service reserve account if it has been drawn down, (4) other reserve accounts (major maintenance reserve, for instance), and only then (5) distributions to shareholders as dividends. The escrow mechanism means the SPV's management never has unrestricted access to the incoming cash — the bank sees every rupee arrive and controls the order in which it leaves, which is the single most powerful practical tool lenders have for making sure they get paid before shareholders do.

Debt Service Reserve Account (DSRA). A DSRA is a ring-fenced cash reserve, funded either upfront from the loan proceeds or built up gradually from early cash flows, sized to cover a defined number of months of future debt service (commonly three to six months of principal and interest). Its purpose is simple: if the plant has a bad month — a landslide damages the intake, the river runs unusually low, a transformer fails — the SPV can draw on the DSRA to keep making loan payments on schedule rather than immediately defaulting. Lenders require the DSRA to be replenished from subsequent cash flows before any dividends can be paid, which is exactly why it sits ahead of shareholder distributions in the escrow waterfall.

Beyond the PPA assignment, escrow, and DSRA, a full Nepali hydropower security package typically also includes: a first-ranking mortgage or charge over all project land, the powerhouse, penstock, and generating equipment; a pledge of the SPV's shares held by the promoter group (so that if the SPV defaults, lenders can effectively take control of the company by enforcing the share pledge rather than only chasing physical assets); assignment of all project insurance policies, so insurance payouts after a loss go first to repair the asset or to the lenders rather than disappearing into the sponsor's other businesses; and assignment of the EPC contract and any performance bonds/liquidated-damages rights against the contractor, so lenders can pursue the contractor directly if construction defects or delay cause loss.

Lesson 42.5 — Construction Risk vs Operational Risk: Why Financing Terms Change Over the Project's Life

Every project financing has a hinge point, and understanding it is essential to understanding why the loan terms, the security package, and the level of sponsor involvement all shift dramatically at one specific moment: the date the plant is actually finished, tested, and generating revenue, commonly called the Commercial Operation Date (COD).

Before COD, the project is in its construction phase, and the dominant risks are what project finance calls construction risk: will the EPC contractor finish on time and on budget? Will unforeseen geology in the headrace tunnel cause months of delay? Will a monsoon flood wash away a cofferdam? During this phase, the project generates zero revenue — there is nothing yet to sell to NEA — which means there is no cash flow for lenders to be repaid from, and no operating track record to judge. This is the riskiest period of the entire project's life, and lenders price and structure the loan accordingly.

After COD, the project enters its operational phase, where the dominant risks become operational risk: will the river's hydrology behave as the feasibility study predicted? Will the machinery run reliably? Will NEA pay on time? These risks are generally far more predictable and far better insured against, because by this point there is an actual track record of the plant generating power and receiving PPA payments.

In practice, lenders manage the gap between these two risk profiles through several specific tools written into the loan agreement:

Completion guarantees and cost-overrun undertakings. Sponsors are typically required to guarantee that they will fund any cost overrun out of their own pocket (rather than asking lenders for more money) up to a specified cap, and sometimes to guarantee the project's physical completion itself — meaning if the SPV cannot finish construction, the sponsor is on the hook to either fund completion or repay the loan. This is the "limited" part of limited-recourse financing: the recourse to the sponsor is limited in time (construction period only) and limited in scope (cost overruns and completion, not general project underperformance).

Contingent equity or standby letters of credit. Rather than requiring sponsors to inject additional cash immediately, many Nepali financings require sponsors to arrange a standby facility — a letter of credit or committed additional equity line — that only gets drawn if an actual cost overrun materialises, keeping the sponsor's capital productively deployed elsewhere until it's actually needed.

Independent engineer oversight. Lenders appoint their own independent engineer (a technical consultant paid by the borrower but reporting to the lenders) to monitor construction progress, certify that disbursement milestones have genuinely been met before releasing further loan tranches, and flag emerging delays or cost overruns early.

Disbursement conditions tied to physical progress. Loan tranches are released against verified construction milestones (foundation complete, penstock installed, turbines delivered) rather than as a lump sum upfront, so lenders are never funding further ahead of the asset's actual physical progress than necessary.

Once COD is reached and the plant has demonstrated stable output over an agreed testing period, the loan typically "converts" into its long-term operational repayment schedule, and covenants shift toward ongoing financial monitoring — chiefly the Debt Service Coverage Ratio (DSCR), a ratio comparing the cash available for debt service in a period to the actual debt service due in that period. A DSCR of 1.3x means the project generated 30% more cash than it strictly needed to make that period's loan payment — the cushion lenders require as protection against a bad hydrology year. If DSCR falls below the covenanted minimum (commonly 1.1x-1.2x in Nepal), the loan agreement typically restricts or blocks dividend distributions to shareholders until the ratio recovers — another example of the escrow waterfall protecting lenders ahead of equity.

Lesson 42.6 — Case Studies: Learning from Nepal's Hydropower Financings

Nepal's own hydropower history offers a useful spread of examples across ownership structures, financing sources, and outcomes.

Upper Tamakoshi (456 MW) is Nepal's largest operating hydropower plant, commissioned in July 2021, and is a valuable case study precisely because it broke from the pattern of foreign-financed mega-projects. It was developed by Upper Tamakoshi Hydropower Limited (UTKHPL), an SPV established by NEA in 2007, and — notably — was financed entirely from domestic Nepali financial institutions and companies, without foreign commercial lenders. Its ownership structure blends public and community interests: NEA holds 41%, with Nepal Telecom, Citizen Investment Trust, and Rastriya Beema Sansthan (the state insurance corporation) holding smaller stakes, alongside a 10% quota reserved specifically for residents of Dolakha district (where the project is located), 15% for the general public, and the remainder for EPF contributors and NEA/company/financial-institution staff. This structure demonstrates how a genuinely large hydropower project can be financed domestically when a syndicate of Nepali banks, provident funds, and public shareholders is assembled at sufficient scale — and it stands as evidence against the assumption that only foreign capital can finance Nepal's largest hydropower assets.

Chilime Hydropower Company (22.1 MW), an earlier and much smaller NEA subsidiary project in Rasuwa district, is often cited as Nepal's original template for the "community-inclusive" project-company model — later becoming one of the first hydropower companies listed on NEPSE, with a portion of shares specifically reserved for local project-affected residents. Many subsequent Nepali hydropower IPOs have followed this local-quota-plus-public-tranche template, which has become close to a market norm.

Tamakoshi V, a follow-on project to Upper Tamakoshi, illustrates the more recent trend of pension and provident funds becoming direct project lenders rather than merely portfolio investors: NEA, the Employees Provident Fund, and the Tamakoshi hydropower company signed a tripartite loan agreement in 2023 to fund its development, alongside conventional bank debt.

Arun-3 (900 MW), by contrast, illustrates the opposite end of the spectrum: a cross-border, foreign-sponsor-led project developed by SJVN (a joint venture of the Government of India and Himachal Pradesh government), where the bulk of project debt — reported at roughly INR 6,333 crore — was arranged through Indian lenders rather than Nepali banks, reflecting the fact that a project of this scale, with a foreign sponsor exporting power to India, sits largely outside the domestic Nepali banking system's capacity and currency comfort zone. It is a useful counter-example showing that when a project's revenue and sponsor structure point outward (an export-oriented PPA, a foreign sponsor), the financing naturally follows that same external orientation.

Rasuwagadhi Hydropower Company, whose IPO grading was assessed by ICRA Nepal, is a useful illustration of how Nepal's rating agencies formally assess project bankability before an IPO — evaluating exactly the factors covered in this chapter: PPA quality and NEA's payment reliability, hydrology and technical risk, the strength (or weakness) of the debt-equity structure and security package, and construction-phase versus operational-phase risk profile. Reading an IPO grading rationale for any NEPSE-bound hydropower company is one of the most efficient ways for an investor to see project finance analysis applied in practice by professionals, rather than in the abstract.

A final, sobering thread runs through nearly all of Nepal's hydropower financing history and deserves an investor's close attention: bankability gaps that persist despite an otherwise sound legal and financing framework. Two issues recur repeatedly in professional legal and rating commentary. First, currency risk: NEA refuses to sign PPAs denominated in US dollars for domestic projects, and Nepal has no meaningful market for hedging long-dated Nepali Rupee exposure, which is precisely why foreign commercial lenders rarely participate directly in domestic-PPA hydropower financings, leaving the field to Nepali banks, provident funds, and multilateral development institutions willing to lend in or alongside Nepali Rupees. Second, PPA termination asymmetry: as covered in Lesson 42.4, domestic-investment PPAs typically leave a developer with no compensation and only wheeling rights if NEA terminates the agreement, while foreign-investment PPAs typically do include termination compensation — a structural inconsistency that Nepali legal commentators have flagged as an unresolved bankability weakness for purely domestic developers and their lenders.