- +977-1- 5706295 / +977-1-5706000 / +977-1-5706392
- 246 Sahayog Marg, Anamnagar, Kathmandu
This note sets out the policy changes Nepal needs in order to accelerate the development of its energy sector.
The note is arranged in five parts. The first describes Nepal’s present position, in terms of both installed generating capacity and the sources from which that capacity has been financed. The second examines the specific risks that decide whether a project reaches financial close, meaning the point at which lenders formally commit their funds, and for each of them identifies what the state can do to reduce it. The third looks at the institutions through which projects are approved, and at the role of the Investment Board Nepal (IBN) within them. The fourth sets out two financing instruments available to the state itself, as distinct from the protections sought by investors. The fifth asks how far the diagnosis and the measures proposed here apply beyond the energy sector.
Throughout, the note describes each problem as investors and lenders experience it, and then names the policy or legal change that would ease it.
How to read this note. A reader who wants the whole position in one view should begin with the annexure at the end. It sets out every issue in a single table: the problem, the measure proposed, and the instrument through which that measure would be delivered, whether the Project Development Agreement (PDA) policy, a change in statute, or administrative action. The sections carry the reasoning behind each entry, and can be read in any order or consulted individually.
Nepal’s energy ambitions are now clearly on record. The Energy Roadmap 2081 sets a generation target of 28,500 MW by 2035 and a per-capita consumption target of 1,500 kWh. The Energy Roadmap estimates that meeting both will require investment of approximately USD 58.9 billion across generation, transmission, and distribution.
The present position falls considerably short of these targets. Installed capacity is roughly 4,200 MW, about fifteen percent of the 2035 target. Per-capita consumption is 465 kWh, about thirty-one percent.
The more serious gap is not the size of the shortfall but the speed at which Nepal is closing it. Over the last three fiscal years the country has added about 500 MW of capacity a year. Reaching the 2035 target requires an average of about 2,700 MW a year, more than five times the rate achieved so far.
A gap of that size cannot be closed through incremental improvements to the existing system. Two things have to change: the amount of capital entering the sector, and the speed at which licensed projects become operating power plants.
The licensing data show where the blockage lies. Close to 11,000 MW of capacity holds a generation license but is not operating, and a further 13,000 MW sits under survey license. Yet only 4,303 MW of licensed but non-operational capacity has reached financial closure. Nepal is not short of projects. It is short of projects that lenders and investors will actually fund and will achieve commercial operation.
Nepal’s own financial system cannot fund a transition of this size. This is a matter of arithmetic, not of willingness.
Nepal Rastra Bank already requires commercial banks to lend a rising share of their loan portfolios to the energy sector, reaching 10 percent of total loans by mid-July 2027. The banks have met each interim target ahead of schedule. As of mid-January 2026, they stood at 8.44 percent, with an outstanding energy portfolio of NPR 472.08 billion. The sector is not short of commitment from domestic lenders. It is close to the limit of what that commitment can produce.
Even with the mandate fully met, the banking system’s lending capacity for the sector comes from two major sources. The first is the principal that projects in operation repay each year, which returns to the banks and is lent out again. The second is the gradual growth of the overall credit base, which has expanded at close to 6 percent a year. Recycled principal is the larger of the two, at roughly three-fifths of the total.
Together these allow the banking system to commit in the region of USD 600 million to the sector this year, rising to around USD 1 billion a year by 2035. Against a requirement of USD 58.9 billion, this falls substantially short. The balance must come from outside Nepal’s own financial system.
Energy, along with information technology, is one of the few sectors to have attracted sustained foreign interest. It accounts for a substantial share of the foreign investment commitments Nepal has realized.
A closer look at where that money came from changes the picture. Much of the equity has come from investors acting under a state-level mandate to put capital into Nepal. Much of the debt has come from development finance institutions lending to hydropower projects on terms, and with a tolerance for risk, that commercial lenders do not share.
Money that arrives because an investor is mandated to invest differs fundamentally from money that arrives because an investor expects a good risk-adjusted return.
The difference can be tested with one question, and it is a question policy makers should put to themselves before designing any incentive. Suppose an investor holds one billion dollars. It can place that money in any sector, in any country. It answers only for the return the money earns. Would it choose Nepal’s energy sector?
The value of the question lies in what it strips away. It sets aside concessional funding, climate mandates, bilateral relationships, and goodwill, and asks whether the sector can compete on commercial merit against every other option available. Wherever the honest answer comes with a qualification, that qualification points to one of the reforms this note addresses.
Foreign investors in the energy sector fall into two broad groups.
The first invests for return. It places capital wherever the risk-adjusted yield is most attractive. The second invests for an agenda, whether that is a climate finance objective, a bilateral or state investment mandate, or a development impact target. This second group will accept returns that a purely commercial investor would decline.
The two behave differently in most respects, but they are identical in one. Both begin by reading Nepal’s risk profile, and both put a price on what they find.
That profile is public and not open to argument. Fitch rates Nepal’s sovereign credit at BB- with a stable outlook, which places the country in speculative grade. The OECD places Nepal in country risk Category 7, its highest band. FATF grey-listing raises the compliance cost of every USD-denominated financing. World Bank government effectiveness scores and political fragility indices point the same way. None of these indicators is discretionary, and none can be negotiated away in a single transaction.
This leads to the organizing premise of this note: every risk is priced.
Any risk the state declines to address does not disappear. It comes back as a higher tariff, a wider debt margin, a higher required return on equity, a larger insurance premium, or a decision not to invest at all. Uncertainty over approval timelines is priced. So is uncertainty over land acquisition, over who will buy the power, over whether profits can be taken out of the country, over which law governs the concession, and over whether a negotiated position will survive a change of government.
Each of these carries a cost. That cost is paid in the end either by Nepali consumers through their tariffs, or the project itself can become financially unfeasible.
The state’s central task in mobilizing private capital is therefore not offering financial incentives, it is removing uncertainty. In practice that requires three things: setting out a defined, step-by-step path a project must follow, guaranteed timeframes for each step, and giving investors solid grounds to expect that the revenue, or the mandate objective, they are underwriting will in fact be delivered and realized.
The question for a policy maker is narrow and practical. For each risk identified, what can the state do to reduce or remove it? The sections that follow set out the main risks in the sector and the measures available to address them.
A Project Development Agreement (PDA) is a contract between the state and the project company. It governs how a project is developed, built, and operated, and it allocates the risk between the two parties.
The PDA is the instrument through which the state gives a financier the assurances that generally applicable law cannot provide. These typically cover the facilities and concessions extended to the project, protection against political and regulatory change, the state’s own obligations on approvals and supporting infrastructure, the treatment of force majeure and termination, the arrangements for repatriation of profits and for taxation, and the governing law and dispute resolution mechanism. For an asset that will operate for decades in a country with Nepal’s risk profile, the PDA is frequently the single document on which the financing decision turns.
Because the PDA is referred to repeatedly in the sections that follow, one point about its present treatment should be made at the outset. Nepal negotiates PDAs project by project, with no published policy framework setting out what the state is prepared to offer. An investor assessing Nepal therefore has no visibility on the terms it will ultimately secure.
The result is that questions which could be settled once, at policy level, are instead re-argued in every transaction. Negotiations are prolonged, comparable projects end up with inconsistent terms, and investors are denied the predictability that a published PDA policy would supply.
Several of the measures set out in this note involve commitments the state may reasonably wish to test before adopting them permanently. A sunset provision is a useful instrument in these circumstances. Under it, a policy or facility is enacted for a defined period and lapses automatically unless it is deliberately renewed.
This allows the government to back a facility firmly enough for investors to rely on it, while keeping a fixed point at which its cost to the budget, its effect on investment flows, and its continued necessity can all be reviewed, and the policy revised or discontinued. The arrangement serves both sides: the state can commit without committing permanently, and investors get a defined and transparent basis on which to plan.
The sections that follow deal, in turn, with each of the principal risks that stakeholders encounter in the energy sector. In each case the section sets out the constraint as investors and lenders encounter it, the current legal and policy position, the cost the uncertainty carries, and the measure proposed.
Land acquisition is the single most persistent constraint on project delivery in Nepal’s energy sector. Its effects are twofold: it extends the project timeline, and it raises project cost.
As Nepal diversifies its generation mix beyond its near-exclusive reliance on run-of-river hydropower, the land intensity of the project pipeline will increase. Solar generation, in particular, requires substantially more land per megawatt than hydropower. For solar projects, land is therefore not merely one constraint among many; it is the principal constraint.
The pressure will be greater still in transmission. The Nepal Electricity Authority (NEA) has now begun inviting private participation in transmission on a build, own, operate and transfer basis. This opens up a part of the sector that has until now been built by the state alone, and it brings private capital into a form of project whose land requirement is different in kind.
The transmission line is not a single site but a corridor. It crosses many separate holdings, often across several districts and several local governments, and the line cannot be energized until every stretch of it is settled. A single unresolved parcel can hold up the whole asset. For a private developer financing that line, the land risk is therefore not one risk among several. It is the risk that determines whether the project can be built at all.
A land acquisition framework that already struggles to support a hydropower-dominated pipeline will face considerably greater pressure as the energy mix broadens and as private capital enters transmission.
The acquisition of land, including compulsory acquisition, is a necessary condition of energy sector development. That should be said plainly at the outset.
The power to take private property for a public purpose against compensation is expressly recognized in the constitutional order. Article 25(2) of the Constitution permits the State to requisite or acquire property where this is done in the public interest. Article 25(3) provides that the basis of compensation and the procedure to be followed shall be as provided by law.
Section 33 of the Electricity Act 2049 gives this effect in the energy sector. It enables a licensee to apply to the Government where the land or house of any person is required for generation, transmission, or distribution. It also provides that government-owned land shall be made available on lease for a period up to the term of the license.
The question before policy makers is therefore not whether the State should hold this power. It already holds it. The question is whether the power is being exercised as its purpose demands. On the present evidence, it is not.
Project land falls into two categories, each requiring a distinct remedy:
Three clarifications of scope are necessary. First, the lease of private land for temporary use during construction has not caused difficulty in practice and is not addressed here. Second, forest land is itself public land, but the approvals attaching to it raise a separate set of issues and are dealt with in section 3 below. Third, the right of way a transmission line requires over land the project neither owns nor occupies is a distinct question, and is addressed in section 11.
The difficulty with public land is not entitlement but process. The legal basis for a lease is settled. The time required to obtain one is not; at present it is at least a year.
The delay comes from the number of authorities that must act before a lease can be issued. An application will typically pass through the Department of Electricity Development, the Land Revenue Office, the Department of Land Management and Archive, and the Ministry of Land Management, Cooperatives and Poverty Alleviation. A decision of the Council of Ministers is then required to conclude the matter. Each step has its own queue, its own documentary requirements, and its own indeterminate duration. The steps run one after another rather than in parallel.
The involvement of the Council of Ministers deserves particular attention. The lease of government land to a licensed energy project is, in substance, an administrative decision against settled criteria. Sending it to the apex of the executive treats it as a matter of national policy. The result is not better scrutiny but slower scrutiny.
The measure. The regulatory process for leasing public land to licensed energy projects should be re-examined with the explicit aim of shortening it. At most, the decision should rest with the relevant ministry. Requiring a Council of Ministers decision for a matter of this kind imposes a burden out of all proportion to what is being decided.
A standing committee would be a workable alternative. It would bring together the authorities whose input is genuinely required, and would be empowered to issue the approval on their collective behalf. That converts a sequential chain into a single sitting.
The outcome sought is straightforward: a quick and efficient route by which a licensed project company can obtain the public land it requires on lease.
Private land is a harder problem. Both available routes are defective, and the defects of each push project companies toward the other.
The institutional problem. Compulsory acquisition is implemented through the District Administration Office, with the Chief District Officer chairing the Compensation Determination Committee constituted under Section 13 of the Land Acquisition Act 2034.
The difficulty is that the same office already carries a substantial set of other functions. These include law and order and district security coordination, the issuance of citizenship and identity documents, licensing and registration, disaster response, and quasi-judicial adjudication under a range of statutes. Most are unavoidable and many are time-bound.
Land acquisition for an energy project competes for attention against all of them. The problem is not that the office is indifferent. It is that the office has no capacity to spare.
The compensation problem. Compensation determined through the statutory mechanism has consistently been set well below what landowners regard as fair. This is the origin of much of the disturbance seen at project sites, where local opposition is often less about the project itself than about the terms on which land is being taken. A process that is both slow and seen as unjust produces resistance, and resistance produces further delay.
The litigation problem. Compulsory acquisition is frequently challenged in court, and the courts have readily granted injunction orders, halting the process until the dispute is settled. A landowner who disputes the compensation has no other forum in which to be heard, and an injunction is the only protection available.
The difficulty is what follows. Case backlogs in Nepal are long, and a land compensation dispute waits its turn alongside every other kind of case. A project can be stopped for years by a challenge to the amount payable on a single parcel.
For a lender this is the most damaging feature of the statutory route. It turns a known cost into a delay of unknown length, and a delay of unknown length is not something a financing model can absorb.
The consequence. Faced with a statutory route that is slow and contested, project companies have increasingly preferred negotiated purchase. This substitutes one problem for another.
Once a project is announced, or becomes known by any other means, land prices in the project area do not rise gradually. They move to multiples of their pre-project level, with increases of seven to ten times observed in practice. The developer is then paying a price that reflects not the land’s value but the project’s existence, and that premium enters project cost and ultimately the tariff.
In short, the statutory route is too slow and too contentious to rely on, and the negotiated route exposes the project to speculative pricing. Neither outcome is acceptable, and neither can be fixed in the Project Development Agreement. These are constraints of general law and general administration, and they require intervention at the policy level.
The objective is to strike a balance. Project costs should not increase merely because a project has been announced, while landowners should receive genuinely fair compensation. This is what the Constitution contemplates, and perceived unfairness in compensation is the main cause of local resistance and resulting project delays.
These two objectives are usually presented as being in tension. They are better understood as depending on the same thing: a credible mechanism for determining compensation.
Differentiate between categories of affected landowner. The present mechanism treats all landowners the same way, which produces outcomes that are formally equal and substantively unequal.
A landowner with a house and a small adjoining plot does not experience the acquisition in the way the holder of a larger parcel in the same area does. Some landowners must relocate and others need not. Some run a business on the land, and the cost of re-establishing that business, together with the revenue lost in the meantime, is a real loss that a per-unit land rate does not capture.
A framework able to recognize these distinctions would produce fairer outcomes without necessarily raising total cost, because it would direct compensation toward those who actually bear the heaviest burden.
Fix the valuation reference point before the announcement effect takes hold. Compensation can be determined by reference to market prices prevailing one year before the acquisition, adjusted by a market-based index of land price inflation.
This preserves the landowner’s entitlement to the real value of the land, including its natural appreciation, while removing the speculative increase that arises solely from the project’s announcement. It is also workable in practice, because it relies on transaction records that already exist rather than on a contested valuation carried out under pressure.
Reconsider which authority determines compensation. Experience to date indicates that the District Administration Office is not the right body to lead compulsory acquisition. That does not mean it should be removed from the process. The office retains a legitimate and necessary role in implementing acquisition on the ground, where its district presence and its law and order mandate are directly relevant.
What should be reconsidered is where the determination itself sits. Either an existing authority should be empowered to lead this function, or a dedicated authority should be established, at central level or on a decentralized basis, with the technical capacity and the single focus the task requires.
Establish a dedicated forum for land acquisition disputes. A judicial or quasi-judicial body should be constituted to hear disputes arising from compulsory acquisition and nothing else. A specialist forum hearing one class of case develops consistent standards, and it does not compete for time against every other kind of litigation. It gives the landowner a real hearing, and it gives the developer a timeline.
Separate payment and possession from the dispute. Once compensation has been determined, the project company should be required to pay the determined amount to the landowner at once, and the acquisition should then proceed within a prescribed timeline. A landowner who disputes the amount takes that dispute to the designated body, where it runs on its own track.
The purpose of this sequence is that neither side waits for the other. The landowner receives, at the outset, an amount sufficient to relocate and re-establish, rather than waiting years for payment while the project stands idle. The developer keeps a construction schedule a lender can rely on. What remains in dispute is the difference between the amount paid and the amount claimed, which is a question about money and can be settled as one.
Re-imagine the form that compensation takes. Compensation need not be exclusively monetary. Equity participation gives landowners a share in the project’s upside rather than a single payment at the point of loss, and it converts a party with a grievance into a party with an interest in the project’s success.
A mechanism of this kind already exists. Hydropower projects are required to offer 10 percent of their shares to the Project Affected Population. Its weakness is in targeting: the Project Affected Population is identified through the environmental impact assessment, and the resulting net is drawn very wide, so the allocation is spread across a population considerably larger than those whose land the project actually takes.
Re-weighting that allocation to give clear priority to landowners whose land is used for the project would direct the benefit closer to where the burden falls, give those landowners a stake in the outcome, and reduce friction at the project site.
Taken together, these measures are directed at two outcomes. The first is that compulsory acquisition should become certain and efficient, so that a developer can plan around it and a landowner knows what to expect. The second is that compensation should be fair to the landowner while artificial price escalation is prevented.
A third benefit follows from the first two and is easily overlooked. Where a credible statutory route exists, negotiated acquisition becomes easier rather than harder, because a landowner who knows what the statutory process would yield has a reference point against which to negotiate, and no reason to hold out for a multiple of it.
A project passes through a long sequence of government approvals before construction can begin, and through further approvals during construction and operation. The difficulty is not that approvals are required. Every country requires them. The difficulty is that a developer cannot tell in advance how long they will take, or what answer they will produce.
These are two separate problems, and they need separate remedies.
The approvals a project needs are spread across many agencies and across all three levels of government. Each has its own queue, its own documentary requirements, and its own internal sequence.
Few of these approvals carry a binding timeline. Where a timeline does exist, missing it usually carries no consequence for the office concerned, and the clock commonly restarts whenever further documents are called for. A developer can therefore list the steps a project must pass through, but cannot put dates against them.
That gap between a known sequence and an unknown schedule is what reaches the lender. Interest accrues during construction against a commercial operation date nobody is able to commit to, and the sponsor carries the cost of every day the date moves.
The second problem is interpretation. The same statutory provision is read differently by different offices, and differently by the same office at different times. Positions on the legal provisions are often given orally, are not recorded, and do not bind the officer who inherits the file.
This is best understood as a response to incentives rather than as an attitude. An official who grants an approval that is later questioned carries a personal exposure, whether through audit observation, departmental action, or investigation. An official who refuses, defers, or refers the matter upward carries none. Where a provision is capable of more than one reading, the reading that is safest for the officer is the narrow one, and the safest course of all is to refer the file elsewhere.
The effect is that a provision means what the office applying it says it means on the day, and there is no way to establish that in advance.
An investor who cannot date the approvals builds contingency into the construction schedule and into the budget, and both raises the tariff the project must earn.
An investor who cannot rely on how a provision will be read discounts the protection that provision appears to offer. The more conservative the reading a sponsor has to assume, the higher the return it must require before committing.
This problem has to be tackled on two tracks at once. The PDA, backed by a published PDA policy, can protect an individual project against the consequences of delay. Structural change to the approval regime is what reduces the delay itself. Neither track is sufficient on its own.
The PDA should list the major approvals a project needs before and during construction. These would include environmental study approval, tree cutting approval, foreign investment approval, work visas for foreign personnel, and foreign exchange approval for payments to foreign contractors, among other major required approvals.
Against that list, the government can take on an obligation to compensate the developer for delay in approval(s) arising from the government’s end. Compensation here does not mean cash. It means time: an extension of the project’s key deadlines, and a matching extension of the revenue term, so that a delay on the government side does not shorten the period over which the project earns.
An obligation of this kind has to be bounded, or no government could responsibly give it. Three limits make it workable.
First, the obligation attaches only to the approvals specifically named in the PDA. It does not extend to every dealing the project has with the state.
Second, the clock starts only once a complete application has been submitted with all required documents. A developer cannot start the clock by filing an incomplete application.
Third, the authority should have a short window, for example ten days, in which to examine the application and call for anything missing. The clock then runs from the point at which the application is complete.
Structured this way, the obligation falls on the government only where the government is in fact the cause of the delay. It does not expose the state to the cost of a developer’s own unpreparedness.
The PDA protects a project. It does not improve the approval process itself; only the second track can do that.
Major approvals should carry a timeline within which the authority must process the application and either approve it or deny it.
A denial should state its reasons in writing, and state them specifically. A reasoned denial can be answered, corrected, or challenged. A file that simply does not move in the authority’s office can be none of those things.
There should also be a designated body, other than a court, able to review whether the reasons given for a denial hold up. The case for placing this outside the court system is practical. Court proceedings in Nepal take long enough that a right of appeal to the courts is, for a project on a construction schedule, close to no remedy at all. A review mechanism has to resolve faster than the delay it exists to correct, or it adds a stage instead of removing one.
Taken together, the two tracks change what an investor is being asked to price. A timeline that binds, a denial that must be explained, and a review that moves faster than a court turn the approval process from an open-ended risk into a bounded one. The PDA obligation then covers what delay remains.
A bounded risk can be modeled. A risk that can be modeled is priced far more cheaply than one that cannot.
Environmental clearance and forest approvals stand between a licensed project and the start of construction. A project cannot begin work without them, and neither carries a date a developer can rely on.
The two gates fail for different reasons and need different remedies. Environmental study approval is slow at the point of decision. Forest land and tree cutting approval is slow because of the route the file has to travel.
Approval of an environmental impact assessment currently takes between six months and a year. This is among the longest single steps in a project’s pre-construction path, and it needs to come down substantially.
What makes this case unusual is that the route is not the problem. The sequence of offices an environmental study passes through is not itself unreasonable, and shortening that sequence would achieve little. The time is spent waiting for the reviewing ministry to reach the file and decide it.
The cause is capacity. The ministry is reviewing a volume of assessments that its present staffing cannot turn around, and as the project pipeline grows the gap widens rather than closes.
The measure. The Ministry of Forests and Environment should be equipped with the number of reviewers, and the specialist expertise, that the volume of work actually requires.
This is a resourcing decision rather than a legal one, which makes it unusual among the measures in this note. It is also among the cheapest. The cost of additional review capacity is small beside the cost of holding a pipeline of projects for six months to a year each.
A second part of the delay is created before the file reaches the ministry at all.
The regulation does not set out clearly what an environmental study report must cover. Developers and their consultants therefore prepare reports against their own reading of what is required. The reviewer finds the report incomplete or wrongly scoped, it goes back for revision, and the cycle repeats.
Each of those cycles costs months. None of them reflects a disagreement about the project’s environmental effects. They reflect a disagreement about what the document was supposed to contain.
The measure. A comprehensive guide for the preparation of environmental study reports should be issued, built around a checklist. The guide should state what each study must contain, in what form, and to what level of detail. A developer should be able to establish from the guide alone whether a report is complete before submitting it.
A checklist also changes the review. It gives the reviewing officer a defined standard to apply, which is faster than forming a view from first principles, and it makes the outcome consistent from one project and one officer to the next.
A project that requires forest land follows a familiar route. The application passes through the Department of Electricity Development and then to the Division Forest Office, which makes its recommendation, and the matter goes finally to the Council of Ministers for approval.
This is the same pattern identified earlier in relation to the lease of public land. The decision turns on technical criteria: the area of forest involved, the number and species of trees, and the compensatory obligation that follows. None of that calls for a decision by the Council of Ministers.
The measure. Final approval for the use of forest land and for tree cutting should rest with a designated authority below the Council of Ministers, applying published criteria. Because the same argument applies to the lease of public land, a single reform could reasonably cover both, and addressing them together would be a smaller legislative undertaking than addressing them separately.
Taken together, environmental clearance and forest approval can consume more than a year before construction begins. That uncertainty enters the financing as contingency, and contingency is recovered through the tariff.
None of the three measures proposed here asks the state to take on a new liability. Staffing a ministry adequately, publishing a checklist, and moving a decision to the level at which it belongs are administrative acts.
These measures also make possible something the previous section depends on. Section 2 proposed that the government take on an obligation in the PDA to compensate a developer for delay in specified approvals. A commitment of that kind is only safe to give if the government is confident that it can meet it. No ministry can sensibly undertake to decide within a fixed period while it lacks the people to do so, and while applicants remain unclear about what they are required to file.
The sequencing therefore runs one way. Adequate staffing and a published checklist come first. The contractual commitment then becomes something the state can give with confidence, rather than a risk it assumes without any assurance that it can be met. Environmental clearance and forest approval are, on that basis, among the first approvals that belong on the PDA list.
The benefit does not stop with the developer and the lender. Every month removed from the pre-construction period is a month earlier that generating capacity reaches the grid. The 2035 target depends on how quickly assets are built and not only on how many are licensed, and the approval timeline is one of the few levers on that pace that the government controls directly.
A project sells its power to the NEA under a power purchase agreement (PPA), and the payments NEA makes under that agreement are the project’s only revenue. The lender is repaid out of those payments and out of nothing else.
The financing therefore rests on one question. Will NEA be able to pay, month after month, for the next twenty to thirty years? Every other aspect of the project may perform as planned, yet the lender will not be repaid if the answer is no.
A sovereign guarantee is the government’s answer to that question. Through it, the government undertakes that if NEA does not pay, the government will.
Three kinds of obligation matter to a lender.
The first is the ordinary energy payment, meaning the tariff NEA owes each month for power delivered.
The second is contingent payments. If a dispute goes to arbitration and an award is made against NEA, NEA has to satisfy it. A lender will not assume that a substantial award can be absorbed out of operating cash flow.
The third is what becomes of NEA’s obligations if NEA does not continue in its present form. If NEA were to become insolvent, or be liquidated or restructured, someone would have to assume its financial obligations. A lender wants to know in advance who that is.
A sovereign guarantee moves the obligation from NEA to the state. The lender stops assessing whether a utility can pay and begins assessing whether a country will.
That is a smaller question and a far better documented one. Nepal’s sovereign position is rated, published, and followed. A lender can price it. NEA’s position over a thirty-year horizon, without the guarantee, is something the lender must form its own view on, and lenders price their own uncertainty expensively.
The guarantee therefore affects three things at once: whether the lender participates at all, the margin it charges, and the tenor it will lend for. For foreign currency debt in particular, it is often the difference between a financing that closes and one that does not.
A guarantee commits no money on the day it is given. It is a liability, nonetheless. It is recorded as a contingent liability, meaning an obligation that becomes real only if a specified event occurs, and it is counted as such by rating agencies, by the multilateral institutions, and in the government’s own fiscal reporting.
A guarantee given today therefore carries a cost today, in fiscal headroom used up. It is the state’s balance sheet being lent to the project, and that balance sheet is finite.
This is why a sovereign guarantee cannot be extended to every project that asks for one. The question of which projects receive one has to be settled deliberately.
The government has given sovereign guarantees to some hydropower projects and has declined to give them to others. There is no published basis on which those decisions were taken.
A developer assessing Nepal therefore cannot tell whether its project will receive a guarantee. It must commit to survey, feasibility, licensing, and the early stages of financing without knowing whether the most important piece of credit support in the structure will be available to it.
This is the same pattern that runs through the PDA as a whole. A question capable of being settled once at policy level is left instead to be resolved project by project, and the uncertainty in the meantime is priced.
A sovereign guarantee policy should be published, setting out which projects qualify and on what basis. Four elements would make it work.
First, the eligibility criteria. These might include installed capacity, strategic importance to the grid, whether the project is export oriented, whether the debt is in foreign currency, and whether a development finance institution requires the guarantee as a condition of lending. The criteria should be stated in terms a developer can apply to its own project before it spends money.
Second, the scope. The policy should state which of NEA’s obligations are covered, whether energy payments alone, or energy payments together with termination and arbitral liabilities.
Third, a ceiling. The policy should cap the total guaranteed exposure the state will carry at any one time, either as an absolute amount or as a share of GDP. A ceiling is what allows the Ministry of Finance to support the policy at all, because it converts an open-ended exposure into a budgeted one.
Fourth, the timing. A developer should be able to establish eligibility early, before pre-investment is sunk, rather than at the point of PDA negotiation.
Once the policy is in place, the guarantee is delivered through the PDA, and the PDA policy can carry the same criteria so that the two documents say the same thing.
A sunset provision would sit naturally here. The policy could be enacted for a defined period, with the ceiling and the criteria reviewed at the end of it against how much exposure has actually been taken on and what it has bought.
A published policy does not commit the state to guarantee any particular project. It commits the state to being predictable about which projects it will consider, and predictability is what the investor is asking for.
A developer that qualifies can raise foreign debt against a credit the lender understands. A developer that does not qualify learns this early enough to structure around it, whether through partial risk guarantees, political risk insurance, or a different debt mix, rather than discovering it after capital has been committed.
The state’s exposure also becomes something it has decided and bounded, rather than something that accumulates through a series of individual negotiations.
A foreign lender lends in US dollars and expects to be repaid in US dollars. The project it is lending to earns rupees.
That mismatch is the whole of the problem, and everything in this section follows from it.
The rupee has lost nearly half its value against the dollar in fifteen years
| Date | Exchange rate, NPR per USD | Rupee value lost | Dollar cost increase |
|---|---|---|---|
| January 2010 | 74.90 | Baseline | Baseline |
| 2015 | 101.56 | 26.3 percent against 2010 | 35.6 percent |
| 2020 | 115.18 | 11.8 percent against 2015 | 13.4 percent |
| 2025 | 137.54 | 16.3 percent against 2020 | 19.4 percent |
Across the fifteen years from 2010 to 2025, the rupee lost 45 percent of its value against the dollar. Expressed the other way, a dollar became 84 percent more expensive to acquire. Compounded, that represents depreciation of approximately 4 percent a year.
The direction matters as much as the magnitude. This is not volatility around a stable level, where a loss in one year is recovered in the next. It is sustained movement in one direction across three consecutive five-year periods. Over a loan of ten to fifteen years, a lender does not treat this as a risk that might occur. It treats it as an outcome that will.
The project’s debt service is fixed in dollars. Its revenue is fixed in rupees, under a PPA at an administratively determined tariff that does not adjust for the exchange rate.
Consequently, each year the rupee depreciates the project requires more rupees to purchase the same dollars. The tariff does not move. Revenue has to extend further each year to meet an obligation that is growing in rupee terms, and nothing in the contract compensates for the difference.
A lender that models this will do one of three things. It will require a substantially larger coverage cushion, which raises the tariff the project must earn. It will shorten the tenor to limit its exposure, which the project cannot service. Or it will decline to lend.
Until 2018, Nepal operated a policy of dollar-denominated power purchase agreements. NEA’s payment obligation to the developer was denominated in US dollars.
The effect was direct. A developer receiving dollars could service dollar debt without currency exposure at the project level. For a foreign lender, the mismatch did not arise at all.
The risk did not disappear under that arrangement. It moved. NEA sells power to consumers in rupees while paying the developer in dollars, so NEA absorbed the currency risk. That was a deliberate allocation of the risk to the state-owned purchaser rather than to the private project.
The policy changed in 2019. The dollar PPA was discontinued and a Hedging Regulation was introduced in its place, offering developers a mechanism through which the exchange rate could be locked. The regulation carried severe flaws, and no developer was able to use it.
A second Hedging Regulation followed in 2021, replacing the 2019 instrument. It envisages a hedging entity that would provide the cover. That entity has still not been established.
The position since 2019 is therefore that a project has neither a dollar PPA nor a functioning hedge. Currency risk sits with the developer and its lenders, and nothing in current law or practice mitigates it.
A foreign lender is being asked to carry a currency risk that is large, one-directional and uncovered, over a tenor long enough for the exposure to compound against it throughout.
Cover can be purchased in the international market, but it is not inexpensive. Hedging a project of this kind commercially raises project cost by five to six percent.
This also helps explain the composition of foreign capital described at the start of this note. A development finance institution operating under a mandate can absorb an uncovered currency risk. A commercial lender answerable to a return generally cannot and does not. The absence of a functioning hedging mechanism is one of the reasons the pipeline remains weighted toward mandate-driven capital.
The immediate task is to establish the hedging entity that the 2021 Regulation envisages.
The reason it has not been established is instructive. The institutions that would have to assume the role were not properly consulted when the regulation was drafted. No market exercise was carried out to establish what terms, what pricing and what risk limits a potential hedging entity could accept. A framework designed without the participation of the parties expected to operate it has attracted no enthusiasm from them, which is the predictable result.
The sequence should be reversed. Consult the institutions capable of acting as hedging entities first. Establish what they are able to carry, at what price, and on what conditions. Then revise the regulation to match what the market can deliver, rather than requiring the market to fit a framework it had no part in designing.
Beyond that, the state faces a choice it has not yet made expressly, and the choice concerns where this risk is to sit.
It can sit with NEA, through a dollar-denominated or dollar-indexed PPA, as it did before 2018. It can sit with a hedging entity, through a functioning facility, against a fee. Or it can sit with the developer, which requires purchasing cover internationally at five to six percent of project cost and recovering that through the tariff.
Each of these carries a cost and none of them is free. What should be clear is that the present position is the least satisfactory of the three. The risk sits with the party least able to carry it, no mechanism exists to move it, and the cost is paid in any event, through projects that do not reach financial close and capital that does not arrive.
Every PDA negotiation reaches the governing law and dispute resolution clauses, and this is where negotiations meet the most resistance. Because there is no policy on the question, the same argument is reopened from the outset on every project.
The position taken by government authorities has consistently been that the contract should be governed by the law of Nepal, and that disputes should go either to arbitration seated in Nepal or to the courts of Nepal. Foreign investors and their lenders ask for something different, and the gap between the two positions does not close easily.
Much of the resistance rests on a misunderstanding of what a governing law clause actually does, and clarifying it is the principal purpose of this section.
A governing law clause does not release the project from the law of Nepal. The project company is incorporated in Nepal. It is subject to Nepali law in full, including licensing, taxation, labor, environmental compliance, foreign exchange, land, company law, and criminal law. There is no exception to this, and no governing law clause can create one.
What the clause decides is narrower. It determines which body of law is used to interpret the contract and to resolve a dispute arising under it. It governs the meaning of the words the parties wrote, and it becomes relevant only once a dispute has arisen.
The distinction matters because nothing about sovereignty or regulatory control is at stake in it. The state keeps every power it has over the project. What is being settled is how a contract between two parties is to be read if they fall out over it.
The preference for an established governing law is not a judgment on the quality of Nepali law. It is a question of how much of that law has been tested in one narrow area.
Nepal has not yet seen many disputes over large infrastructure concessions. The principles that govern contracts of this kind have therefore had few occasions to be settled here, and precedent developed in other commercial contexts may not translate to projects of this scale.
One example makes the point. The Supreme Court has taken the view that liquidated damages cannot take the form of a penalty. PDAs in Nepal regularly provide for a termination payment. Whether a Nepali court would treat a termination payment as a penalty, and so decline to enforce it, has not been decided. The question is open.
This is not a defect. It is the ordinary position of a jurisdiction in which such disputes have been rare. The consequence for an investor is nonetheless concrete. A clause it has negotiated, and priced, may or may not be given effect, and it will not find out until the day it needs that clause.
An established governing law offers the opposite assurance: reasonable confidence that the words of the contract will be applied as the parties intended, without the validity of the clauses themselves becoming the subject of the dispute.
The forum is a separate question from the governing law, and it turns on a different concern.
The state, or an entity the state owns, is a party to the contract. Where the forum deciding the dispute belongs to one of the parties, it is not read as neutral to the other, whatever the merits of how it would in fact decide. This is a structural feature of the arrangement rather than a comment on any particular institution.
Nor is it particular to Nepal. It is why states with long established judiciaries routinely agree to foreign-seated arbitration in concession contracts. What an international forum supplies are neutrality, and a procedure that both sides know in advance and neither control.
This is worth stating plainly, because it is often left out of the discussion.
Nepal acceded to the New York Convention on the Recognition and Enforcement of Foreign Arbitral Awards in 1998, subject to the reciprocity and commercial reservations. The Arbitration Act 2055 provides for the recognition and enforcement of foreign awards, and the High Court has jurisdiction over enforcement.
A foreign arbitral award is therefore already recognized and enforceable in Nepal. Agreeing to foreign-seated arbitration in a PDA is not a new concession by the state. It is consistent with a treaty commitment Nepal took on nearly thirty years ago.
The PDA policy should state the government’s position on governing law and on dispute resolution, so that the question is settled once rather than argued afresh on every project.
The policy should set out the distinction described above expressly. If it is recorded as policy that a governing law clause does not displace Nepali regulatory law, the point does not have to be re-established at each negotiation, and officials are not asked to accept it on a developer’s assurance.
The policy should then state what the state is prepared to agree: the governing law of the contract, the seat and institutional rules of arbitration, the number of arbitrators, and the language of the proceedings. It may reasonably differentiate, offering foreign governing law and a foreign seat for projects above a stated size or with foreign debt, and Nepali law and seat below it.
A published position removes the negotiation from the shoulders of the individual official. An officer who agrees to foreign governing law project by project is exposed personally if the choice is later questioned. An officer applying a published policy is not. That is the same incentive problem identified in relation to government approvals, and it is answered the same way.
For the investor, the effect is that the enforceability of the contract stops being a matter of negotiation and becomes a known feature of investing in Nepal. A contract whose clauses will be read as written is worth more than one whose clauses may not be, and the difference is reflected in what lenders will lend and at what price.
Force majeure covers events outside the control of either party that prevent a project from performing. The question the contract has to answer is who bears the cost and the delay when one occurs.
The starting point is not the contract but the insurance market. An investor or a lender does not look to the state first. It looks first for an insurance product, and it turns to the state only for what the market will not cover. Force majeure allocation is therefore shaped by what can be insured, and it changes as the insurance market changes.
Cover for natural events is commonly available. Flood, landslide, and earthquake can be insured, the premium is a known cost, and a lender is satisfied by the policy rather than by the state.
Cover for political events is not. War, insurgency, civil unrest, and blockades at the border are either not offered in the market or offered at a price that moves the cost of the project materially. This is where foreign lenders and investors seek protection from the state, and they seek it because the market does not supply it rather than as an opening position in a negotiation.
In some PDAs the government has agreed that political force majeure sits wholly within the state’s obligation. If war, insurgency, civil unrest, or a border blockade occurs, the state carries it.
Relief in those agreements runs in two currencies. The state reimburses cost the project incurs as a result of the event. It also restores time, by extending the deadlines in the contract and by extending the revenue term, so that a delay the project did not cause does not shorten the period over which it earns.
These are the same two currencies proposed in relation to approval delay, and the parallel is not accidental. Where the state is the party better placed to carry a risk, cost and time are the two things it can give. This is a correct approach to deal with political force majeure risks as there are no commercial insurance products available for it.
Where insurance generally answers the risk, the state can reasonably offer a lower degree of protection, and it should not be asked to duplicate cover that the project can buy.
The time dimension survives all the same. Insurance pays for damage. It does not give back the months lost to repair and reconstruction, and it does not extend the period over which the project earns. A project rebuilt at the insurer’s expense and handed back a shorter earning life has still lost money.
The state should therefore take a flexible view on time even where it declines to take cost. Extending contract deadlines and the revenue term after an insured natural event requires no cash outlay from the state, and it removes a risk the project cannot insure against.
The recent floods in Nepal matter here, and their effect is not yet visible. Insurers reprice after losses of that kind. Premiums for natural catastrophe cover may rise by a wide margin, which either raises the cost of the project or, at the limit, leaves some cover unavailable at any commercially sensible price.
Any position the state fixes today against today’s market may need to be revisited as the market moves.
The government has taken different approaches on force majeure from one project to the next, and there is no published basis for the difference. A developer cannot tell in advance which treatment its project will receive.
The PDA policy should set out the state’s position on force majeure. It should state the distinction between force majeure coverage on political and natural events, and which heads of relief attach to each, meaning cost and time for political events and time alone for natural events.
A developer can then price the risk it is being asked to carry, insure what the market will insure, and rely on the state for what it will not. The lender can see the same picture before committing.
What the state gives up is the discretion to treat comparable projects differently. What it gains is that its exposure follows a stated rule, moves only as the insurance market moves, and can be reviewed on a schedule rather than in the middle of a negotiation.
A project is priced once. A developer commits equity and a lender commits debt on the basis of the laws in force at that moment, covering what the project will cost to build, what it will cost to run, and what it will be allowed to keep. Those assumptions are fixed at financial close, and the project then operates under them for twenty-five or thirty years.
The state keeps the power to change any of those laws at any time, and will legitimately exercise it. Change in law protection answers the question this creates. When the rules change after the money has gone in, who absorbs the difference?
It cannot be the lender. A loan agreement does not forgive repayment because the law changed, so a change that raises cost or cuts revenue falls first on the developer’s return, and then, if it is large enough, on the project’s ability to service its debt at all. The risk does not disappear by being left unallocated. It settles on the parties least able to have foreseen it, which is why change in law is ordinarily treated as a risk that should be absorbed by the state. The state is the only party that decides whether the change happens.
The tax regime is where this risk is most acute. When a lender comes into a project it reads the statutory position as it stands, in particular the income tax holiday period and the customs duty exemption period, and prices the cost and return of the project against it. Those concessions are not incidental to the numbers. They are a large part of what makes the numbers work.
A change to that regime after the money is committed changes the basis on which the loan was made. The principle a lender is protecting here is a simple one: a change in the tax regime cannot undo what was priced in.
The first question is who is entitled to the protection, and the answer needs to be broader than it usually is.
Some protection already exists for the project company under the income tax laws. The project company is not the only party exposed.
Contractors to the project need the same protection, because a change that raises their cost reaches the project through the construction contract. Lenders need it. Shareholders need it, particularly in relation to withholding tax on dividends, which determines what they actually receive.
The reason for drawing the scope widely is that a change in law is an act of the state. Where the state has said, at the point of investment, that this is the legal regime which will govern the project, a later change is not something any private party could have predicted or priced. Any change that moves the numbers for the developer, the lender, or the shareholders should fall within the protection.
A change in law can also raise the cost of building and operating the project.
The state may introduce a requirement that did not exist when the project was priced: a new category of environmental study, an assessment of glacial lake outburst flood risk, a different standard of hydrological instrumentation. Each of these is a legitimate regulatory decision. Each also costs money the project did not budget for.
Changes of this kind belong within the protection for the same reason tax changes do. The developer could not have priced them, and the state decided them.
This point needs to be stated carefully, because it is easily misread.
Change in law protection does not exempt the project company from the new law. When the state makes a law, the company has to comply with it. If a new study is required, the project has to carry out the study. If a new standard applies, the project meets it.
What the state promises is narrower than that. It promises that the cost of compliance will be reimbursed. The state’s regulatory power is untouched and so is the obligation to obey. What moves is who pays.
A protection of this kind cannot respond to every change in the statute book. Countries legislate constantly, and much of what they pass is minor: a new registration, a new annual filing, an updated compliance requirement. Changes of that size are an ordinary cost of operating anywhere, and no developer expects to be compensated for them.
Two devices keep the protection proportionate, and they work together.
The deductible sets a threshold below which an individual change does not count. A change whose cost effect falls under the threshold is absorbed by the project and does not enter the calculation at all.
The basket sets a cumulative threshold at which the protection is triggered. Qualifying changes accumulate, and the developer bears them until their combined cost effect passes the basket amount. Only then can a claim be made. The basket might be set, for example, at NPR 20 million.
The reasoning behind the basket is that a developer entering any country expects some margin of legal change and prices for it. The basket is that expected margin, made explicit. Below it, the developer carries what it always assumed it would carry. Above it, the change has gone beyond what could reasonably have been anticipated, and the state carries it.
Between them, the deductible decides which changes count, and the basket decides when the protection is triggered.
The PDA policy should set out the state’s position on change in law protection: which projects receive it, which parties are protected, which changes are covered, and at what deductible and basket levels.
A lender that can rely on the fiscal regime it priced will lend against it. A lender that cannot will either discount those concessions entirely, which raises the return the project has to generate, or decline to lend for the tenor the project needs.
The state gives up no regulatory power by providing this protection. It keeps the ability to change any law at any time. What it accepts is that where it does so, and where the change is large enough to matter, it will meet the cost rather than leave it with parties who had no way to foresee it.
A power project is built mostly with borrowed money, and the lender’s security is the project itself. Security of that kind is effective only if the asset can be sold.
This is where a power project differs from most of what a bank lends against. A building or a vehicle can be taken back and sold on an open market. A hydropower plant cannot. Its value does not come from its turbines and its concrete. It comes from its license to generate and its agreement to sell power to NEA. Take those away and what is left is a structure in a river gorge that nobody can operate and nobody will buy.
Consequently, if the project’s agreements come to an end, the lender’s security can be worth close to nothing on the very day it is needed. The plant may be complete and in working order, and still be worth almost nothing.
A transfer price is the answer to that problem. It is a price, fixed in advance and written into the PDA, at which the government will take the project over if the project comes to an end early.
What it achieves is straightforward. It gives the project a known buyer at a known price. The lender is no longer left to speculate on what a stranded power plant might realize. It knows who will buy it, and it knows roughly what will be paid.
The price is not a single figure. It depends on why the project came to an end. A project can end for many reasons. The company may default. The government may default. A force majeure event may make the project impossible to complete, or a change in law may make it unviable. The PDA sets out each of these categories and the price payable in each.
The concept is not new to Nepal. It already appears in PDAs the government has signed. What is missing is a settled position on when it applies and at what level.
An infrastructure project is built on a high gearing ratio, which means most of the money is borrowed. A ratio of eighty percent debt to twenty percent equity is normal.
This is not comparable to ordinary commercial lending. A bank funding seventy or eighty percent of a project’s cost earns only its interest margin. It takes none of the equity upside. If the project does extraordinarily well, the bank receives exactly what it would have received if the project had merely done adequately.
A lender with limited upside and very large exposure will insist on the one thing that is left, which is protection on the downside. It cannot put in most of the money, take a fixed and modest return, and also carry the risk of losing the principal if something goes wrong. The developer, having put in twenty percent, has far less at stake in absolute terms. This is why termination is a bigger question in project financing than it appears from the outside.
Nepal has very little true project finance. Lending to projects rests largely on corporate guarantees and personal guarantees. If the project fails, the bank has recourse to the sponsor, and behind the sponsor to personal property and other assets.
A foreign lender will not lend on that basis. It will not assess the personal wealth of individuals in Nepal, and it has no practical way to enforce against it. It lends against the project itself, which is what non-recourse financing means. The project’s assets and its contracts are the whole of the security.
That is why the transfer price matters so much to foreign capital in particular. It is what gives the project asset a value on termination. Without it, the security a foreign lender is relying on can fall to nothing on the day it is most needed.
The PDA states the transfer price payable for each category of termination event.
Where the government is in default, established practice has been for the price to cover the sunk cost of the project, all project loans together with accrued interest, and the equity contributed together with a return on it.
The treatment of other events varies, and this is where negotiations concentrate.
This is asked in every PDA negotiation, and it is a fair question. If the project company has defaulted, why should the government pay anything at all?
The answer is that the lenders will not rely on the developer to protect them against that risk, and if nobody carries it the financing does not happen. What the state is undertaking, in substance, is that the lenders will be made whole whatever the cause. That undertaking is the thing foreign debt is lent against.
The undertaking sounds open-ended. It is not, because of how project loans are drawn.
A loan agreement normally requires the sponsor to inject its equity before the first drawing of debt, or to inject it proportionally alongside the debt. The developer’s money goes in first.
Take a project on an eighty to twenty ratio that reaches fifty percent completion and then terminates. The twenty percent of equity has already been spent on the works. Only thirty percent of the cost comes from the lenders. The government takes over a project that is half built and pays out the debt portion, which is thirty percent of the total cost.
The state is therefore not paying more than it receives. It receives a half-completed asset and pays a little over half of what that asset cost to reach that point.
The point of that arithmetic is not that the state comes out ahead. It is that the state is not overpaying, and the difference between those two readings matters a great deal.
A government that approached termination as a chance to acquire assets cheaply would be taking the wrong lesson from it. Nothing would undermine the mechanism faster. A transfer price works only because lenders believe the state will honor it, and a state seen to be looking for an advantage in the arrangement is not one they will rely on.
The posture has to be forward-looking. The state should be willing to make the lenders whole, to pay properly for what it receives, and to pass the project on rather than hold it. In practice that means aiming for a back-to-back sale, in which the asset taken over from the original developer and its lenders is transferred onward on terms that recover what the state has paid out.
The mechanism assumes a further step. Having taken the project over, the government re-tenders it to another party.
The transfer price the government paid is recovered from the incoming developer, who takes the project forward and carries the risk from that point. The state’s role is to stand between one developer and the next, not to become the owner of a half-built power plant.
The PDA policy should state the transfer price the government will pay for each category of termination event, and the basis on which it is calculated.
A defined transfer price is what converts a project into something a foreign lender can lend against. It gives the project asset a floor value that does not depend on the PPA surviving, on the sponsor’s other wealth, or on how a dispute is eventually resolved.
Without it, a lender looking at Nepal is asked to take construction risk, operating risk, and hydrology risk, and to accept that its security may be worth nothing if the offtake agreement falls away. That is not a risk that gets priced higher. It is a risk that stops the financing altogether.
Termination and transfer price is the most complex matter this note deals with, and what is set out above does not cover all of it.
The arithmetic offered earlier assumes there is still a project to take over. There are situations in which there is not. A project can be destroyed outright, as happened to projects swept away in the recent floods. Had a PDA been in place, the government could have found itself obliged to make the lenders whole while receiving nothing at all in return.
That exposure is real and cannot be argued away. It is a judgment about risk appetite which the government has to make for itself, and it is a contingent liability that sits on the state’s books whether or not the event ever occurs.
The right treatment of cases of this kind, and of the many other situations a transfer price regime has to anticipate, needs careful work with legal and financial expertise. What this note sets out is the shape of the mechanism and the reasoning behind it. It is not the whole of the subject.
Project finance rests on the project. A lender that will not look to the sponsor’s other wealth has to look to the project’s assets instead, and it has to be able to take security over them and enforce that security if things go wrong (events of default).
This only works if the security package is complete. A lender does not assess a security package by what it contains. It assesses it by what is missing. If a material part of the project cannot be secured, the package is incomplete, and a project finance lender treats a gap as a reason not to lend rather than as a point to price.
The Land Act sets a ceiling on how much land a single holder may own. The limits are 11 bigha in the Terai, 30 ropani in the Kathmandu Valley, and 75 ropani in the hills.
Energy projects routinely need more than this, and the requirement is growing. Solar generation is land intensive by nature, and transmission lines run across long corridors.
A project can hold land above the ceiling, but only with approval, granted by the Ministry of Land Management, Cooperatives and Poverty Alleviation or, in some cases, by the Council of Ministers. What matters for financing is the condition attached. Land held above the ceiling cannot be mortgaged to raise bank finance. It also cannot be freely sold, transferred, or leased.
The consequence is direct. The land on which the project is built, and the works constructed on that land, sit outside what a lender can take security over. For a bank lending on project finance terms this is not a matter for negotiation. It is a reason to decline.
This is not a provision aimed at energy projects. The ceiling exists for land reform reasons and was framed with individual landholders in mind. It nonetheless operates today as one of the clearest legal obstacles to project finance in the sector.
Where the project asset cannot secure the debt, the lender secures the debt against something else, and in Nepal that means corporate guarantees and personal guarantees from the sponsors.
That is the practice described earlier in relation to termination, and it is the practice foreign lenders will not accept. A rule that makes the project asset unsecurable therefore does more than inconvenience a transaction. It forces the market back into the very model that foreign capital has declined to use.
Exemption above the land ceiling joins (i) the lease of public land, (ii) the approval of forest land use and, (iii) tree cutting as a matter escalated to the Council of Ministers.
These are three separate approvals, each turning on criteria that can be written down, each requiring a decision by the full Cabinet. This strengthens the case made earlier for addressing them together rather than one at a time.
The PDA policy should state the security package the state is prepared to facilitate. At minimum a project finance lender will expect the following.
Beyond the assets themselves, two mechanisms matter. The first is a direct agreement between the state, NEA, and the lenders, recording what the lenders may do and what the state will recognize. The second is confirmation that a foreign lender, or a security trustee acting for a group of lenders including foreign ones, can hold and enforce that security in Nepal.
One element deserves separate attention, because it serves the state at least as well as it serves the lender.
Step-in rights, sometimes called substitution rights, allow lenders to replace a failing developer with another operator rather than let the project collapse. A lender facing a defaulting sponsor would almost always prefer to substitute it than to terminate and claim a transfer price.
For the state, the difference is substantial. Substitution keeps the project alive, keeps construction moving, and costs the government nothing. Termination requires the state to pay out the transfer price, take over a partly built asset, and re-tender it. Step-in rights are therefore the state’s least expensive route out of a failing project, and they should be treated as such rather than as a concession extracted by lenders.
Substitution depends, however, on something the law does not presently provide. A lender that steps in has to be able to transfer the generation license to the incoming operator, and there is no clear statutory provision governing when that transfer takes place or who is eligible to receive it.
The same gap runs through the project’s land rights. A project holds its forest land and its public land on lease from the state, and the law provides no mechanism for assigning those leases. An incoming operator that acquires the license has no clear route to the land on which the license is exercised, which leaves the substitution incomplete even where the license itself can be transferred.
These gaps have so far been filled by direct agreement between the government and the lenders on individual projects. A PDA can address it as well, and the PDA policy should say that it will.
The better answer is statutory. If the Electricity Regulation, or another instrument of the same standing, set out the grounds on which substitution may be exercised, the conditions on which the license and the project’s land leases transfer, and the criteria an incoming operator must satisfy, the question would not need to be negotiated in each PDA at all.
A statutory right would also reach considerably further. It would be available to every licensed project, including the large majority that have no PDA and no direct agreement. Those projects are financed domestically today, and the same mechanism would give their lenders an alternative to enforcement and their projects an alternative to collapse.
The PDA policy should set out the security package the state will facilitate and the direct agreement it is prepared to enter into.
The PDA alone cannot address the remainder. This is one of the few areas in this note where a contractual commitment is not enough, because the obstacle is statutory. Legislative change is required on three points: permitting security over land held above the ceiling by a licensed energy project, confirming the position of foreign lenders and security trustees taking and enforcing security in Nepal, and aligning the mortgage restriction with the exemption that allowed the land to be acquired in the first place.
A complete security package is the precondition for project finance, and project finance is the form in which foreign capital comes to this sector. Without it the state can publish a PDA policy, guarantee NEA’s obligations, and fix every approval timeline, and a foreign lender will still find that it cannot take security over the asset it is being asked to fund.
This is the least visible of the constraints in this note and among the most binding. It is also the one that a PDA on its own cannot solve.
Right of way is a transmission question. It arises where a line crosses land the project does not own and does not need to own, and it is not a material issue for generation projects.
It becomes more pressing now that NEA has begun inviting private participation in transmission. A state utility building its own lines can absorb an unclear legal position. A private developer financing a line cannot, and nor can its lenders.
The regulation currently in place is light. It prescribes clearances, meaning how much space must be left unbuilt on either side of a line, and how close to it structures may stand or tall trees may be planted.
What the regulation does not do is say what kind of legal right a right of way actually is.
Nepali law recognizes two principal rights over land. Ownership is acquired by purchase. Possession is acquired by lease. A right of way is neither. The project does not take the land and does not occupy it. What it takes is a permanent restriction on how the owner may use it.
That is a distinct kind of right, and it has no name in Nepali law at present. The agreements signed for it therefore rest on a category the statute does not recognize.
Right of way agreements are signed to run in perpetuity, which reflects the plain fact that a transmission line is not built to be removed.
The Civil Code limits the term of a lease for infrastructure projects to forty years, subject to renewal. Because a right of way sits closer to a possessory right than to ownership, there is a real question whether a perpetual right of way can stand against that limit.
For a lender this is not an academic concern. The right of way forms part of the security package it expects to be able to enforce. A right of indefinite duration and uncertain legal character, resting on an agreement that may conflict with the Civil Code, is not security a project finance lender will rely on.
Present practice is to pay right of way compensation that is between fifteen and thirty percent of the value of the land.
The owner loses considerably more than that fraction suggests. Building on the land is no longer possible. Its agricultural use is limited. Banks will not accept it as security, so the owner cannot mortgage it. Land beneath a transmission line loses much of its market value and much of what made it worth holding.
A right restricted this severely, compensated at a fraction of the loss, is the principal reason transmission projects meet resistance on the ground. That resistance is not a separate problem from the compensation. It is the consequence of it, and it reaches the project as delay and cost.
Two things are needed here, and neither is a matter for the PDA or the PDA policy. Both require statutory change.
First, the law should create and define the right of way as a distinct legal right over land. It should state what the right permits the project to do, what it prevents the owner from doing, how long it lasts, how it is registered, and how it may be transferred. Setting its duration expressly would also settle the conflict with the Civil Code, rather than leaving a perpetual agreement to be tested one day against a forty-year limit.
Second, the basis on which right of way compensation is fixed should be reconsidered. A flat percentage of land value does not measure what the owner has given up. The considerations set out earlier in relation to land acquisition apply here with equal force. The loss differs from one holding to another, and compensation determined without regard to that difference will continue to be seen as unfair, with the consequences for delivery that follow from it.
For the lender, a defined statutory right is what allows the right of way to form part of the security package at all. For the landowner, compensation that reflects the real restriction is what removes the grievance. For the state, the two together are what make private participation in transmission financeable.
Beyond the individual risks addressed above, the institutional arrangements through which a project is approved and administered are themselves determinants of investor confidence.
IBN is the institution a large foreign investor deals with. Much of what this note proposes assumes that some authority can make the approval chain work as a whole, and IBN is the candidate for that role.
IBN was established to serve as a one-stop solution. The intention at its creation was an institution that would give developers ready access to every approval a project requires, deal with them directly, and resolve their difficulties rather than add to them.
The power required to discharge that function was never conferred.
This is the central difficulty, and everything else follows from it.
IBN’s mandate covers investment promotion, project screening, transaction coordination, and approvals. In PDAs, IBN takes on obligations to procure the government approvals a project requires.
IBN has no regulatory power to direct the authorities that actually hold those approvals. It cannot require the Division Forest Office to decide within a period, or the reviewing ministry to complete an environmental study assessment. It has undertaken to deliver an outcome it has no means of compelling.
That is a defect in the design rather than in the institution. No agency can be held to a timeline it has no authority to enforce.
Mandate has outrun capacity. Staff numbers, transaction-advisory bandwidth, and the ability to retain specialist talent on market-aligned terms have not kept pace with what IBN is asked to do.
Federal mandates overlap. The Ministry of Finance holds fiscal policy authority, the Department of Industry handles smaller investments, line ministries control sectoral approvals, and the Office of the Prime Minister exercises informal coordination. No single agency is accountable for a project from start to financial close. Recent changes have narrowed IBN’s remit further, with foreign investment approvals now routed through the Department of Industry where IBN previously held authority over high-value cases.
The Council of Ministers remains the final clearance point. Projects IBN has already approved are escalated to Cabinet, on matters both material and routine.
Jurisdiction is fixed by rigid thresholds. The Public-Private Partnership and Investment Act 2075 demarcates IBN’s remit by generation capacity for hydropower and by estimated project cost for everything else. IBN cannot take on a project below those thresholds even where the project would benefit from PPP delivery, including demonstration projects in emerging sectors.
There is no policy defining what IBN may offer. Because the facilities, concessions, and protections available to an investor are not set out anywhere, the negotiating team has no guidance on the level it is authorized to agree.
The result is reflected in how the market now perceives IBN. Among sponsors and lenders, it is increasingly viewed not as a facilitator of investment but as an additional layer in the approval process, one that adds cost and time without a corresponding gain in coordination, certainty, or speed.
The consequence extends beyond perception. Developers now structure their projects to remain outside IBN’s jurisdiction where they are able to, rather than seeking it out. The approvals required under the Public-Private Partnership and Investment Act 2075 are experienced as a second set of approvals rather than as a substitute for approvals obtained elsewhere. An institution created to be the preferred route has become one that sponsors seek to avoid.
The obvious response is to let IBN issue every approval itself. That is the wrong model, for two reasons.
The first is expertise. Some approvals require judgment IBN does not hold and should not be expected to develop. The Department of Electricity Development has the experience to assess how many trees a given type of project requires to be felled. Moving that decision to IBN would not make it better or faster.
The second is legal exposure. Environmental study approvals, tree-felling approvals, and approvals relating to explosives attract challenge. An IBN that issued them would carry the litigation risk that goes with them, and a body defending its own clearances in court is not a body facilitating investment.
The workable alternative is narrower and stronger. IBN should be empowered to direct the authority that holds the approval, and that authority should be accountable for complying with the direction. IBN’s own power is limited to ensuring that applications are examined within the prescribed time.
This matches the diagnosis. The problem was never that IBN lacks approval powers. It is that IBN cannot compel the bodies which have them.
The FY 2026/27 Budget proposes an Investment Express single-window platform. The commitment is welcome, but its design will determine whether it succeeds.
Comparable single-window systems established elsewhere in government have functioned as document collection centers. Applications are received and passed on, and nothing about the underlying timeline changes. A window of that kind does not remove a step. It adds one.
The difference is authority. The platform needs the power to direct an authority to examine an application, to track where every application stands, and to escalate when one stops moving. Digital tracking supplies the visibility. The power to direct is what makes the visibility matter.
The Budget has committed to ending secondary Cabinet clearance for IBN-approved projects, and that commitment should be codified rather than left as policy intent.
IBN approval should be complete and binding on line agencies for the ancillary clearances tied to the project, including transmission line and access road permissions. A sponsor that has been approved should not have to restart the process with individual agencies afterwards.
This would also address the pattern running through the earlier sections of this note. The lease of public land, the use of forest land and the felling of trees, and exemption from the land ceiling each currently require a decision of the Council of Ministers. All three concern projects the state has already decided to support.
Land acquisition is the constraint this note identifies as the most persistent, and the question left open in that section was which authority should lead it. IBN is the appropriate answer, but not in its present form.
IBN should be empowered to formulate a land acquisition policy for projects within its remit. Under that policy it would operate a price determination mechanism that is faster, more certain, and more equitable than the present one, a dedicated unit to approve the land a project requires, a mechanism for distributing compensation, and the coordination function needed to ensure that acquired land is actually cleared and handed over.
This is a larger change than adjusting IBN’s approval powers. It gives IBN an operational function rather than a coordinating one, and it requires both staffing and statutory authority. It is proposed because the alternative, which is leaving compulsory acquisition where it currently sits, has not worked.
The PDA policy proposed throughout this note is, in institutional terms, a document for IBN.
At present an IBN negotiator asked to agree a governing law, a termination payment, or a change in law threshold has nothing to point to. The position is established afresh in each transaction, and the officer who agrees it carries the exposure personally if it is later questioned.
A published policy protects the negotiator as much as it informs the investor. It converts each of those questions from a judgment an individual has to defend into the application of a position the state has already taken.
The measures above assume an IBN able to carry them. Staffing levels, transaction-advisory capability, and the ability to recruit and retain specialists on market-aligned terms all require attention, and none of it follows automatically from a change in mandate.
The measures above address specific failures. Behind them sits a larger question, which is whether IBN’s design still matches the purpose for which it was created.
A review should begin where the institution began. What was IBN established to do, what powers would an institution require in order to do it, and which of those powers does it hold today? On the present evidence the distance between the first question and the third accounts for the whole of the problem. An institution given a mandate without the authority to discharge it is not corrected by being given more to do.
Two matters in particular call for reconsideration: how a project enters IBN’s jurisdiction, and what happens to it once inside. Four examples indicate what such a review would have to confront.
Jurisdiction is fixed where it should be flexible. A project either clears the threshold or it does not. Nothing about the project’s wider significance enters the calculation.
Consider what that excludes. If Nepal intends to establish itself as a location for data centers, the sector has to begin somewhere, and it will not begin at the cost threshold the Act prescribes. A data center costing a billion rupees falls well outside IBN’s remit. A first project of that kind, delivered well, would demonstrate to every subsequent investor that it can be done here. An institution whose purpose is to create an investment environment should be able to take on a project of that size precisely because of what it would establish. At present it cannot.
It is unclear whether IBN may license where a licensing authority already exists. The Department of Electricity Development and the Ministry issue generation licenses for hydropower. Whether IBN may issue a license in parallel, or whether it is in fact required to, has not been resolved. A sponsor inside IBN’s jurisdiction therefore cannot state with confidence which body licenses its project.
The fees charged to projects within IBN’s jurisdiction are disproportionate. A project pays materially more for being inside the regime than it would outside it, without a corresponding gain in speed or certainty. A fee that secures nothing the sponsor values is read as a penalty for having qualified.
Some prescribed procedures create difficulty rather than discipline. The Swiss challenge method is one example. Under it an unsolicited proposal is published, third parties are invited to better it, and the original proponent is given a right to match. The sponsor that carried the cost and the risk of developing that proposal may lose the project, or retain it only by matching a competing bid that its own work made possible. Whatever the merits in principle, a developer deciding whether to bring a project forward at all reads this as a reason not to.
The Public-Private Partnership and Investment Act 2075 contributes to the same difficulty. A number of its provisions are unclear, and uncertainty in the statute that governs the approval route is read by investors as uncertainty about the approval itself.
Taken together these point to one conclusion. A project should not be worse off for falling within IBN’s jurisdiction than for falling outside it. Where it is worse off, in cost, in time or in certainty, the institution produces the opposite of what it was created to produce, and no adjustment at the margin will correct that.
The distinction that matters is between an institution that adds a step and one that removes several. IBN was created to be the second and is currently experienced as the first.
An IBN that can direct other authorities, whose approval is final for the project, which negotiates from a published policy, and which has the capacity to do those things, is a materially different proposition for an investor deciding whether to commit to Nepal.
The two measures set out in this part do not concern risks to be reduced. They are instruments available to the state and to the domestic financial system for mobilizing capital that is presently immobilized. Both depend on the protections discussed in the previous part being in place, but the case for each rests on what the state stands to gain rather than on what an investor requires.
Asset monetization, also called asset or capital recycling, raises money from infrastructure the state has already built and is already operating.
The government grants a private party the right to operate an asset for a fixed period, in return for an upfront payment or a stream of periodic payments. Because the asset is already built and already earning, the transaction concerns operating rights and revenue rather than construction. The private operator runs and maintains the asset to agreed service standards and earns its return by operating it well. The state receives capital it can deploy elsewhere.
This distinction needs to be made early and clearly, because the objection will be raised.
Asset monetization is a time-bound transfer of operating rights. Ownership remains with the state, and the asset returns to public control at the end of the concession period. It is a structured partnership for a defined term, not a sale.
The point carries particular weight in Nepal, where public perception of privatization has been shaped by the experience of the 1990s. A program that is designed to retain ownership, to specify service standards, and to hand the asset back is answering that concern in its structure rather than in its presentation.
The case for asset monetization is strongest where public borrowing capacity is tightening while the investment requirement is rising. That is Nepal’s position.
Public debt has reached 44.69 percent of GDP as of 2025, up from around 25 percent in FY 2015/16. Capital expenditure in the FY 2025/26 budget is NPR 407.89 billion, which is 20.8 percent of the national budget, and NPR 86.01 billion of that has been allocated to the energy sector. Against the Roadmap requirement, those figures fall short of what is required, and the room for additional sovereign borrowing is narrowing rather than widening.
Asset monetization does not create new debt. It converts value already locked in operating assets into capital available for new ones.
In the energy sector the relevant assets are overwhelmingly in transmission, because that is what the state owns.
NEA holds more than 6,700 circuit kilometers of transmission line as of FY 2024/25. Rastriya Prasaran Grid Company Limited is undertaking eleven high voltage transmission projects of roughly 260 kilometers and plans 16,000 circuit kilometers by 2035. Independent power producers generally own only the lines connecting their plants to the grid.
There is therefore a substantial stock of mature, operating public assets on one side, and an expansion pipeline of a size that will compete with every other national priority for public capital on the other. Asset monetization is the mechanism that connects them.
The asset is already built, which removes design and construction risk from the transaction entirely.
That changes who will invest. Institutional investors, including pension funds and infrastructure funds, look for stable and predictable cash flows and have little appetite for greenfield construction risk. Operating transmission assets suit them, and they are a pool of capital Nepal has not yet reached. Over time, transactions of this kind help establish infrastructure as an investable asset class in the country, which supports deeper financing options including local currency structures.
Australia ran a federal asset recycling program after 2013 which had unlocked more than USD 17 billion for new infrastructure by 2018, across ports, electricity generation, transmission and distribution, and roads.
Indonesia introduced a Limited Concession Scheme under a 2020 regulation, granting limited concession rights over existing assets held by the central government or state-owned enterprises, in return for an upfront concession fee or revenue sharing. Partners are selected through prequalification-based tender.
In the United States, Indiana monetized a 157-mile toll road in 2005 and 2006 through a 75-year concession, receiving USD 3.8 billion upfront while retaining ownership. The proceeds funded a ten year statewide road program.
Three structures cover most cases.
Operate Maintain Transfer suits an asset that is built and earning. The concession is awarded against an upfront or periodic payment, the private party takes over operations and maintenance, and the asset returns at the end of the term.
Operate Maintain Develop suits an asset that is operating but needs augmentation. The private party operates the asset and carries out the expansion during the concession, financing the work against the strength of the existing revenues. Where the asset needs upgrading before normal operation can resume, the same structure appears as Rehabilitate Operate Maintain Transfer.
The third route is through capital markets. An infrastructure investment trust pools money from many investors to hold assets with predictable cash flows. Revenue-generating assets are transferred into the trust, which issues units to investors, and the proceeds fund new projects or repay existing debt.
The legal basis on which transactions of this kind would be structured should be settled at the outset, whether under existing legislation or through a purpose-made framework. Beyond that, a program depends on four things.
A pipeline comes first. The government has to publish a list of brownfield assets, meaning assets already built and operating, that are mature, de-risked, and generating stable revenue. Isolated transactions do not attract serious investors. IBN is the natural institution to prepare and hold that pipeline, given the role proposed for it earlier in this note.
Asset-level disclosure is the requirement most likely to be underestimated. An investor pricing a concession needs to see what the specific asset earns. NEA at present publishes financial metrics at company level and does not report revenue asset by asset. Without disclosure at that level, and without the revenue stream being ring-fenced so that it can be identified and protected, an investor has nothing to price and no comfort that the cash flow will reach it.
Standard concession documents matter more here than in most contexts. Negotiation of project agreements in Nepal has been protracted, and an asset recycling program depends on doing many transactions rather than one. Model documents for each class of asset, with performance standards and measurable service levels already settled, are what make the program repeatable.
Procurement has to be open and competitive, and the use of proceeds has to be defined before the first transaction rather than after it. A dedicated fund, with published objectives and a published list of the projects the proceeds will build, is the arrangement most likely to hold public confidence. The link between the asset monetized and the infrastructure built has to be visible.
Asset monetization is not a substitute for public investment and it is not a quick source of funds. What it offers is a way to fund new transmission without new sovereign debt, while bringing operating discipline to assets the state already holds.
The constraint it addresses is specific. Transmission is the part of the system most likely to determine whether generation capacity reaches consumers, and it is the part most exposed to competition for public capital. Recycling value out of the existing grid to build the next section of it is the one financing route that does not depend on the Roadmap’s foreign investment arriving first.
Nepal has a substantial number of projects that were financed by domestic institutions through construction and are now operating assets, selling power to NEA.
In Nepal, as elsewhere, construction is where project risk concentrates. Geology, land, approvals, hydrology, cost overrun, and delay all materialize before the plant is commissioned. Once a project is built and generating, most of that risk is behind it.
This matters because it changes who will lend. A foreign lender that will not take Nepali construction risk will look seriously at a Nepali operating asset. Refinancing an operating project is therefore the point at which foreign capital is easiest to attract, and it is available now rather than after a decade of reform.
The view generally taken by government authorities is that the state need not facilitate refinancing. The risk has already been mitigated, the argument runs, and the state’s attention belongs on creating new assets.
The premise is correct. The conclusion does not follow from it.
Refinancing does not leave capital where it is. It replaces domestic debt with foreign debt and releases the domestic bank’s balance sheet. The capacity that comes back is then available to finance the next project under construction.
This connects directly to the constraint identified at the start of this note. Nepali banks are close to the limit of what the mandated energy allocation can produce, and the larger part of their annual lending capacity comes from principal repaid by operating projects rather than from growth of the overall credit base. Refinancing accelerates exactly that mechanism. Instead of waiting fifteen years for a loan to amortize, the bank recovers its capital at once and lends it again.
Refinancing is therefore a means of creating new assets, albeit at one remove, which is why it is easy to overlook, but the effect on new construction is direct.
Domestic lenders in Nepal carry a materially larger appetite for Nepali risk than foreign lenders do. That is a real asset, and the question is where it is worth the most.
It is worth most where risk is highest and where foreign capital will not go, which is construction. Foreign capital is worth most where risk is lowest and where scale and long tenor matter, which is an operating asset.
Refinancing is the mechanism that puts each in its place. It also changes the role of a domestic bank. A bank that holds a fifteen year loan to maturity finances one project with that money. A bank that originates, funds construction, and refinances out on commissioning finances several. The constraint on domestic lending is the speed at which capital turns over, and refinancing is the most direct lever on it.
There is a practical point here that determines how the measure has to be designed.
The projects best suited to refinancing are largely those financed domestically, holding a PPA with NEA and nothing else. They do not have a PDA and never did.
A refinancing framework delivered only through the PDA would therefore miss most of the assets it is meant to reach. Whatever is done has to be available to a project that has no PDA, which points toward a standalone policy or statutory treatment rather than a PDA limb alone.
A lender refinancing an operating project need most of the same assurances as a lender financing construction. It needs to know that NEA will pay, that the law governing its contract will not shift, that a change in law which alters the economics will be compensated, that its security is complete and enforceable, and that it will be made whole if the project is terminated.
The protections set out throughout this note therefore carry across to refinancing almost in full. What falls away is the construction-specific part, which is precisely what makes refinancing the easier case for the state to concede. The state is being asked for less than it is asked for on a greenfield project.
A refinancing policy, whether as a limb of the PDA policy or as an instrument in its own right, should state which projects are eligible, which of these protections are available on refinancing, and the approvals required to bring foreign debt into an existing project and to transfer security to the incoming lenders.
Of everything proposed in this note, refinancing asks the least and delivers soonest. The assets are built and operating. The lenders exist and are already willing to look at assets of this kind. What is missing is the framework that allows the transaction to happen.
It also produces a result the state wants for reasons of its own. Every refinanced project returns capital to a domestic banking system that is close to its ceiling, and that capital goes back into construction. The foreign money arrives at the safest point in a project’s life, and the domestic money is freed to do the part only it will do.
This note has dealt with energy throughout, and the reason is worth stating.
Energy is the segment of infrastructure into which the most private capital has gone towards in Nepal, as equity and as debt, and as foreign private investment in particular. Projects have been built, financed, negotiated, and in some cases have failed, and the difficulties described in this note are difficulties that real transactions have run into.
That gives the sector something the others do not yet have. The problems are visible because projects have hit them, and there is experience of them both inside government and among the advisers who have worked on those transactions. A diagnosis drawn from what has actually happened is more reliable than one drawn from first principles.
Almost everything in this note concerns how the state conducts itself toward a private investor.
How long approvals take and whether a timeline binds. How land is acquired and what the owner is paid. How forest land is released and trees are felled. What happens when the law changes after money has been committed. Who carries events outside either party’s control. What is paid when a project is terminated. Whether security can be taken over project assets and enforced by a foreign lender.
None of that is particular to electricity. A road, a transmission line, an airport, a water system, and an urban facility all meet the same authorities, the same statutes, and the same delays.
Several of these are better prepared once, for infrastructure as a whole, than separately for each sector. A land acquisition policy is the clearest case: the problems set out in this note are the problems of acquiring land for any public purpose, and a policy written for energy alone would have to be rewritten for transport within a few years. Forest land and tree-cutting approvals are the same. So, in large part, are change in law protection, force majeure allocation, termination and transfer price, and the creation and enforcement of security.
There is also a reason of principle for doing it this way. A sector-by-sector approach would reproduce exactly the fragmentation this note has identified as the underlying problem.
One thing in this note rests entirely on a feature energy has and other sectors do not.
An energy project has a buyer. NEA purchases the output under a power purchase agreement at an administratively determined tariff. Whatever else is uncertain, the developer knows before construction begins who will take the electricity and what will be paid for it. Much of the commercial risk is resolved at the outset.
A highway has no such buyer. Its revenue depends on how many vehicles use it and what they are charged, and neither is known in advance. That is demand risk, it is the central commercial question in transport, and nothing in this note addresses it.
Other countries answer it through instruments Nepal has not yet adopted: minimum revenue guarantees, availability payments under which the state pays for the asset being available regardless of use, and viability gap funding for projects that are economically justified but not commercially viable on their own. Each carries a fiscal consequence, and each requires a policy decision of its own.
Accordingly, a measure proposed here for energy will in some cases need a sector-specific counterpart rather than direct replication, and the revenue model is where that applies most sharply.
The practical conclusion is a division.
The measures that concern the state’s own conduct should be built as infrastructure-wide instruments from the beginning, even where energy is the sector driving them. Nothing is lost by drafting them that way and a great deal of later work is avoided.
The measures that depend on how a project earns its revenue need separate work, sector by sector, and that work should be done before another sector is opened to private investment rather than during the first transaction in it.
What carries across without qualification is the principle this note began with. Every risk is priced. An investor deciding whether to build a road in Nepal reads the same sovereign rating, the same country risk classification, and the same record of how the state has treated private counterparties as an investor deciding whether to build a power plant.
What the state does in energy is therefore read as evidence of what it will do elsewhere. A government that makes approval timelines binding, pays fairly for land, and honors its contracts in the energy sector has not only made energy investable. It has told every other investor looking at Nepal something about what to expect.
This table summarizes the whole note. The first column names the issue, the second states the problem, the third states the measure proposed, and the fourth identifies the instrument through which that measure would be delivered. The sections above give the reasoning; this table gives the position.
| Issue | The problem | The measure proposed | Instrument |
|---|---|---|---|
| 1. Land Acquisition | Lease of public land takes about a year through five authorities and ends at the Council of Ministers. Compulsory acquisition sits with District Administration Offices that have no capacity to spare. Statutory compensation is set well below what owners regard as fair. Court injunctions halt acquisition, and case backlogs leave the delay open-ended. Negotiated purchase attracts price increases of seven to ten times on announcement. | Shorten the public land chain to ministry level or a standing committee. Differentiate compensation by category of affected owner. Value land at market prices one year before acquisition, indexed for land price inflation. Move determination away from the District Administration Office. Offer equity participation weighted toward owners whose land is taken. Establish a dedicated forum for acquisition disputes. Pay the determined amount at once and let any dispute run separately. | Statutory change. Land acquisition policy. |
| 2. Government Approvals | Approvals are spread across many agencies and three tiers of government. Few carry binding timelines, and the clock restarts whenever further documents are called for. A developer can list the steps but cannot date them. The same provision is read differently by different offices and at different times, and positions are often oral and do not bind a successor. The incentives facing an officer favor the narrow reading or referral upward. | List the major approvals in the PDA, with the state compensating delay by extending deadlines and the revenue term. Bound the obligation to named approvals, starting only on a complete application, with a short window for the authority to call for missing documents. Set statutory timelines within which approvals must be granted or refused. Require written reasons for a refusal. Designate a body other than a court to review refusals. | PDA policy. Statutory change. |
| 3. Environmental Studies, Forest Land and Tree Cutting | Approval of an environmental impact assessment takes six months to a year, and the delay is at the point of decision rather than in the route. The cause is reviewing capacity. The regulation does not state what a study must contain, so reports are returned for revision repeatedly. Forest land and tree cutting approval passes through the Department of Electricity Development and the Division Forest Office and ends at the Council of Ministers. | Resource the reviewing ministry to the volume of work. Publish a checklist-based guide to the preparation of environmental study reports. Move final approval of forest land use and tree cutting below Council of Ministers level. | Administrative action. Published guidance. Statutory change. |
| 4. Sovereign Guarantee | Repayment depends on NEA paying over twenty to thirty years, and the exposure extends to energy payments, arbitral awards, and NEA’s obligations if it does not continue in its present form. Guarantees have been given to some projects and refused to others with no published basis. A developer cannot know before committing pre-investment whether one will be available. A guarantee is a contingent liability counted against the state, so it cannot be extended to every project. | Publish a sovereign guarantee policy setting eligibility criteria, the obligations covered, a ceiling on total guaranteed exposure, and a point early enough for a developer to establish eligibility before it spends. Deliver the guarantee through the PDA. Apply a sunset provision. | Sovereign guarantee policy. PDA policy. |
| 5. Foreign Currency Exchange Risk | A foreign lender lends in dollars and is repaid in dollars, while the project earns rupees under a tariff that does not adjust for the exchange rate. The rupee has lost 45 percent of its value against the dollar since 2010, approximately 4 percent a year and in one direction. Dollar-denominated PPAs were discontinued in 2018. The 2019 Hedging Regulation proved unusable, and the 2021 Regulation that replaced it envisages a hedging entity that has never been established. Cover purchased in the international market raises project cost by five to six percent. | Consult the institutions capable of acting as hedging entities before revising the regulation further, establishing what they can carry, at what price and on what conditions. Then settle expressly where the risk is to sit: with NEA through a dollar-denominated or dollar-indexed PPA, with a hedging entity against a fee, or with the developer through cover purchased internationally. | Hedging Regulation. PPA policy. |
| 6. Governing Law and Dispute Resolution | There is no policy, so the question is argued afresh on every project and meets the greatest resistance. Much of that resistance rests on a misunderstanding, since a governing law clause does not displace Nepali regulatory law. Contract principles have not been tested here for projects of this scale, and it is undecided whether a Nepali court would treat a termination payment as an unenforceable penalty. A forum belonging to one party does not read as neutral to the other. | State the government’s position in the PDA policy: governing law, seat, institutional rules, number of arbitrators, and language. Record expressly that a governing law clause does not displace Nepali regulatory law. Differentiate by project size or by the presence of foreign debt. | PDA policy. |
| 7. Force Majeure | Cover for political events is unavailable, or priced at a level that moves project cost materially. Cover for natural events is available, but insurance restores cost and not time. Different projects have received different treatment with no published basis. | State in the PDA policy the distinction between political and natural events and the relief attaching to each: cost and time for political events, time alone for natural events where cover exists. | PDA policy. |
| 8. Change in Law Protection | A project is priced once against the laws in force, and the state retains the power to change them. A loan is not forgiven because the law changed, so the cost falls on parties that could not have foreseen it. Statutory protection reaches the project company but not contractors, lenders, or shareholders, including on withholding tax on dividends. Changes that raise construction or operating cost are not covered. | State in the PDA policy which projects are protected, which parties are protected, and which changes are covered. Use a deductible to determine which changes count and a basket to determine when the protection is triggered. Reimburse the cost of compliance without excusing compliance. | PDA policy. |
| 9. Termination and Transfer Price | A power project’s value rests on its license and its PPA, and without them the asset cannot be sold. With no defined transfer price, a lender’s security can be worth nothing on the day it is needed. Lenders fund seventy to eighty percent of cost for an interest margin alone, so downside protection is the condition of their lending. Nepali practice relies on corporate and personal guarantees, which foreign lenders will not take. The treatment of company default is contested in every negotiation. | State in the PDA policy the transfer price payable for each category of termination event and the basis on which it is calculated. Pay properly for what the state receives, and aim for a back-to-back onward sale. Extreme cases, including total destruction of the asset, remain unresolved and require further legal and financial work. | PDA policy. |
| 10. Security Creation | Land held above the Land Act ceiling cannot be mortgaged, so the project land and the works on it fall outside the security package. Exemption above the ceiling can require a Council of Ministers decision. There is no clear provision governing transfer of the generation license when a lender exercises substitution, and no mechanism for assigning forest land and public land leases. Lending therefore falls back on corporate and personal guarantees. | State in the PDA policy the security package the state will facilitate and the direct agreement it will enter. Permit security over land held above the ceiling by a licensed project. Confirm the position of foreign lenders and security trustees. Provide statutory grounds for substitution, license transfer, and lease assignment, so that they reach projects without a PDA. | Statutory change. PDA policy. |
| 11. Right of Way | The law recognizes ownership and possession but does not say what kind of right a right of way is. Agreements run in perpetuity while the Civil Code caps infrastructure leases at forty years. Compensation at fifteen to thirty percent of land value does not reflect the loss, since the land cannot be built on, is limited in agricultural use, cannot be mortgaged, and falls in market value. This is the principal cause of local resistance to transmission projects. | Create and define the right of way as a distinct legal right over land, stating what it permits, what it restricts, how long it lasts, how it is registered, and how it transfers. Reconsider the basis on which compensation is fixed. | Statutory change. |
| 12. The Role of the Investment Board Nepal | IBN undertakes in PDAs to procure government approvals but has no power to direct the authorities that hold them. Its mandate has outrun its staffing, advisory capacity, and ability to retain specialists. Federal mandates overlap, and foreign investment approvals now route through the Department of Industry. Projects IBN has approved are still escalated to Cabinet. Jurisdiction is fixed by rigid thresholds. No policy defines what IBN may offer. The market reads IBN as an additional layer in the approval chain. | Empower IBN to direct the responsible authority rather than to issue approvals itself, with its own power limited to ensuring applications are examined in time. Give the single window authority to direct, track, and escalate. Codify the end of secondary Cabinet clearance and make IBN approval binding for ancillary clearances. Reimagine IBN’s role in land acquisition, covering policy, price determination, a dedicated approval unit, compensation distribution, and clearance coordination. Publish the PDA policy IBN negotiates from. Address staffing and retention. | Statutory change. PDA policy. Administrative action. |
| 13. Asset Monetization | Public debt has reached 44.69 percent of GDP, and the capital budget falls far short of the Roadmap requirement. A large stock of mature public transmission assets sits alongside an expansion pipeline competing for the same public capital. NEA reports at company level and not asset by asset, so an investor has nothing to price. Private participation in transmission remains limited. | Publish a pipeline of mature brownfield assets, prepared and held by IBN. Require asset-level financial disclosure and ring-fencing of revenue. Issue standard concession documents by class of asset. Procure openly and competitively. Define the use of proceeds in advance, preferably through a dedicated fund. Settle the legal basis before the first transaction. | Program policy. Legal basis to be settled. |
| 14. Refinancing | Operating projects financed domestically hold capital on bank balance sheets that could otherwise fund construction. The view generally taken is that refinancing needs no state facilitation because construction risk has already passed. Most candidate projects hold only a PPA and have no PDA, so a remedy delivered through the PDA would not reach them. | Adopt a refinancing policy stating which projects are eligible, which of the protections set out in this note apply on refinancing, and the approvals required to bring foreign debt into an existing project and to transfer security to incoming lenders. Make it available to projects without a PDA. | Standalone policy or statutory change. |
