A Briefer on the IMF’s Resilience and sustainability facility for pakistan
There is a central and well-recognized tension at the heart of the International Monetary Fund’s climate based lending through its Resilience and Sustainability Facility (RSF): the tension between debt and climate. Pakistan has been involved in successive programs under the Fund since 2019, ranging from Stand By Arrangements (SBA), an Extended Fund Facility (EFF), and now the RSF. The current EFF, approved in September 2024, targets IMF’s longstanding reform priorities of privatisation, tariff reforms, higher sales taxes, and subsidy reform via targeted transfers; remaining within a grid protectionist frame. In May of 2025, the IMF also approved a 29-month RSF to target climate resilience and decarbonisation. These “climate” priorities were to be addressed through disaster-shock buffering, green finance and taxonomy reforms; which included a carbon levy, electric-vehicle incentives, targeted power subsidies. Such priorities were tagged to the helm of energy efficiency and mitigation efforts.
The RSF however, ends up introducing climate under a newer macrofiscal risk category present in the conventional adjustment framework centered on fiscal consolidation and revenue generation. Policy-level conditionalities such as the Petroleum Development Levy, the Carbon Support Levy, and the removal of electricity subsidies that protect the most vulnerable households are no longer presented as fiscal consolidation. They are re-framed as climate-mitigating measures under the guise of reducing inefficiencies, reducing incentive to consume petroleum (in spite of its inelastic nature), and reducing wasteful consumption of electricity and water.
This article argues that the fiscal benefits of energy subsidy reform are overstated in the short and the long term, and the mechanisms meant to replace subsidies are inept at ensuring vulnerable sections of the population achieve affordable energy, undermining the Fund’s just transition goals in tandem with Pakistan’s Nationally Determined Contribution (NDC) targets. Under Reform Measure 12 of the RSF, the Fund argues that Pakistan's power subsidies are “poorly targeted”. Resulting in over-consumption in upper income quintiles, which leads to Pakistan’s high emissions state. As a solution, the RSF supports transferring power subsidies from the budget and tariff structure to budget-financed, targeted subsidies delivered through BISP – aiming at fiscal savings and effective social protection in one fell swoop. This will support the poorest forty percent of energy consumers while wealthier consumers will face higher tariffs, providing an incentive to “conserve energy.” The causal chain is identified from the under-pricing of electricity, which reduces incentives for more efficient electricity generation, echoing more than a decade of World Bank reporting that universal energy subsidies cause wasteful consumption by upper-income households. To reduce circular debt, the IMF has called for reducing subsidies to 0.6% of GDP, eliminating the tariff differential subsidy (TDS) that protects vulnerable consumers from high energy prices, and privatising distribution companies.
According to the Fund, a total domestic electricity subsidy of Rs 521 billion is treated as fiscally unsustainable given the emphasis on primary balance consolidation, with the TDS alone costing Rs 249 billion a year. The Rs 272 billion annual cross-subsidy from consumers above 300 units is blamed for distorting price signals and keeping industrial tariffs uncompetitive. Around 73% of power sector costs are fixed, while only about 7% of revenue was recovered through fixed charges before the reform. And the protected consumer class grew from 28% of domestic consumers in 2022 to 59% in 2026, which the Fund reads as subsidies leaking to households that keep grid consumption low through solar and metering adjustment.
However, this diagnosis leaves out the role of the IMF in building the existing unjust cost structure and the subsequent circular debt problem. The IPP policies of the 1990s, developed through extensive involvement of the Fund and the World Bank, continue to weigh down the energy sector. Capacity payments, which contribute heavily to the circular debt of the gas and energy sector, are often left out of the analysis, which focuses instead on low recoveries and high levels of consumer-end subsidies.
Capacity payments remain 61% of Pakistan's power purchase cost. In FY2024–25, a capacity payment price of Rs 14.3 per unit made up more than half of the new Rs 25.29 tariff; adding the Rs 3.23 per unit debt service surcharge, 69.6% of the tariff consists of the grid's structural inefficiencies.
The current conditionality aims to stop the accumulation of circular debt, but it does not target legacy contracts with independent power producers. For the ordinary grid consumer, those contracts meant capacity payments even for non-performance, sovereign guarantees, and dollar-indexation in a devaluing currency: in effect, a bundled tariff paid in an 'unbundled' regime.
The first phase of the reform arrived through NEPRA's tariff rationalisation in February 2026. It replaced the flat consumer charge for high users with a capacity-based fixed charge of Rs 200–675/kW/month, applied fixed charges of Rs 200–300/kW/month to protected consumers for the first time, and cut variable rates by Rs 0.49–1.53/kWh only for households consuming above 300 units.
The bottom 70% of consumers, those using 300 units or fewer, receive no variable rate relief but now pay new sunk fixed charges. A protected household consuming 80 units saw its bill rise from Rs 843 to Rs 1,043 (+23.7%), while a household above 700 units saw its bill fall by roughly Rs 6,700 (−17.1%). Some 20.2 million protected households now pay the fixed charge regardless of how little they consume, so careful consumption management no longer lowers the bill.
This is where the climate rationale collapses. The bottom 70% of consumers, those using 300 units or fewer, receive no variable rate relief but now pay new sunk fixed charges. A protected household consuming 80 units saw its bill rise from Rs 843 to Rs 1,043 (+23.7%), while a household above 700 units saw its bill fall by roughly Rs 6,700 (−17.1%). Some 20.2 million protected households now pay the fixed charge regardless of how little they consume, so careful consumption management no longer lowers the bill. The reform punishes the behaviour it was designed to reward, as the reform incentivises overconsumption from higher quintiles as their variable rate decreases, shifting the burden further onto lower income households. The planned rollback of the TDS compounds this. At full removal, the effective rate for lifeline consumers using up to 50 units rises from Rs 3.95/kWh to the Rs 36.13/kWh cost of service, a bill increase of around 815%; a protected household on 80 units moves from Rs 843 to Rs 2,890 a month (+243%). Consumers above 300 units, already paying near or above cost of service, see bills fall by 8–28%.
Fiscal savings are often cited as the reasoning behind TDS removal, but these savings are overstated, as they assume static grid conditions, discounting solar exit and assuming constant demand and collection. A modelling of 2,304 scenarios was done, varying the share of TDS realised as savings, recovery-rate deterioration, BISP coverage, the size of the compensating transfer, and administrative cost. Only 312 (13.5%) produce a net fiscal gain; the median outcome is a net loss of Rs 184.65 billion, and the worst case a loss of Rs 658 billion.
The scenarios that produce a positive fiscal outcome share a specific and demanding profile: realised TDS savings above 80%, which would result in massive tariff hikes; recovery rate deterioration below 1 percentage point; BISP coverage held to 30% rather than the 40% poverty headcount; compensation at the floor of PKR 1,000/month ($3.6/month); and administrative costs at 3.3%. Under these conditions, the net fiscal gain is PKR 115.7 billion ($416.9 million), roughly 46% of the gross TDS saving. Under the current assumptions of this policy shift, fiscal gains necessarily precede and take precedence over welfare coverage. The conditions necessary for fiscal viability are the same conditions that maximise welfare loss.
The proposal is to use the Benazir Income Support Programme (BISP) as a targetted subsidy transfer. While the BISP is an important and progressive initiative, but has never fully reached those below the poverty line: this year it reached only 16 million individuals. By design, it also does not cover those just above the poverty line who remain financially vulnerable.
Expanding BISP to 30% of the population with a Rs 1,000/month electricity top-up costs an incremental Rs 141 billion, more than half the Rs 249 billion TDS budget, before any collection shortfall, and still leaves most of the bill increase uncompensated. At Rs 2,000/month and 40% coverage, the cost rises to Rs 384 billion, a net fiscal loss of Rs 284 billion. Even Rs 2,000 falls short: full TDS removal raises the protected 0–100 unit bill by Rs 2,047 a month. A package that actually protects the poorest is fiscally ruinous; one that is fiscally manageable leaves most of the welfare loss uncompensated.
BISP's proxy means test was predicted by the World Bank to produce exclusion errors of up to 88% for the poorest decile; observed inclusion and exclusion errors have reached 73.6% and 52.3% respectively. The near-poor households consuming 200–400 units, sit above the PMT threshold of 32 but lack the buffers to absorb sustained tariff increases. Transfers are disbursed quarterly against monthly bills, for households already spending around 13.8% of income on electricity, so the predictable result is arrears, informal borrowing, or cuts to food, health and education. On implementation, the authorities, working with the World Bank, aim to finish linking electricity consumers to the National Socio-Economic Registry, along with validity checks, by end-November 2026, and will only then determine eligibility criteria. This is a slip from the original commitment to define eligibility criteria by end-July 2026.Initial rebates are still due by end-January 2027. Verification so far has run through Power Division QR-code registration printed on bills, which checks consumers against BISP and NSER data, allows one subsidised meter per CNIC, and requires the name on the bill to match the CNIC exactly, following earlier proposals for biometric verification at financial institutions before the subsidy is deducted from the bill. These methods are out of touch with gaps in on-ground access; NADRA registration and verification remains a major barrier to qualifying recipients.
The other route out of high grid tariffs has been the solar rush, which has seen the country import over 50 GW of solar panels over five years valued at $18 billion. But exit is not equally available. HIES 2024–25 shows 16.4% of households in the poorest consumption quintile own solar, mostly as solar-only systems (11.8%) averaging 1.5 panels; in the richest quintile, 31.8% own solar, mostly as grid-hybrid systems (16.9%) averaging 4 panels. For poorer households, solar is a coping mechanism on a weak or absent grid; for richer households, it is optimisation. At Rs 110,000–120,000 for a 1 kW system, lifeline and protected consumers are structurally grid-captive.
The policy response has been to penalise exit rather than address the conditions that make it rational. Under the EFF and RSF-backed budget, a 10% sales tax was placed on imported solar panels, reduced from an initially proposed 18% after backlash. The net metering buyback rate was cut in December 2025 from Rs 25.98 to Rs 11 per unit, stretching payback for new installations from 3–5 years to 10–12 years. Draft prosumer regulations would have DISCOs buy solar exports at roughly Rs 8–9 per unit while selling grid power at roughly Rs 35. Yet net-metered exports in FY2024–25 were 3.28 TWh, only about 1% of national generation; the narrative of solar "free-riding" is not borne out by the data.
This is why the savings are overstated in the long run too. Grid demand contracted by 3.6% in FY2024–25, with agricultural consumption falling 18% as farmers moved to solar irrigation.³⁰ As better-off consumers defect, fixed grid costs are recovered from a shrinking, poorer base that cannot exit, raising tariffs and accelerating the next round of defection. The result is not grid protection but regressive cost shifting, while the legacy contracts that drive costs go untouched.
The recent solar boom, sustained alongside a glut of Regasified-LNG from ill-planned long-term contracts, exposes gaps in the IMF's planning paradigm that helped lock Pakistan into fossil-fuel dependence. Instead of addressing these gaps directly, the programme accommodates greater use of expensive RLNG in national energy planning, while taxing solar panels imported from China. The closure of the Strait of Hormuz has pushed momentum toward domestic coal reserves over renewable options, and the IMF's most recent review is silent on these developments. Nor has the programme absorbed a post-floods world: Nor has the programme absorbed a post-floods world: [V] the 2023 SBA effectively turned a blind eye to the impact of the 2022 floods, and the Fund's DIGNAD stress test applies a 5%-of-GDP shock only after a five-year period of growth and investment, leaving out delayed-onset climate impacts.
Country programme findings
Energy-sector conditionality under the EFF focuses on tariff cost recovery, DISCO privatisation, and the migration of captive power users to the grid. Energy-sector conditionality under the EFF focuses on tariff cost recovery, DISCO privatisation, and the migration of captive power users to the grid. The RSF layers on a climate support levy, initially set at PKR 2.5 per litre and scheduled to rise to PKR 5 per litre, framed as reducing emissions and dependence on imported fossil fuels, alongside power-subsidy retargeting through BISP under RM12. alongside power-subsidy retargeting through BISP under RM12. [V] The Fund's CPAT modelling claims the petroleum and carbon levies would reduce projected 2030 CO2 emissions by 17 percent from the no-policy baseline. Yet crude oil imports rose by 11.3%, or 9 million barrels. In a country where most citizens rely on fuel-dependent vehicles, especially motorbikes, petroleum is highly price-inelastic, and the levy acts as a regressive and extractive source of revenue.
National taxation policy has followed the EFF-RSF strategy of raising revenue through energy. The Petroleum Development Levy has driven non-tax revenue from about PKR 834 billion to over PKR 1.2 trillion, projected to reach PKR 1.6 trillion, and is now the second-highest source of non-tax revenue after State Bank profits. Because it is classified as non-tax revenue, it falls outside the divisible pool under the NFC Award, leaving the federal government with the entire take, readily available for sovereign debt servicing rather than shared with provinces or dedicated to climate action. On the clean energy side, the 10% sales tax on solar panels hampers the transition in order to achieve marginal revenue increases.
Reform Measure 3 of the RSF focuses on federal green budgeting by enhancing budget transparency and extending green budgeting practices to the provincial governments. In contrast to these provincial and federal green targets, the PSDP has been cut from PKR 1.096 trillion (FY25) to PKR 820 billion (FY26), though this was exceeded by PKR 92 billion in FY 26; PKR 10 billion allocated for Thar coal; offshore oil and gas exploration resumed after two decades.
The programme treats energy-sector debt as a problem of under-recovery rather than of contracts, imports and systemic. Capacity payments to legacy IPPs, the largest single component of power purchase cost, are left outside conditionality. Import dependencies persist: petroleum products made up 19.5% of total imports, costing $11.5 billion in July–March FY26 and long-term RLNG contracts continue to shape planning. These sectoral liabilities sit within a national budget where debt servicing accounts for nearly half of total budgeted expenditure.
Conclusion
The RSF adds climate language and selected climate-relevant reforms, but it does not stand apart from the austerity and restructuring already being imposed through the EFF. Instead, it largely extends the same planning paradigm into climate policy. In energy, subsidy reform, DISCO privatisation, carbon and climate levies are framed as climate measures, while their distributional consequences and the persistence of fossil-fuel lock-in receive little systematic attention
On the fiscal front, the savings from subsidy removal are real but overstated: solar exit erodes DISCO revenue, affluent consumption rises with cheaper marginal units, and the state cannot meaningfully compensate welfare losses through BISP without eroding the fiscal space the reform was meant to create. On the climate front, the reform rewards consumption at the top, taxes the cheapest available exit at the bottom, On the climate front, the reform rewards consumption at the top, taxes the cheapest available exit at the bottom, and leaves the drivers of high tariffs, overcapacity, guaranteed returns and fossil lock-in, untouched. A reform that wanted affordable, low-carbon electricity would start with the cost of the grid itself: capacity payments, legacy contracts and fuel-import exposure.
