25 Power sector

Chapter 25 of 60 in the Bangladesh 2036 research base. Contents of the series.

Thesis

The power sector is the megaproject machine's largest exhibit, and the stored record shows the build succeeded while the economics inverted: derated capacity nearly quintupled from 5,271 MW in 2010 to 24,171 MW in 2023 [BPDB 2023] while maximum demand grew from 6,454 to 17,100 MW [BPDB 2023], leaving the system running at 41 percent utilisation [BPDB 2023], paying capacity charges to idle plants, importing 10,425 million kWh across the borders in FY23 [BPDB 2023], and carrying an operating loss of 6,131 crore BDT with 192.7 days of payables [BPDB 2023]. The claim this chapter defends is that the power decade to FY36, the horizon the chapter 15 scenarios assume, is a contract-repricing problem wearing a generation-problem costume: the wires and plants exist, the demand exists, and the binding constraint is the stock of take-or-pay contracts, the fuel price pass-through and the tariff path that determine whether the sector's losses stay inside a subsidy line or crystallise as the debt chapter 42 prices.

Where Bangladesh stands

The capacity record is the boom in one series. Derated capacity rose from 3,718 MW in 2007 to 5,271 in 2010, then compounded: 8,100 in 2012, 12,771 in 2017, 15,410 in 2018, 21,280 in 2021, 24,171 in 2023 [BPDB 2023]. Maximum demand followed, but slower: 6,454 MW in 2010, 11,200 in 2018, 17,100 in 2023 [BPDB 2023]. The gap between the two curves is the idle fleet: installed capacity of 24,911 MW on the panel chapter 10 cites [BPDB 2023] against a maximum served of 15,648 MW, and a capacity utilisation of 41.0 percent in FY23 [BPDB 2023]. Chapter 10 frames the paradox; this chapter prices it: every idle megawatt under a take-or-pay contract earns its capacity charge, and the payments the largest generator bills, 2,062.77 crore of capacity payments against 1,705.00 crore of energy payments in FY25 [NWPGCL 2025], are the sector's arithmetic in miniature. The idle fleet is not waste in the engineering sense, it carries the reserve margin that the FY23 fuel crisis showed the system needed, but it is a liability in the accounting sense, and the decade's task is to convert an expensive reserve into a cheap one through contract structure rather than concrete.

The demand and supply record below the headline is healthier than the capacity gap suggests. Energy sales reached 84,450 million kWh in FY23 from 76,323 in FY21 [BPDB 2023], and per capita generation climbed from 426 kWh in 2020 to 517.89 in 2023 [BPDB 2023], a level still far below the region's middle-income norm and the demand headroom the decade's growth assumes. Distribution system loss fell to 7.92 percent in FY23 from 8.5 in FY21 [BPDB 2023], the slow efficiency gain that makes every sold kWh cheaper to deliver. The generation mix moved as chapter 10 records: private plants generated 46,104.5 GWh in 2023 [BPDB 2023], the majority share, and the gas burn fell from 267,768 million cubic feet in 2020 to 227,924 in 2023 [BPDB 2023] while generation rose, the fuel switch toward imported oil, coal and LNG that chapter 26 prices.

The import record is the new dependency. Power imports reached 10,425 million kWh in FY23, up from 7,712 in FY22 [BPDB 2023], the cross-border flows from India that include the dedicated Godda plant's contracted supply; the Adani contract's terms, price formula and take-or-pay structure could not be confirmed; the Power Division and the contract disclosures would resolve them. Imports now supply a double-digit share of the energy the system delivers, and their currency denomination ties the retail tariff to the exchange rate path chapter 05 documents [BIS 2026].

The financial record is the loss the boom accumulated. BPDB's operating loss of 6,131 crore BDT in FY23 [BPDB 2023] sits alongside 192.7 days of accounts payable [BPDB 2023], the payables being the subsidy-and-cash-flow gap made visible: the utility buys dollars-priced fuel, sells taka-priced power, and waits for the budget to reconcile the difference. The tariff adjustments of the transition period, the retail price steps of 2023 whose full schedule could not be confirmed here, the BERC documents being the resolving source, narrowed the gap, and the sector's subsidy line inside the chapter 04 budget remains the reconciliation mechanism. The unserved energy collapse chapter 10 dates, 3,818 million kWh in FY23 against 84 the year before [BPDB 2023], is the record's proof that the constraint was fuel and cash, not capacity.

The tariff and subsidy record is the sector's fiscal face, and the stored pieces frame it without the full schedule. BERC's retail adjustments arrived in steps during 2023, each step narrowing the cost-recovery gap whose residual is the subsidy line chapter 04 carries, and the full step schedule could not be confirmed here. The fuel side of the tariff is the dollar side: the sector's marginal cost moved with the LNG spot price, the coal price and the exchange rate together, and the retail tariff moved in taka steps months behind, which is the 192.7 days [BPDB 2023] made. The subsidy arithmetic the budget carries peaked with the 2022-23 price cycle, the fossil subsidy measure chapter 04 cites at 7.39 percent of GDP in 2022 [UNSD 2026] being the widest definition, and the power-specific line inside it is the gap the tariff steps are closing. The chapter's decade claim is that the steps must sum to cost recovery before the contract stock matures, because a repriced contract under an unrecovered tariff simply moves the loss from the capacity payment to the retail gap, and the retail gap is where the political economy of tariffs has always been won or lost.

The demand composition is the sector's growth map. Energy sales of 84,450 million kWh [BPDB 2023] split across categories whose shares could not be confirmed at the stored level, the BPDB category tables being the resolving source, but the drivers are ordered: industry, the garment clusters chapter 17 maps and the SME tier chapter 22 counts, residential demand riding the electrification near-universal coverage chapter 01's access series records [WB WDI 2026], irrigation from chapter 16's boro machine, and commercial cooling load that the heat record chapter 11 documents compounds every summer. The peak-demand series, 17,100 MW in 2023 [BPDB 2023], grows with the cooling load as much as with the industrial one, and the evening peak's solar-free timing is the integration problem chapter 27 owns.

Mechanism

The mechanism that built the overbuild was the quick rental legacy compounding into contract stock. The 2009-2016 shortage was answered with rental and IPP plants on short-tenor, high-tariff, take-or-pay contracts, the state bought reliability with option value, and when demand growth slowed after 2019 the options matured into obligations: capacity charges became the sector's fixed cost, the idle fleet's price. The 2,062.77 crore [NWPGCL 2025] capacity-versus-energy inversion is the stock's signature, and the contract expiries running through the decade are the repricing window, and every expiry that lapses without renewal is a capacity charge the tariff never has to carry again. The mechanism's second gear is the fuel switch: domestic gas production fell from its 981.8 bcf peak to 711.4 bcf [EIA 2025], the gas burn data above shows the system's gas engines demoting to swing duty, and oil and LNG imports filled the gap at dollar prices, which is the transmission channel from the exchange rate to the retail tariff.

The third gear is the distribution utility's cash machine. The 192.7 days of payables [BPDB 2023] are the sector's ledger of unpaid bills and unreconciled subsidies; the receivables from the distribution utilities and ministries, and the arrears to the IPPs, form the circular debt pattern the region knows, and its Bangladeshi size as an aggregate could not be confirmed from the sources this chapter draws on, the BPDB audited accounts being the resolving source. The tariff mechanism is the reconciliation's price: BERC adjusts retail rates in steps, each step is politically priced, and the gap between cost and tariff is the subsidy chapter 04 measures and the arrears chapter 41 details.

The fourth gear is the nuclear and import bet. Rooppur's 94,689.26 crore spent of 138,685.77 crore [Planning Commission ADP 2025] with zero generation through 2025 [Ember 2025] is the sector's largest single capital commitment, and its synchronisation, whenever it arrives, adds baseload capacity the system's surplus already holds, which is why the tariff and contract repricing must happen before the plant lands, not after. The Indian import lines, 10,425 million kWh [BPDB 2023], are the reverse bet: capacity abroad on contract, paid in dollars, delivered reliably.

The cross-border power mechanism deserves its own accounting because it is the sector's newest structural layer. The 10,425 million kWh imported in FY23 [BPDB 2023] includes the dedicated Godda supply that chapter 45's India account prices geopolitically, the older grid links through Bheramara and Cumilla, and the Nepali hydro entry whose first deliveries mark the regional trade turn chapter 27 anticipates. The imports' contract terms, dollar-denominated, take-or-pay elements, fuel-indexed, could not be confirmed here, and their domestic consequence is direct: every imported kilowatt-hour is a retail tariff liability at the exchange rate of the settlement day. The sector's fuel security question and its tariff question are the same question wearing two denominations.

The decade ahead

The forces the FY36 scenarios assume are four. First, the contract repricing cycle: the rental and IPP contracts' expiries through the window are the one instrument that fixes the arithmetic without new money, and the NWPGCL comparison [NWPGCL 2025] is the template; the revealing indicator is BPDB's loss line falling below 4,000 crore, chapter 10's threshold. Second, the demand growth: per capita generation at 517.89 kWh [BPDB 2023] has years of catch-up demand in it, and the 6.5 percent growth path chapter 15's reform scenario assumes raises peak demand faster than the fleet can idle gracefully. Third, the fuel path: the gas decline chapter 26 details and the coal and LNG import shares set the sector's marginal cost, and the energy security question chapter 26 owns is this chapter's cost curve. Fourth, the renewable obligation: the 2.14 percent share [Ember 2025] against buyer decarbonisation demands chapter 17 faces makes the grid integration question chapter 27 owns a power sector variable.

No stored power-sector projection targeting 2036 exists in the sources this chapter draws on beyond the older master plan chapter 10 cites as stale [Power Division 2023]; all FY36 figures are not established, and the named authors whose work would resolve them are the Power Division's integrated planning successor, the IMF programme's sector reviews, and the JICA and World Bank sector assessments. The decision points: the contract repricing sequence, the tariff path's honesty, the Rooppur synchronisation disclosure, and the import contract terms' publication.

The jobs and transition layer closes the sector's social account. The coal plant buildout chapter 10 records at Matarbari, Rampal and Payra, the rental fleet's gas-oil engines and the solar rooftops chapter 27 counts employ different workforces on different horizons, and no stored series tracks the sector's employment by technology, the Power Division establishment tables being the resolving source. The transition question the buyer economies chapter 17 faces, decarbonised supply chains, lands on this sector as a demand for a renewable share and a coal exit path, and the 2.14 percent renewables reading [Ember 2025] is the gap those buyers will price. The just-transition financing that the climate chapters 11 and 27 frame is therefore a power sector fiscal question before it is an environmental one.

Risks and upside

Risks. First, the capacity-payment spiral: if demand growth disappoints while contracts renew on old terms, the 41 percent utilisation [BPDB 2023] falls and the loss compounds past 8,000 crore, chapter 10's failure mark [BPDB 2023]; the revealing indicator is the loss line and the utilisation pair. Second, the exchange-rate pass-through: fuel and import contracts priced in dollars against a taka that chapter 05 shows moved 43 percent [BIS 2026] put the retail tariff on a collision path with the inflation regime, and the revealing indicator is the tariff step schedule, which could not be confirmed here. Third, the circular debt crystallisation: the 192.7 days of payables [BPDB 2023] growing into a stock that forces IPP defaults would convert a utility problem into a banking exposure chapter 06 inherits; the revealing indicator is the payables days series.

Upside. First, the repricing dividend: chapter 10's threshold, a durable loss below 4,000 crore [BPDB 2023], releases fiscal space larger than the ADP's social sector annex. Second, the reliability dividend: the unserved energy line returning below 500 million kWh [BPDB 2023] is the growth input chapter 01's investment stall names as the business climate's first constraint. Third, the import leverage: the 10,425 million kWh [BPDB 2023] of contracted imports is reliability the system did not have to build domestically, and negotiating its terms alongside the domestic contract stock is the combined repricing the decade requires, with the exchange rate as the price both legs of the sector share.

What to watch

1. BPDB operating loss. Current value 6,131 crore BDT in FY23 [BPDB 2023]. Threshold: a durable fall below 4,000 crore is the repricing proof; persistence above 8,000 marks the contract stock unreformed.
2. Capacity utilisation. Current value 41.0 percent in FY23 [BPDB 2023]. Threshold: a rise above 50 marks the overbuild digesting; a fall below 35 marks the overbuild deepening.
3. Power imports. Current value 10,425 million kWh in FY23 [BPDB 2023]. Threshold: growth above 15,000 million kWh without published contract terms marks the import dependence regime; a stable share with published terms marks the managed dependency.
4. Distribution system loss. Current value 7.92 percent in FY23 [BPDB 2023]. Threshold: below 6 percent marks the efficiency turn; above 10 marks the network's maintenance deficit.
5. Per capita generation. Current value 517.89 kWh in 2023 [BPDB 2023]. Threshold: sustained growth above 6 percent a year marks the demand catch-up the reform scenario assumes; stagnation marks the demand-side constraint chapter 01's stall names.

Sources used

[BPDB 2023] Bangladesh Power Development Board annual report panel and financial series via bdpolicy.db, series: bpdb_derated_capacity_mw, bpdb_maximum_demand_mw, bpdb_capacity_utilization_pct, bpdb_energy_sales_mkwh, bpdb_per_capita_generation_kwh, bpdb_distribution_system_loss_pct, bpdb_private_gen_gwh, bpdb_fuel_gas_mmcft, bpdb_power_import_mkwh, bpdb_power_capacity_generation_panel, bpdb_financial_performance_2023. [NWPGCL 2025] Northwest Power Generation Company Limited audited accounts, capacity and energy payments, cited from chapter 10. [EIA 2025] US Energy Information Administration via the energy/eia_international_bd parquet, series: eia_international_bd. [Ember 2025] Ember yearly electricity data via the climate/ember_yearly_bd parquet, series: ember_yearly_bd. [Planning Commission ADP 2025] RADP FY2025-26 project table via bdpolicy.db adp_projects, Rooppur rows. [Power Division 2023] Power System Master Plan status, cited from chapter 10. [BIS 2026] Bank for International Settlements USD/BDT series via bdpolicy.db, series: bis_usd_bdt_monthly_eop. [UNSD 2026] UN Statistics Division, SDG indicator 12.c.1 fossil fuel subsidies as a proportion of GDP, via bdfacts snapshot, series: SDG.12.ER_FFS_CMPT_GDP, cited from chapter 04. [WB WDI 2026] World Bank World Development Indicators, series: wb_electricity_access. [Not confirmed, chapter 25] Adani contract terms, retail tariff step schedule, circular debt aggregate, integrated plan successor; resolving sources Power Division, BERC, BPDB audited accounts.

---

Verified line by line against primary sources: 27 claims checked, 0 corrected.

Previous: 24 Transport megaprojects and PPP

Next: 26 Gas, LNG and fuel security

All sixty chapters and the ten narrative chapters.