The Smallest Energy Budget: A Manufacturing Economy on 287.8 Kilos of Oil Equivalent

Executive Summary. Bangladesh has the lowest energy use per person of any economy in its comparison group, and it is not the poorest of them. Energy use was 287.8 kg of oil equivalent per person in 2023, last of six Asian economies and 29% of the peer median of 982.4. Thailand used 6.4 times as much. Income does not account for the gap: at 9,647 international dollars per person at PPP in 2024 Bangladesh ranks fifth of the six, above Pakistan's 6,252, and yet Pakistan uses 415 kg of oil equivalent per person. Industrial structure does not account for it either. Manufacturing was 22.44% of GDP in 2025, third of the six and 3.38 points above the peer median of 19.07. So the picture is a manufacturing economy of 173.6 million people running on the lowest energy use per person of the six, and buying an unusually large share of that fuel abroad: net energy imports were 44.31% of energy use in 2022, second-highest of the six. Whether 287.8 records efficiency or a ceiling is the question this data cannot settle, and almost everything downstream depends on which it is.

Under a third of the peer median

In 2023 Bangladesh used 287.8 kg of oil equivalent per person. The comparison group: Thailand 1,851, Vietnam 1,018, Indonesia 982.4, India 763.2, Pakistan 415. Bangladesh is last of the six, and the distance is not a matter of a few percent. The peer median is 982.4, so Bangladesh runs on 29% of what a median comparator runs on, a gap of 694.7 kg of oil equivalent per person. Thailand, the group's highest, uses 6.4 times the Bangladeshi figure.

Source: World Bank World Development Indicators, energy use (kg of oil equivalent per capita, EG.USE.PCAP.KG.OE), 2023.

This is a physical quantity, not a financial one. It counts all primary energy entering the economy: fuel burned in power stations, gas delivered to factories and kitchens, diesel in trucks and irrigation pumps, biomass in rural households. All of that comes out of 287.8 kg of oil equivalent per person.

It is not the income ranking, and not the industry mix

The obvious explanation is that Bangladesh is poor and poor countries use less energy. The panel does not support that as a complete account.

GDP per person at purchasing power parity was 9,647 international dollars in 2024, fifth of the six against a peer median of 16,386, a shortfall of 6,739 and about 59% of the median. India 11,160, Indonesia 16,448, Vietnam 16,386, Thailand 24,712, Pakistan 6,252.

Source: World Bank WDI, GDP per capita, PPP (NY.GDP.PCAP.PP.CD), 2024.

Read those two rankings together. On income Bangladesh is fifth of six and Pakistan is sixth. On energy Bangladesh is sixth and Pakistan is fifth. Pakistan is the poorer country on this measure, 6,252 against Bangladesh's 9,647, and it still uses 415 kg of oil equivalent per person against Bangladesh's 287.8. Income alone does not put Bangladesh at the bottom of the energy table; something else does.

Nor is it that Bangladesh has stayed out of the energy-hungry business of making things. Manufacturing value added was 22.44% of GDP in 2025, third of the six, 3.38 percentage points above the peer median of 19.07 and 1.18 times it. Vietnam 24.53 and Thailand 23.74 sit above Bangladesh; India 13.47, Pakistan 13.02 and Indonesia 19.07 sit below. Bangladesh is, in proportion to the size of its economy, more of a manufacturing country than India or Pakistan, on a per-person energy budget smaller than either.

That combination is the finding. A country that makes things at that share of output, for that many people, on that little fuel, is doing something the standard relationship between industry and energy does not predict.

Electricity ranks better than total energy

On the wire, Bangladesh is not at the bottom.

Electricity consumption was 602.9 kWh per person in 2023, fifth of the six, ahead of Pakistan's 518.3 and behind India's 1,182, Indonesia's 1,445, Vietnam's 2,585 and Thailand's 2,965. Against the peer median of 1,445 that is a gap of 842.3 kWh and a ratio of about 42%.

Two things follow, and only two. First, Bangladesh's position on electricity (42% of the peer median) is less extreme than its position on total energy (29%), and on this measure Pakistan's 518.3 sits below it. Second, that pattern is what a country looks like when the grid has been extended faster than the wider fuel economy: electricity gets built deliberately, planned, financed and counted, while transport fuel, industrial heat and household cooking accumulate from millions of separate decisions.

The two series sit on different bases, primary energy in one and delivered electricity in the other, and this essay does not convert between them. What compares is each country's position within each ranking.

More than doubled since 1990, still last

The long series says the low figure is not a stable national characteristic. It is a level that has been rising the whole time.

Energy use per person went from 114.3 kg of oil equivalent in 1990 to 287.8 in 2023, an increase of 173.4 and a rise of 151.7%, taking it to 2.5 times its 1990 level. Over the same span the country's measured income per person rose from 940.7 international dollars at PPP to 9,647, a multiple of 10.26.

Source: World Bank WDI, energy use (kg of oil equivalent per capita, EG.USE.PCAP.KG.OE), Bangladesh against comparators, 1990-2023.

The income multiple needs a warning label. The WDI series used here is PPP in current international dollars, so 10.26 carries global price change as well as real growth, and it is not a statement that Bangladeshi living standards rose tenfold. The energy series carries no such problem: kilos of oil equivalent are kilos in both years. What survives the caveat is the direction. Fuel use per person has grown substantially since 1990 and Bangladesh still finishes last of the six on energy use per person. Whatever restraint the 287.8 figure represents, it has been loosening, not holding.

A small budget with a large import share

The other half of the energy story is where the fuel comes from, and here Bangladesh is near the top of the table rather than the bottom.

Net energy imports were 44.31% of energy use in 2022, second-highest of the six, behind Thailand's 57.12 and ahead of Pakistan's 39.79, India's 36.53 and Vietnam's 34. Indonesia is a net exporter at -82.92, which is why the peer median of 36.53 sits where it does. Bangladesh is 7.78 percentage points above that median and at 1.21 times it.

Bangladesh consumes the least energy per person of the six and imports the second-largest share of what it does consume. Those combine badly. A large energy budget with a large import share is a trade exposure. A small energy budget with a large import share is a trade exposure with nothing left to trim, because a system running on 287.8 kg of oil equivalent per person has very little discretionary consumption to shed when the import bill moves against it. The volume is small; the vulnerability per unit is not.

Renewables do not change that arithmetic much. Renewable energy was 25% of final energy consumption in 2021, third of the six, 0.8 points above the peer median of 24.2, behind Pakistan's 41.6 and India's 34.9 and ahead of Vietnam's 24.2, Indonesia's 20.2 and Thailand's 19. That indicator measures all renewable energy in total final consumption, which includes biomass burned in households, so a share slightly above the median is not evidence of installed solar or wind capacity, and this data carries no separate figure for either. Where the renewable share and the traditional-fuel share are counted together, a number near the peer median says little about the modern energy transition.

What this data cannot separate

The whole argument turns on a distinction the World Bank indicators are not built to make.

A consumption series records energy delivered, not energy wanted. A country can arrive at 287.8 kg of oil equivalent per person by using energy efficiently, or by not having enough of it to use, and the two look identical in this series. Nothing in the six-country panel separates them. The indicators here contain no measure of unserved demand, no outage or load-shedding hours, no queue for industrial gas connections, no suppressed-demand estimate of any kind.

Three confounders deserve naming, because each would move the interpretation and none can be tested inside this comparison. Climate and cooling load differ across the six, and this data holds no temperature or cooling variable. The composition of manufacturing differs: a garment-heavy industrial base and a steel, cement or petrochemical base do not consume energy at the same rate per unit of value added, and there is no sectoral energy intensity here, so the composition hypothesis stays a hypothesis. And the measurement of household biomass, which sits inside the energy series, varies in method across countries in ways a single ranking cannot correct for.

These are six countries observed side by side, not a sample and not an experiment. Every relationship here is an association, and none of them identifies a cause. The years also differ by indicator because coverage does: energy 2023, imports 2022, renewables 2021, income 2024, manufacturing 2025.

What would change this conclusion

A specific, checkable test. The claim in this essay is that 287.8 records a constraint at least as much as an achievement. If Bangladeshi income per person converges toward the current peer median of 16,386 international dollars while energy use per person stays below Pakistan's current 415, then the efficiency reading was right and this essay is wrong. If instead energy use per person climbs toward the peer median of 982.4 as income rises, the low reading was a ceiling, and the planning done on the assumption that Bangladesh is structurally a low-energy economy will have underbuilt supply for a decade.

Three moves, with owners and signals.

1. Publish unserved demand next to supply. The entire ambiguity above would collapse against a national series for load shedding and unmet industrial gas demand, because the efficiency reading and the ceiling reading predict opposite values for it. Owner: Power Division, with the Bangladesh Energy Regulatory Commission on the reporting standard. Success signal: a published quarterly series of unserved energy by division and by industrial category, on a stable definition.
2. Report energy per unit of manufacturing value added, by sector. Manufacturing is 22.44% of GDP and the composition hypothesis is currently untestable from published data. Owner: BBS, with SREDA on the energy side. Success signal: an industry-level energy intensity table in the annual statistical release, which would show whether the low national figure is a garment-sector composition effect or something broader.
3. Treat the import share, not the volume, as the security number. At 44.31% of energy use, second-highest of the six, exposure is the binding constraint rather than consumption. Owner: Energy and Mineral Resources Division. Success signal: net energy imports below the peer median of 36.53% in a future WDI vintage while energy use per person rises, which is the combination the current data does not show.

The counterargument

The strongest objection is not that the numbers are wrong. It is that they are good news read as bad.

On this reading, Bangladesh delivers 4.22% growth, a US$450 billion economy and a manufacturing share above the peer median while burning less fuel per person than any comparator, which is precisely what a world trying to cut emissions says it wants. Treating the lowest energy use per person of the six as a deficiency is asking a poor country to consume more fossil fuel, and the essay's own admission that the data cannot separate efficiency from scarcity cuts against the pessimistic reading exactly as much as the optimistic one.

Three answers. First, the argument here is not that more energy is better. It is that a number capable of two opposite meanings is currently being read as though it had one, and the optimistic reading is the one that requires no new power stations, no new gas contracts and no new capital. That asymmetry is worth noticing in any plan that rests on it.

Second, the efficiency reading makes a prediction the series has not borne out. A structurally low-energy economy should hold its level as it industrialises. Bangladesh's energy use per person is 2.5 times its 1990 level, and the manufacturing share is third of the six and still above the median. The trajectory looks like a country climbing a curve from a low base, not one that has found a way around it.

Third, and this is the concession the objection earns: if the efficiency reading is correct, then Bangladesh's growth path costs less energy than any planner assumes, and nothing in this data rules that out. But the import exposure survives either reading. At 44.31% of energy use, second-highest of the six, the fuel that does get burned is bought in a market Bangladesh does not set the price in.

---

Data sources: World Bank World Development Indicators, retrieved from the BDPolicyLab data lake, 2026-08-10. Comparators are India, Pakistan, Indonesia, Thailand and Vietnam; "peer median" is the median of the comparators reporting in the stated year, excluding Bangladesh. All six countries report on every indicator used here. Years differ by indicator because coverage does: energy use and electricity consumption 2023, net energy imports 2022, renewable share 2021, income and the population and output context figures 2024, manufacturing value added 2025. The GDP per capita series is PPP in current international dollars, so its 1990-2024 multiple includes price change; the energy series is in physical units and does not.

Sources

1. World Bank WDI, Energy use (kg of oil equivalent per capita), EG.USE.PCAP.KG.OE: https://data.worldbank.org/indicator/EG.USE.PCAP.KG.OE
2. World Bank WDI, Electric power consumption per capita (kWh), EG.USE.ELEC.KH.PC: https://data.worldbank.org/indicator/EG.USE.ELEC.KH.PC
3. World Bank WDI, Energy imports, net (% of energy use), EG.IMP.CONS.ZS: https://data.worldbank.org/indicator/EG.IMP.CONS.ZS
4. World Bank WDI, Renewable energy consumption (% of total final energy consumption), EG.FEC.RNEW.ZS: https://data.worldbank.org/indicator/EG.FEC.RNEW.ZS
5. World Bank WDI, GDP per capita, PPP (current international $), NY.GDP.PCAP.PP.CD: https://data.worldbank.org/indicator/NY.GDP.PCAP.PP.CD
6. World Bank WDI, Manufacturing, value added (% of GDP), NV.IND.MANF.ZS: https://data.worldbank.org/indicator/NV.IND.MANF.ZS