Nine findings

A building is an energy decision

Every building is an energy decision. Once a roof is set, walls enclosed and windows placed, a pattern of energy use is fixed for as long as the building stands. Roof colour, shading, window size and orientation determine how much heat a home absorbs. The appliances fitted at handover determine how much electricity it takes to remove that heat. Most of these choices are made once, at construction, and are expensive to reverse.

That is the global picture. This investigation narrows it to where the problem is sharpest: fast-growing housing markets with weak codes, unreliable grids and thin housing finance. Nigeria is the principal case. There, the economics of efficiency are shaped less by the grid tariff than by the petrol and diesel generators that fill the gaps in supply.

The findings run against two common assumptions. The first is that efficiency always pays in hot, poorly supplied markets. It pays handsomely where a generator is running, and barely at all for a grid-only household on a subsidised tariff. The second is that the split incentive between developer and occupant is the main barrier. It is a barrier for the building fabric. It is not a barrier for the air conditioner, which is where most of the saving lies.

The global buildings boom

The world's building stock reached 273 billion m² of floor area in 2024, after growing 1.7% in the year. That is about 12.7 million m² of new floor area every day (GlobalABC). Residential buildings account for 77% of floor area and around 70% of buildings' energy demand.

Between 2015 and 2024, floor area grew 20% while buildings' final energy demand rose by around 11% (GlobalABC). New buildings are more efficient than old ones, and global building energy intensity has fallen 8.5% since 2015. But total demand is still rising, and renewables supplied only 17.3% of buildings' energy in 2024. Operational emissions rose 1% to 9.9 GtCO₂.

The next wave of construction is larger still. The IEA expects global floor area to grow by about 45% by 2050, with around 80% of that growth in developing economies (IEA, Breakthrough Agenda Report 2025). Those are the markets with the fewest enforced energy codes and the least housing finance.

For a longer view, the IPCC put total building greenhouse-gas emissions at 12 GtCO₂e in 2019, 21% of the global total. Of that, 57% was indirect emissions from electricity and heat generated off site, 24% direct on-site emissions, and 18% from cement and steel used in construction (IPCC AR6, Working Group III, Chapter 9). Accounting boundaries differ between the IPCC and GlobalABC, so the two figures should not be compared directly.

The energy hidden in construction

A building has emissions before, during and after occupation. Embodied carbon is the emissions from extracting, manufacturing and transporting materials. Operational carbon comes from the energy used while the building is occupied. Whole-life carbon brings these stages together over a defined assessment period. GlobalABC's 37% covers both operations and construction. Cement, steel and aluminium used in buildings alone accounted for about 9% of global emissions in 2024 (GlobalABC). That figure excludes glass, insulation, timber and other materials.

The useful question is not "which material is green?" but "which building function, in which climate, over which period, with which supply chain?" A material that raises upfront emissions can lower lifetime emissions if it cuts cooling demand for decades, and the reverse is also true.

In Nigeria, material costs are shaping construction as much as carbon. The price of a bag of cement rose 442%, from ₦2,400 to ₦14,000, between 2019 and 2024 (CAHF). When the structure itself becomes that much more expensive, developers have little room to add anything beyond the minimum.

The technology changing how homes work

Air conditioners and electric fans use nearly 20% of the electricity consumed in buildings worldwide, about 10% of all global electricity. Without efficiency improvements, energy demand for space cooling would more than triple by 2050 (IEA, The Future of Cooling). In hot countries whose incomes are rising, cooling is the load that grows fastest once a household can afford it.

How efficiently that cooling is delivered depends mostly on the air conditioner. Laboratory testing by Lawrence Berkeley National Laboratory found that in extremely hot-humid climates, a variable-speed (inverter) unit rated at 3.5 W/W achieves a seasonal efficiency of 4.90, against 3.72 for a fixed-speed unit (Park and Shah, LBNL, 2022). Rated efficiency understates the inverter's advantage because most cooling hours are at part load.

How major economies are building for the energy transition

These are explanatory profiles, not rankings. Climate, household size, floor area, fuel mix, tariffs, reliability and enforcement differ too much for league tables.

Africa's housing boom

Africa's housing challenge is also an energy challenge and a finance challenge. The homes being built now will set cooling demand for decades, and most are being built without mortgage finance or energy standards. The three markets below show how differently that plays out.

South Africa has the continent's deepest mortgage market: 1.67 million residential mortgages outstanding in early 2024 and a mortgage-to-GDP ratio of 17.6% (CAHF). The Department of Human Settlements delivered 29,424 houses in 2023/24, and CAHF notes that the department does not publish a current backlog estimate. South Africa also shows that green housing finance can exist on the continent. In December 2021, Nedbank issued a R1.09 billion green bond, anchored by the IFC, to finance EDGE-certified residential developments (IFC). Nedbank also markets home-loan benefits to buyers of EDGE-certified properties, including cash back and reduced bond registration fees (Nedbank). By 2024, 52,180 homes were EDGE-certified (CAHF).

Kenya faces annual housing demand of around 250,000 units against supply of about 50,000 (CAHF). Some 30,015 mortgages were outstanding in 2023, with mortgage-to-GDP of 1.86% and an average mortgage rate of 14.3%. The country has adopted a new National Building Code with standards for efficient and sustainable construction. As of August 2024 it had 89 green-certified buildings and 2,380 EDGE-certified homes (CAHF). Nearly 90% of Kenya's electricity came from renewable sources in 2023, led by geothermal (IEA), so efficiency there matters more for affordability and grid capacity than for carbon.

Sources: CAHF Housing Finance in Africa Yearbook 2024 (country profiles); IFC; Nedbank. Years differ by indicator, as shown.

Nigeria: building for an unreliable grid

Nigeria is short of housing and short of power at the same time. How short depends on the measure. In January 2026 the National Housing Data Technical Committee put the 2025 housing deficit at 14.9 million units, measured on a persons-per-room basis (ThisDay, reporting the committee). The Centre for Affordable Housing Finance in Africa (CAHF) cites about 28 million units for 2023, a figure it says has doubled since 2010. The two are different measures, not a correction of one by the other, and we use the official figure as the benchmark.

Formal supply is a fraction of need: about 50,000 units a year against roughly 700,000 required (CAHF). Only 24,654 residential mortgages were outstanding in 2023, mortgage lending was 0.6% of GDP in 2021, and the typical non-subsidised mortgage rate was 18% (CAHF). The median house price was ₦107 million. Only 13.2% of property owners hold title deeds. With mortgage finance reaching so few buyers, most owned homes are financed from savings and built in stages, outside the reach of lenders' standards.

Who occupies these homes is contested, and it matters for the argument. CAHF reports that homeownership fell to 20% in 2024 and that over 80% of Nigerians live in rented housing. An older NOI Polls phone survey found 51% renting and 31% owning. Neither source separates cities from rural areas, and ownership is widely understood to be higher outside the cities. We therefore treat Nigeria as two housing markets: rented and developer-built urban homes, where the builder and the bill-payer differ, and owner-built homes, where they are the same person and the barrier is capital, not incentive.

What the grid delivers

The most recent household survey is the NBS Residential Energy Demand-Side Survey, produced with IEA and EU support. It covered 7,706 households in nine states between April and May 2024. Lagos was not among them. In those states, 58.2% of households were connected to the grid, and connected households received power for an average of 6.61 hours a day. The survey also found 85.2% of households on estimated billing (NBS). These are the best available household figures, but they are not national figures and should not be read as describing Lagos.

Tariffs depend on the service band a customer's feeder is assigned, which sets a minimum number of supply hours a day. Under NERC's July 2025 order for Ikeja Electric, one of the two distribution companies serving Lagos, residential Band A customers pay ₦209.50 per kWh. Band B pays ₦62.48, Band C ₦45.80, and Bands D and E ₦31.24. The Band B to E rates have been frozen since December 2022. NERC put Ikeja Electric's average cost-reflective tariff at ₦185.9 per kWh against an allowed average of ₦114.7, a subsidy of about ₦26.3 billion a month for that one distribution company (NERC).

Sources: NERC Order/NERC/2025/064 (Ikeja Electric, effective 1 July 2025), Table 2, residential non-maximum-demand customers; band hours from NERC's service-based tariff definitions. Tariffs for Lagos's other distribution company, Eko, are set in a separate order. Our illustrative household T is on Band B; that rate is not a general Nigerian grid tariff.

The generator economy

Households fill the gaps with generators. A 2019 study by the Access to Energy Institute and Dalberg estimated 22 million small petrol generators in use by Nigerian households and small businesses, with around 42 GW of combined capacity against a grid of about 5.4 GW. It found an average load factor of 22% and a working life of at most five years (A2EI/Dalberg). A January 2026 study by ZE-Gen estimates that 40% of households own a generator, mostly under 4 kVA, and that about 2 litres of fuel are needed per kWh. It puts average household spending at about $12.1 a month on fuel, around $2.5 billion a year across 17 million households (ZE-Gen).

Fuel prices have since jumped. The national average petrol price reached ₦1,596 per litre in May 2026, up 55% on a year earlier. Diesel reached ₦3,277, up 86% (NBS).

What a kilowatt-hour costs

To compare these sources on equal terms, we built a new model (Section 8). It rebuilds generator costs from fuel use, maintenance and equipment cost. At May 2026 prices and with our assumed generator performance, petrol-generator electricity for the illustrative household costs about ₦2,080 per kWh ($1.57), and diesel about ₦1,643 ($1.24). A household that mixes a Band B grid supply with a petrol generator pays an average of about ₦567 per kWh served.

The cooling market and the code

The air conditioners on sale are mostly the less efficient kind. In a 2024 market assessment for United for Efficiency (U4E), 77% of models used fixed-speed compressors and 23% were inverters. 19.6% fell below Nigeria's minimum standard of 2.8 W/W, and 19% still used the banned refrigerant R22. Price was only weakly related to efficiency, so buyers cannot use price as a signal of quality (U4E).

The building fabric is not regulated in practice either. Nigeria's 2017 Building Energy Efficiency Code sets requirements such as minimum roof insulation. In a 2023 survey of 150 building professionals working on housing estates in the Federal Capital City, only 18% agreed that the estates complied with the code, and only 12.7% agreed that roofs were insulated. The authors describe the code as seemingly "just paper work without enforcement or compliance" and note that Nigeria has no legalised building code (Macaulay et al., Open Journal of Energy Efficiency, 2023).

Who pays for an efficient home?

The question is not whether efficient homes use less energy. They do. The questions are who pays, who saves, and whether the saving is worth the cost. To answer them for Nigeria, we built a model of two illustrative Lagos households. It is a decision tool, not a building simulation. Every input is cited or labelled as an assumption and tested across a range.

TCL model v2, central case. These are illustrative households, not national averages; most Nigerian households do not own an air conditioner.

Household T shows something a single blended price hides: a third of its potential demand goes unmet. When grid supply is off and the generator is not running, the fan and the air conditioner simply stop. Savings on those hours buy comfort, not naira, and the model counts them as kWh rather than money.

Measure by measure

TCL model v2. Real terms, 10% real discount rate, no real price escalation. Air conditioners and the roof coating are assessed over 10 years, solar over 20 years with a battery replacement at year 10. Solar surplus is not paid for, because no residential export payment was documented. Equipment prices from 2026 Nigerian retailer and installer price guides.

Three results stand out.

First, the inverter air conditioner is the best buy, and it is the occupant's decision. For household T, the inverter premium pays back in about five and a half months. It stays profitable when each input is moved one at a time across its tested range, including the lowest fuel use, the lowest fuel price and the highest discount rate. The price premium would have to rise about 13-fold before it stopped paying. Because tenants usually buy their own units, the split incentive does not block this measure. What blocks it is the upfront premium, the weak link between price and efficiency, and the one in five models on sale that fails the minimum standard.

Second, efficiency does not pay everyone. For a household without a generator on a frozen Band C, D or E tariff, the inverter premium loses money over the unit's life, because the kWh it saves are cheap grid kWh or hours without power. On Band B the central-case NPV is only ₦21,280 over ten years, a marginally positive result that small changes in assumptions could reverse. Frozen tariffs protect these households from bills, but they also erase the price signal for efficiency.

Third, solar with a battery pays only where generators actually run. We assess it for a household that has already switched to an inverter AC. For household T, the system displaces generator and grid electricity and serves about 1,238 kWh a year of demand that previously went unmet. In the central case the system has a net present value of about ₦5.36 million over 20 years and pays back in about 3.7 years. But the result is sensitive. The model's break-even point is about 1.9 generator hours a day, solved directly from the model and falling between the one-hour and two-hour cases in the chart below. It is a modelled threshold, not a measured household operating result. The investment also turns negative if petrol falls below about ₦748 per litre, or if the buyer's real cost of capital exceeds about 26%. For households borrowing at Nigerian consumer-credit rates, that last condition is a real risk.

Where the split incentive bites

The roof is different. A developer or landlord chooses it, and the occupant pays to cool the rooms beneath it. Measured against a home with a fixed-speed air conditioner, the model finds a reflective roof is worth up to about ₦6,800 per square metre over 10 years to a typical household that runs a generator, and about ₦19,400 to the upper-income home. The developer recovers none of that unless buyers or tenants can see it and pay for it. That requires energy labelling of homes, which Nigeria does not have, or a planning rule that makes the specification standard.

For owner-builders, who pay for the roof and the fuel bill alike, the constraint is capital. Building in stages from savings, with mortgage rates around 18% for the few who can borrow, pushes spending toward the structure and away from anything that pays back only over time.

The policy and investment gap

Globally, the money is growing but not fast enough. GlobalABC puts investment in building energy efficiency at around $275 billion in 2024 and estimates that about $592 billion a year is needed to 2030 (GlobalABC). On a broader boundary that also counts electrification and renewables in buildings, the IEA puts 2024 investment at about $380 billion, and notes that growth in efficiency spending "is now stalling" (IEA).

In Nigeria, the evidence in this investigation points to six interventions. Each is matched to an owner, a way to pay for it and a way to know whether it worked.

These measures involve trade-offs. Enforcing AC standards will raise the entry price of cooling for poorer households, which the first recommendation must offset with labelling that shows the running-cost saving. Cost-reflective tariffs will strengthen the efficiency signal but raise bills for households that cannot afford efficiency upgrades.

What the next generation of buildings will mean

Five conclusions follow from the evidence.

First, buildings are increasingly an electricity story. Cooling already takes nearly 20% of building electricity worldwide and grows fastest in hot, emerging economies (IEA). Efficiency lowers each home's demand but does not remove the need for new generation and networks.

Second, construction locks in energy use. The roof, shading and layout are chosen once. In Nigeria's two housing markets, developer-built homes face a split incentive on those choices, and owner-built homes face a capital constraint.

Third, the value of efficiency depends on the marginal kWh. In Nigeria, a kWh saved during a generator hour is worth about 33 times a kWh saved on a Band B grid supply. Efficiency is a strong investment where generators run and a weak one for grid-only households on frozen tariffs.

Fourth, the fastest win is at the appliance shop, not the building site. The inverter air conditioner pays back within six months for generator-using households. Standards, labels and consumer finance can move that market faster than building codes can move construction.

Fifth, the evidence base is thin where it matters most. There is no Lagos household survey of generator hours, no local price data for reflective roofing, and no measured generator fuel use for the equipment households actually run. Each of these gaps changes the answers in this investigation.

Image register and sources

Image register

Sources

All links accessed 27 to 28 September 2026. Page, table and figure references are to the cited document.