Grows where the grid hasn’t.
Carbon per connection
Africa's Mission 300 energy compacts promise about 101 million new household electricity connections by 2030. Every one of them is often counted as carbon saved: a kerosene lamp put out, a credit to sell. We checked that claim for all 35 compact countries, first with the default assumptions carbon methodologies allow, then with what household surveys say people actually use for light. The two answers are seven times apart.
What a developer would claim, and what the surveys support
Carbon methodologies let a project assume that every newly connected household stops burning a kerosene lamp. Household surveys from the last census, MICS, DHS and Multi-Tier Framework rounds tell a different story: in 25 of the 30 surveyed countries, the main light in homes without electricity is already a dry-cell or rechargeable torch, which counts as zero. The kerosene transition largely happened on its own, without carbon finance.
-3.3 Mt
net CO2 a year, central case: 2.9 Mt avoided, 6.3 Mt added
Lighting baseline per country from surveys (kerosene lamps are the main light for a median 5% of unelectrified households; candles counted too). New grid connections draw power from grids that are mostly thermal in West Africa and coal in Southern Africa, so across the 35 countries they add more CO2 than they avoid. Off-grid connections alone still avoid a net 0.6 Mt a year. Range: -1.4 Mt (low) to -2.9 Mt (high).
+14.9 Mt
net CO2 a year, central case: 21.2 Mt avoided, 6.3 Mt added
One kerosene lamp per household (the CDM small-scale default), applied to every new connection. This is what a project developer could claim under current methodologies, and it is roughly what headline "carbon value of electrification" figures rest on. Same connections, consumption and grid factors as the survey view; only the lighting baseline differs.
Both views use the same connections, the same consumption per connection and the same grid factors; only the lighting baseline changes. Small-generator displacement (5% of a new household's electricity replacing petrol or diesel generation in the central case) is an assumption with no survey anchor yet; it is now the largest avoided term in the survey view, and the page says so wherever it appears.
Who still lights with kerosene
Among households without electricity, a dry-cell or rechargeable torch is the main light in 18 of the 27 countries whose survey separates it, and a torch or a solar lantern in most of the rest. Kerosene is the main source for one household in five or more only in Gabon, Sao Tome and Principe, Lesotho, Madagascar, Kenya, Uganda and Botswana. Candles matter in Burundi, Lesotho, Sao Tome and Principe, Botswana, Zambia, Uganda, Zimbabwe and Gambia.
Latest census, MICS, DHS or ESMAP Multi-Tier Framework survey per country, 2014 to 2024; share among households with no grid or off-grid electricity. Dashed outlines: no lighting survey found, the median of the others is used. Hover a country for its figures and source.
Tonnes of CO2 a year per 1,000 new connections, net. Hover or tap a country. Angola, Eswatini and South Africa have no published compact.
Where a connection still avoids carbon
The countries that come out ahead in the survey view are the ones with a clean grid (hydro or geothermal: Central African Republic, DR Congo, Ethiopia, Kenya, Malawi, Namibia, Sierra Leone and Uganda) or where kerosene is still common among households without power (Gabon (67%), Sao Tome and Principe (48%), Lesotho (45%), Madagascar (44%), Kenya (23%), Uganda (21%) and Botswana (20%)). Countries that lean on coal, gas or oil add CO2 with every grid connection: Comoros, Niger, Botswana and Gambia add 156 to 191 t a year per 1,000 connections.
Off-grid connections, through solar home systems and mini-grids, avoid carbon almost everywhere. Grid connections on a thermal grid do not. The carbon case for electrification is a case for clean grids, not for connections as such.
The gap between avoided tonnes and tonnes with a buyer is nearly the whole total. Only 9 countries have an Article 6 deal that covers electricity access (Tanzania, Zambia, Ethiopia, Malawi, Kenya, Gabon, Benin, Senegal and Rwanda), and no African country has yet authorized a solar home system, lantern or mini-grid program under Article 6.2; the authorized portfolios are cookstoves first. Renewable energy certificates (I-RECs) could certify about 7.5 TWh a year of the new load in issuing countries, but a certificate and a carbon credit cannot cover the same megawatt-hour, so they are kept out of the gap.
The claim, the evidence and the net, per 1,000 connections
Each row is one compact country. The hatched bar is what a default one-lamp baseline would claim as avoided; the solid bar is what the lighting surveys support; the sand bar to the left is what the new grid draw and mini-grid diesel add; the dot is the net. Hover a row for the numbers.
Tonnes of CO2 per 1,000 new connections and year, sorted by the net in the central case. * No lighting survey; median used.
| Country | Connections (M) | Grid share | Kerosene homes | Grid factor | Avoided t | Added t | Net t | Per 1,000 | Article 6 route | Monetizable t |
|---|---|---|---|---|---|---|---|---|---|---|
| Gabon | 0.09 | 62% | 67% | 0.38 | 12,649 | 7,641 | +5,008 | +57 | Open | 3,038 |
| Madagascar | 4.1 | 20% | 44% | 0.48 | 393,309 | 172,180 | +221,129 | +54 | In prep. | 0 |
| Uganda | 2.7 | 70% | 21% | 0.02 | 161,540 | 23,439 | +138,101 | +50 | In prep. | 0 |
| Kenya | 5.1 | 96% | 23% | 0.07 | 292,852 | 116,168 | +176,684 | +35 | Open | 5,037 |
| Central African Republic | 0.4 | 44% | 9% | 0.00 | 10,771 | 2,725 | +8,046 | +19 | None | 0 |
| Ethiopia | 9.2 | 70% | 5% | 0.00 | 191,811 | 27,700 | +164,111 | +18 | Open | 11,134 |
| DR Congo | 11.4 | 44% | 8% | 0.00 | 301,639 | 127,533 | +174,106 | +15 | None | 0 |
| Malawi | 2.7 | 43% | 2% | 0.03 | 39,234 | 15,506 | +23,728 | +9 | Open | 9,646 |
| Namibia | 0.2 | 83% | 1% | 0.02 | 3,419 | 1,832 | +1,587 | +8 | None | 0 |
| Sierra Leone | 0.7 | 22% | 1% | 0.04 | 5,750 | 5,233 | +517 | +1 | None | 0 |
| Mozambique | 4.9 | 71% | 5% | 0.09 | 105,360 | 119,808 | -14,448 | -3 | In prep. | 0 |
| Burundi | 3.1 | 60% | 0% | 0.19 | 111,368 | 131,300 | -19,931 | -6 | In prep. | 0 |
| Zambia | 3.2 | 45% | 5% | 0.19 | 81,070 | 104,442 | -23,373 | -7 | Open | 21,336 |
| Cameroon | 1.6 | 62% | 5% | 0.15 | 32,452 | 57,694 | -25,241 | -16 | In prep. | 0 |
| Sao Tome and Principe | 0.01 | 62% | 48% | 0.68 | 1,141 | 1,413 | -272 | -29 | None | 0 |
| Lesotho | 0.4 | 58% | 45% | 0.73 | 44,673 | 55,422 | -10,748 | -29 | None | 0 |
| Guinea | 1.2 | 71% | 5% | 0.19 | 24,298 | 59,916 | -35,617 | -29 | None | 0 |
| Tanzania | 8.3 | 70% | 13% | 0.35 | 300,792 | 742,606 | -441,814 | -53 | Open | 36,952 |
| Liberia | 0.5 | 60% | 3% | 0.35 | 7,198 | 36,274 | -29,076 | -62 | None | 0 |
| Nigeria | 20.0 | 62% | 7% | 0.38 | 465,548 | 1,729,163 | -1,263,614 | -63 | In prep. | 0 |
| Zimbabwe | 3.1 | 62% | 5% | 0.42 | 79,267 | 282,043 | -202,777 | -66 | None | 0 |
| Republic of Congo | 1.0 | 62% | 5% | 0.41 | 18,362 | 87,339 | -68,976 | -72 | None | 0 |
| Ghana | 0.9 | 73% | 2% | 0.33 | 13,485 | 81,405 | -67,919 | -75 | None | 0 |
| Cote d'Ivoire | 3.9 | 66% | 1% | 0.36 | 46,795 | 338,609 | -291,814 | -75 | None | 0 |
| Rwanda | 0.6 | 75% | 1% | 0.35 | 9,982 | 57,864 | -47,883 | -78 | Open | 99 |
| Togo | 0.8 | 82% | 7% | 0.36 | 20,752 | 87,455 | -66,704 | -79 | None | 0 |
| Mauritania | 0.6 | 48% | 5% | 0.59 | 11,453 | 62,359 | -50,905 | -87 | None | 0 |
| Senegal | 0.8 | 58% | 4% | 0.61 | 15,522 | 102,095 | -86,573 | -111 | Open | 1,033 |
| Benin | 0.9 | 84% | 15% | 0.55 | 37,806 | 151,576 | -113,771 | -123 | Open | 2,462 |
| Chad | 2.3 | 56% | 1% | 0.72 | 23,920 | 341,949 | -318,030 | -136 | None | 0 |
| Burkina Faso | 3.0 | 70% | 1% | 0.63 | 36,526 | 486,487 | -449,962 | -149 | None | 0 |
| Gambia | 0.08 | 62% | 0% | 0.76 | 1,131 | 13,382 | -12,251 | -156 | None | 0 |
| Botswana | 0.2 | 72% | 20% | 0.97 | 15,010 | 54,876 | -39,865 | -181 | None | 0 |
| Niger | 2.9 | 72% | 1% | 0.76 | 30,994 | 561,035 | -530,042 | -184 | None | 0 |
| Comoros | 0.06 | 78% | 5% | 0.76 | 1,286 | 12,687 | -11,401 | -191 | None | 0 |
Connections: new household connections by 2030. Kerosene homes: share of unelectrified households whose main light is a kerosene lamp, from the latest national survey (five countries without one use the median). Grid factor in tCO2/MWh. Article 6 route: open means a signed deal that covers electricity access; in preparation means Article 6.4 activities but no deal. Grid factors: Aloe Energy 2024 production averages for all 35 countries. Low and high cases, the import-adjusted grid view and every flag are in the data, available on request.
Method, sources and updates
How we calculate it
Connections. People reached in 2030 from each compact (its own figure, or its access target times today's population), divided by UN household size, split into grid, mini-grid and standalone solar by the compact's stated shares. Where a compact gives no split, the median of the others is used and flagged.
Consumption. 120 / 300 / 600 kWh a year per new grid household (low / central / high), 75 / 200 / 420 for mini-grids, 10 / 40 / 75 for standalone solar, from the ESMAP Multi-Tier Framework and observed post-connection use in Rwanda and Kenya.
Avoided. Lighting: the share of a country's unelectrified households whose main light is a kerosene, paraffin or oil lamp, times 0.092 / 0.2 / 0.375 tCO2 a year (CDM one-lamp default to the AMS-I.L rationale), plus candles at 0.027 / 0.053 / 0.107 t (our own factor from IPCC paraffin wax; no methodology credits candles). Torches, firewood and "none" count zero. Small generators: 0 / 5 / 15% of the household's kWh at 0.8 to 1.0 tCO2/MWh (CDM defaults). This generator share is an assumption; no survey separates generator households for most countries yet.
Added. Grid kWh divided by 0.85 for losses, times the country's 2024 production-average grid factor from our grid emission factors. Mini-grids: 0 / 10 / 25% of kWh from diesel at 1.0 tCO2/MWh. Standalone solar: zero.
Monetizable. Off-grid avoided tonnes count where the country has a signed Article 6 deal or authorized activity AND its framework covers electricity access, times a creditable share of 50 / 70 / 90%. Grid connections are not creditable to the household, because the baseline is the grid itself. Countries hosting Article 6.4 activities without a deal are "in preparation". Price context, not a valuation: Article 6 units have traded at about US$10 (renewables) to 25 (cookstoves) a tonne; I-RECs at US$0.6 to 1.5 per MWh in Southern Africa.
Sources
The 35 published Mission 300 national energy compacts (see the compacts compared); national census, MICS, DHS, ESMAP Multi-Tier Framework and household budget surveys for lighting fuel (30 countries, 2015-2024); Aloe Energy grid emission factors 2024 (Ember and EIA data, CC BY 4.0), compared with the IFI harmonized default grid factors v3.2 (2022); CDM small-scale methodologies AMS-III.AR, AMS-I.L, AMS-I.A and TOOL33; ESMAP Multi-Tier Framework and mini-grid reports; AMDA benchmarking 2022 and 2024; UNEP Copenhagen Climate Centre Article 6 pipeline; I-TRACK Foundation issuer list; UN DESA household size; Natural Earth outlines.
What would change the answer. A survey anchor for small-generator use (Gabon, Liberia, Cote d'Ivoire and Benin publish one) would firm up the largest avoided term. Weighted survey tables would replace the unweighted frequencies used for most countries. A life-cycle view of dry-cell torches would add roughly 0.01 to 0.02 t per household a year, of the same order as the kerosene term now. Grid factors falling by 2030 would shrink the added tonnes; see where the factors are heading.
Indicative, not a registered baseline and not carbon market advice. Numbers are estimates with stated assumptions. The data, including the low and high cases and every source, is available on request: email [email protected].
What changed
Refreshed once a year with the grid emission factors, and whenever a country signs an Article 6 deal that covers electricity access. Each change keeps its source.
- 12 November 2026 - First edition: 35 compact countries, default baseline and survey views, Article 6 and I-REC routes, 2024 grid factors.
Common questions
How much CO2 does a new electricity connection avoid in Africa?
Far less than carbon methodologies assume. Using the default of one kerosene lamp per household, the 101 million connections promised in the Mission 300 compacts would avoid about 21.2 Mt CO2 a year. Using what household surveys show people actually use for light, about 2.9 Mt. Once the electricity the new households draw is counted, the net is slightly negative across the 35 countries, and positive only where the grid is clean or kerosene is still common.
Can electrification projects earn carbon credits?
Off-grid ones can in principle, where a country has an Article 6 deal that covers electricity access; today that is 9 of the 35 compact countries, and no African country has yet authorized a solar home system or mini-grid program under Article 6.2. Grid connections are not creditable to the household, because the grid itself is the baseline. The avoided tonnes with a buyer route today come to about 0.09 Mt a year.
Why does a grid connection add CO2?
Because the power comes from somewhere. On a gas, oil or coal grid, a household using 300 kWh a year adds more CO2 than the kerosene or candles it stops burning. On a hydro or geothermal grid (Ethiopia, Kenya, Uganda, DR Congo, Malawi) it adds almost nothing. See our grid emission factors.
Do people in Africa still use kerosene lamps?
Mostly not. In 25 of 30 surveyed compact countries, dry-cell or rechargeable torches are the main light in households without electricity (60 to 99%). Kerosene is still the main source in Gabon (67%), Sao Tome and Principe (48%), Lesotho (45%), Madagascar (44%), Kenya (23%), Uganda (21%) and Botswana (20%); candles matter in Burundi, Botswana, Sao Tome and Zambia.
Can I get the data or a figure for my project?
The data is available on request, and we prepare project-specific baselines as a paid study. Email [email protected].