Two decades of Environmental, Social, and Governance data, 217 countries at a time

Renewables are gaining in electricity, but the average country's total energy mix hasn't gotten greener

Averaged across every reporting country, renewable energy's share of total final energy consumption slipped from 31.2% in 2002 to 29.1% in 2021 — a 2.1-point decline. This is a simple average of each country's own share, so every country counts equally. But the World Bank counts traditional biomass (firewood, charcoal) as renewable, and the decline sits in the poorest countries: lower-middle-income countries fell from 55.9% to 45.2% and low-income countries from 74.5% to 66.6%, while high-income countries rose from 10.5% to 15.6%. That pattern most likely reflects households moving off traditional biomass as they gain modern energy, rather than a retreat from modern renewables. (The set of reporting countries changes from year to year, so as a check: among the 209 countries with data in both 2002 and 2021, the average moved from 31.2% to 29.2%, a 2.0-point decline — the trend isn't an artifact of who reported.)

Electricity tells a different story. Among the 195 countries that reported in both 2002 and 2021, the average renewable share of electricity generation rose from 26.9% to 32.8%, and it rose in 136 of them. The power-sector transition is broad-based; it just hasn't yet moved countries' total energy mix, which still includes heating, transport, and industry.

World Bank indicator EG.FEC.RNEW.ZS (includes traditional biomass), unweighted mean across reporting countries; income-group and electricity figures (EG.ELC.RNEW.ZS) in the text are computed in scripts/analyze.js. 2022–2023 excluded: only ~35% of the usual country panel had reported yet at data-pull time (see methodology).

Electricity access still splits almost exactly along income lines

In 2023, high-income countries averaged 99.9% electricity access and upper-middle-income countries 97.6% — both essentially universal. Lower-middle-income countries averaged 79.0%, and low-income countries just 43.5%: a 56.4-point gap between the richest and poorest income tier. This isn't a story of gradual convergence so much as two different worlds — access is a solved problem for three of the four income tiers and an open one for the fourth.

World Bank indicator EG.ELC.ACCS.ZS, mean by World Bank income classification, 2002–2023.

High-income countries emit more than seven times the CO2 per person of low-income countries

In 2023, emissions per person stepped down with income: the average high-income country produced 7.2 times the CO2-equivalent per person of the average low-income country. Read alongside the electricity-access finding above, the pattern is uncomfortable for ESG framing that treats "emissions" and "development" as separable: the countries furthest from full electricity access are also the ones with the smallest carbon footprint per person.

• High income — 11.17 t per person. The largest footprint of the four tiers, 7.2× the low-income average.
• Upper middle income — 5.65 t. Roughly half the high-income level.
• Lower middle income — 2.36 t. About a fifth of the high-income level.
• Low income — 1.55 t. The smallest footprint — and, from the finding above, the tier with the lowest electricity access (43.5%).

World Bank indicator EN.GHG.ALL.PC.CE.AR5 (total GHG excluding land-use change), mean by income group, 2023.

Coal's share of the electricity mix has drifted down, but slowly

The average country generated 12.5% of its electricity from coal in 2002, falling to 10.7% by 2021 — a real but modest 1.85-point decline over two decades. Holding the country set fixed gives the same picture: the 205 countries with data in both years went from 12.6% to 10.7%. That's a much smaller move than the renewable-share finding above would predict if coal were being displaced directly by renewables; more likely, natural gas and other sources are absorbing some of coal's lost share too.

World Bank indicator EG.ELC.COAL.ZS, unweighted mean across reporting countries. 2022–2023 excluded for the same thin-coverage reason as the renewables chart.

Effective government and electricity access go together — but the link runs one way

Across 4,370 country-year observations, the correlation between the World Bank's Government Effectiveness score and electricity access is r = 0.574. On the 203 country averages plotted below (each dot is one country's 2002–2023 average on both measures), r = 0.610 — a strong positive relationship for cross-country social data. It is an association, not a demonstration that governance causes access: income, geography, and governance move together, and the country-years behind the first figure are not independent observations.

The pattern is lopsided. Of the 91 countries with a governance score at or above zero, only 2 average below 80% electricity access — the bottom-right of the chart is nearly empty. But weak governance is not a barrier on its own: 43 of the 112 countries scoring below zero average above 90%. Good governance looks close to enough for near-universal access; its absence doesn't rule it out.

World Bank indicators GOV_WGI_GE.EST and EG.ELC.ACCS.ZS. Each point is one country's own multi-year average (203 countries with data on both). Dashed line: least-squares fit to the plotted country averages; the r shown on the chart is for those same points.

Latin America & the Caribbean has had the highest income inequality of any region in every year of the panel

Gini data is reported unevenly, so each country counts once, at its most recent reading between 2015 and 2023 (151 countries in all). On that basis Latin America & Caribbean averages 43.4 (22 countries), followed by Sub-Saharan Africa at 40.5 (44 countries). South Asia sits lowest at 30.3 (6 countries). North America shows 36.6, but that rests on only 2 countries, so it shouldn't be read as a regional pattern. The top spot isn't a fluke of the window: comparing whichever countries reported in each year, Latin America & Caribbean had the highest regional average in 22 of 22 years.

World Bank indicator SI.POV.GINI, unweighted mean by region of each country's most recent value in 2015–2023; the number of countries behind each bar is in its label. Gini figures come from household surveys that many countries run only every few years, so region averages with few countries are fragile.

Women's parliamentary representation has nearly doubled since 2002

The average share of parliamentary seats held by women rose from 13.6% in 2002 to 25.5% in 2023 — an 11.9-point gain. It's the clearest steady, broad-based trend in this entire dataset: unlike the renewable energy and coal figures above, the average rose in every one of the 21 year-over-year steps in the panel. Holding the country set fixed, the 156 countries with data in both years went from 13.7% to 26.6%.

World Bank indicator SG.GEN.PARL.ZS, unweighted mean across reporting countries, 2002–2023.

Stronger regulatory quality tracks with cleaner air — a moderate but real relationship

Regulatory Quality and PM2.5 air pollution correlate at r = −0.333 across 4,294 country-year observations, and at r = −0.348 across the 199 country averages plotted below — weaker than the governance/electricity relationship above, but still a meaningful negative relationship: better-regulated countries tend to have less fine-particulate air pollution, even though regulation is far from the only thing driving air quality (geography, industrial mix, and population density all matter too). As above, this is an association rather than proof that regulation causes cleaner air.

World Bank indicators GOV_WGI_RQ.EST and EN.ATM.PM25.MC.M3. Each point is one country's own multi-year average. Dashed line: least-squares fit to the plotted country averages; the r shown on the chart is for those same points.

Building an ESG composite score from this data — and where it lands 195 countries

Every other finding above looks at one indicator at a time. This one combines all 24. Each country's value on each indicator is converted to a percentile rank against every other country that reported it that year (0–100); indicators where a high value is a bad outcome are inverted so 100 always means the best outcome. A country's years are averaged within each indicator, the indicators are averaged into an Environmental, a Social, and a Governance score, and those three are averaged into the composite. Every indicator carries equal weight within its pillar and every pillar carries equal weight overall — a transparent choice, not a claim that this is the right way to weight ESG.

Countries that report only a handful of indicators can't be ranked fairly, because their score depends on which few happen to be available. So a country is scored only if it reports at least 18 of the 24 indicators and at least two in each pillar; that leaves 195 countries and sets aside 22, mostly small economies and territories. As a sanity check, Norway, Sweden, Iceland, Finland, and Switzerland lead the list and Somalia, Yemen, Afghanistan, South Sudan, and Libya sit at the bottom — a mix of wealthy, stable countries at the top and conflict-affected or fragile states at the bottom, which is what a face-valid score should show. Two cautions: the top five sit within 1.5 points of each other, so their exact order means little; and the environmental pillar is where even leaders lose ground (for Norway, 61.6 on Environmental against 96.0 on Governance), because wealthy countries do poorly on per-capita emissions and fossil-fuel reliance.

Composite = average of the Environmental, Social, and Governance pillar scores. Hover a bar to see how many of the 24 indicators the country reports. Full method in the closing section below.