Energy sources, carbon emissions and electricity prices

This post investigates the relationships between the percentage of a country's electricity from different (alternative) sources, the country's carbon emissions, and the price its households pay for electricity. The sample comprises the 34 OECD countries. Data on the percentages of electricity from different sources were obtained from the Energy Information Association. Data on carbon emissions were obtained from the World Bank. And data on electricity prices were taken from the OECD. All figures are for 2010; electricity prices were unavailable for Australia and Iceland. 

The chart below displays histograms of the distribution of countries by percentage of total electricity generated from six sources: hydroelectric, nuclear, geothermal, wind, solar, and biomass. The distributions for geothermal, wind, solar and biomass are heavily concentrated at the low end of the scale. Mean percentages of electricity from these four sources are, respectively, 1.3%, 3.4%, 0.3% and 3.9%. The distributions for hydro and nuclear are slightly less concentrated at the low end. Mean percentages of electricity from these two sources are 21% and 16%, respectively. Iceland is an outlier in the case of geothermal, generating fully 26% of its electricity from that source, while most other countries generate none.

The first table (below) displays coefficients from linear regression models of carbon emissions per unit of energy used on percentage of electricity from different sources. Carbon emissions per unit of energy used was calculated simply as carbon emissions per capita divided by energy use per capita. For investigating which energy sources are associated with the highest carbon emissions across countries, emissions per unit of energy is a preferable measure to emissions per capita, since the former will be unaffected by differences in emissions that derive from wealth, population density and lifestyle factors.

The first three columns contain significant negative coefficients, indicating that emissions per unit of energy tend to be lower in countries that generate more electricity from hydro, nuclear and geothermal. By contrast, the next three columns contain non-significant coefficients (two positive and one negative), indicating that emissions per unit of energy do not tend to be lower in countries that generate more electricity from wind, solar and biomass. As the seventh and eighth columns show, respectively, percentage of electricity from hydro, nuclear and geothermal explains 73% of the variance in emissions per unit of energy across OECD countries, whereas percentage of electricity from wind, solar and biomass explains 0% of the variance. (I suggest why further down.)

The second table (below) displays coefficients from linear regression models of electricity price for households on percentage of electricity from different sources. The first three columns contain non-significant negative coefficients, indicating that the electricity price does not tend to be higher in countries that generate more electricity from hydro, nuclear and geothermal. By contrast, the next three columns contain significant positive coefficients, indicating that the electricity price tends to be higher in countries that generate more electricity from wind, solar and biomass. As the seventh and eighth columns show, respectively, percentage of electricity from hydro, nuclear and geothermal explains only 1% of the variance in the electricity price across OECD countries, whereas percentage of electricity from wind, solar and biomass explains 48% of the variance.

The chart below displays, on the left, emissions per unit of energy against percentage of electricity from hydro, nuclear and geothermal (corresponding to the seventh column in the first table), and on the right, the electricity price against percentage of electricity from wind, solar and biomass (corresponding to the eighth column in the second table). Both relationships are strong as well as linear, though the correlation on the right does drop from r = .69 to r = .57 when Denmark is omitted. 

In conclusion: use of hydro, nuclear and geothermal tends to coincide with lower emissions per unit of energy but not with a higher price of electricity, whereas use of wind, solar and biomass tends to coincide with a higher price of electricity but not with lower emissions per unit of energy. It is important to note, of course, that these are simply cross-country associations, not causal effects. One possible reason why use of wind and solar does not tend to coincide with lower emissions is that, because these are intermittent sources of energy, they generally have to be backed up by alternative sources, which often means fossil fuels. On the other hand, it could simply be that countries which generate more electricity from wind and solar tend to generate more of the rest of their electricity from fossil fuels, rather than from hydro, nuclear or geothermal.