Energy/Capital Ratio

There doesn't seem to be much of relation between energy intensity and GDP per capita. But this chart shows that there is a relationship between energy per dollar of (estimated) capital and GDP per capita:



The points off to the top right are mostly oil producers. Apart from those points there seems to be a clear downward trend. High income countries use less energy per dollar of capital than poor countries do. Some poorer oil producers (e.g. Libya and Algeria) also have very low energy/capital ratios. These countries have invested a huge proportion of GDP each year and so supposedly have a large capital stock but have little to show for it in terms of increased output.

In a so-called "putty-clay" model, improvements in energy efficiency can only be implemented by investing in new, more energy efficient, capital goods. Once that equipment and infrastructure is in place, energy use is pretty much predetermined. So in countries with less energy efficient installed capital energy use per dollar of capital will be higher and vice versa. Intuitively we'd expect high income countries, ceteris paribus, to have higher quality capital installed and, therefore, lower energy use per dollar of capital.

Technical stuff :): It is also attractive to try to estimate a model for the energy capital ratio rather than energy intensity for econometric reasons. Given the putty-clay theory, capital per capita and the level of energy efficiency technology are highly correlated. If we run a regression where energy intensity (energy per $ GDP) is on the LHS explained by RHS variables that include capital, the unobserved state of technology will be correlated with the capital variable on the RHS leading to biased estimates of the parameters. Shifting capital to the LHS of the equation gets rid of this potential (and actually very problematic) source of bias. The challenge is now to come up with a theoretically rigorous model of the E/K ratio...