Australian Energy Efficiency in Context

Slides for a presentation I will be giving on Tuesday on putting Australia's level of energy efficiency into context. Here's a quick guide...

Slide 3 compares Australia's energy intensity (energy/PPP GDP) to 5 developed and 2 developing economies. Australia is mid-way between the US and Canada on the one hand which are relatively inefficient by this measure and Germany, Japan, and the UK which are relatively efficient. Energy intensity has declined in all countries over time but slower in Australia and Japan. Based on this measure, India and China are just as energy efficient now as Western European countries.

Slide 4 shows some of the variables people usually think of as important in this context and some I found to be important. Australia has a big mining sector but so do Canada and China. Unlike those two countries Australia is warm in the winter. More of Australia's fuel mix is made up of coal than is the case for the other developed economies. But the percentage is as high or higher in China and India. The other three factors here affect the choice of underlying efficiency after controlling for factors such as in the first three columns. In terms of total factor productivity and exchange rate overvaluation Australia isn't an outlier. It does have very high fossil fuel reserves relative to the size of the economy.

Slides 5 through 7 discuss the econometric model which I also presented at my presentations in Adelaide. Taking the above into account, I find Australia to be relatively inefficient.

Slides 8 and 9 compare my results to Filippini and Hunt (2009). They estimate relative energy efficiency across OECD countries while control for GDP and the price of energy as well as some of the same structural factors that I consider. In my model, countries with low energy prices will likely be less efficient. But Filippini and Hunt are estimating energy efficiency controlling for energy prices, which is a big difference. Also, in my model efficiency is a function of explanatory variables, whereas in their model it is just random. As a result it has a bigger variance in my model. They find the US and Canada to be relative inefficient. But I find Australia and Canada to be the two more inefficient countries.