Thursday, 13 August 2009

Did electricity companies really overcharge by $4.3b? Part 3

I am still reading the report by the Electricity Technical Adviosry Group, released yesterday, but it looks like they have rejected the conclusions of the Wolak report and instead focused their attention on other areas.

I'll blog on that when I have digested the report. In the meantime, a final post on the Wolak report:

Previously, I blogged here and here on why I don’t believe the figure of $4.3b from the Wolak report on how much electricity companies ostensibly have overcharged themselves through the exercise of market power. The calculation of the $4.3b figure, however, was just one part of the empirical analysis in the report. A significant earlier part of the report is devoted to calculating estimates of the ability and incentive of firms to exercise market power, and showing that it correlates very highly with price.

Now correlation does not imply causality. While the results in that part of the report are consistent with firms having exercised market power, they are also consistent with perfect competition, since the very conditions that lead to a high ability of firms to exercise market power (firms running up against capacity constraints) are the very conditions that would lead to high prices in competitive markets. The details here become somewhat technical, but I want to make one general point: Sometimes empirical results can be too good to be believable. Consider the following graph from the Wolak report:

The red lines shows average prices across the day between 2001 and 2007, and the blue line shows the average estimated ability of firms to exercise market power. One can see just by eyeballing the graph that one could explain close to 100% of the within-day variation in prices by the within-day ability to exercise market power. There are two conclusions one can draw from this: Either the exercise of market power is the only factor influencing price variation; or, the measure of the ability to exercise market power is highly correlated with other determinants of price.

It is pretty hard to believe that market power is the only determinant of within-day variation in price: for instance, there are fixed costs to starting and shutting down thermal power stations, so once a thermal station is started in order to supply power during the peak hours, the marginal cost of continuing it running during low-demand hours is relatively low.

If, on the other hand, the ability to exercise market power is highly correlated with other determinants of price, we cannot infer from the high correlation between it and price that it was the cause of any of the price variation.

The bottom line here is that we should be very very wary of drawing any conclusions from the correlations presented in this section of the report.

No comments:

Post a comment