Measuring Market Power in Wholesale Electricity Markets

Price spikes do not necessarily indicate market power.  Spikes should be expected even in competitive markets, where they provide crucial incentives for new investment.

In practice, it is difficult to conclusively identify and measure market power in wholesale electricity markets. It is important to recognise that high prices do not necessarily indicate market power. In fact, there is reason to expect bouts of high prices in wholesale electricity markets. As we will see below, episodes of high prices, during periods of high demand (or low supply), play a crucial role in providing incentives for long term investment. These episodes may last for hours (for example, due to a particularly cold day) or weeks (for example, due to drought).

Over the last several years, motivated partly by growing recognition of the role of market power in the UK in the 1990s and the California electricity crisis of 2000-01, scholars have accumulated a number of empirical approaches to measuring market power in wholesale electricity markets. Some are simple to construct, while others are highly complex and require data that is difficult to obtain. Each has strengths and weaknesses. In practice, several approaches should be used in conjunction in order to paint a sufficiently comprehensive picture.

Economists have developed a number of approaches to measuring market power.  None of them are perfect.

There are three basic empirical approaches to identify market power. (Twomey et al (2004) give an excellent overview.) Each approach can be summarised in the form of a question.

Do any firms have the ability to exercise market power? A number of measures assess each firm’s share of the total supply (usually expressed as megawatts of generating capacity) in a given market or country. The idea is straightforward: a market dominated by a few large firms will be more susceptible to market power than a market with numerous relatively small firms. To make sense of this information – and to enable meaningful comparisons across time or perhaps across countries – the shares can be aggregated into a Herfindahl-Hirschman Index (the sum of squares of each firm’s share). Higher values of this index indicate greater concentration and scope for market power. Some studies look at a “pivotal supplier” measure which indicates how often a given firm has to run at least some of its capacity – that is, the measure looks at each firm’s capacity relative to demand.[10] There are other more complex variations, but the basic idea is the same: evaluate whether any firm is large enough relative to the market to allow it the ability to change its own output in a way that will affect the market price.

Unfortunately, concentration measures can give an incomplete picture of the ability of firms to exercise market power. First, these measures generally do not reflect the effect of transmission constraints. As noted above, transmission constraints effectively change the size of the market by limiting the amount of competition at various locations on the network. For example, a given firm may only have a small fraction of the overall generation capacity in a country, but transmission congestion may emerge during certain periods, effectively giving the firm a large share of a region that is temporarily cut off from competition. Second, concentration measures don’t consider the scope for entry by new firms. For a given level of concentration, a market where new investment is very slow (eg, due to heavy permitting procedures) will be more susceptible to market power compared to a market where entry is relatively easy.

Have any firms actually exercised market power in a given period? To answer this question, researchers look at detailed data on plant characteristics and input prices, and attempt to estimate a marginal cost curve for each generator. These estimates of marginal costs are then compared to each generator’s actual bid prices. Deviation of bid prices from estimated marginal cost indicate market power – if, of course, the estimate is correct. This approach requires a lot of data and is sometimes controversial because estimates of marginal costs will always carry a degree of imprecision. Getting an accurate estimate of marginal cost can be particularly problematic in the case of hydro generation where the marginal cost of an extra unit of production includes complex considerations about future prices (Evans 2006). (Hydro generators are typically faced with the thorny challenge of choosing when to use water from limited reservoirs to generate electricity. Thus, the true marginal cost of generation for each hydro generator includes the “opportunity cost” of not using the water at some other time. Accordingly, estimates of the marginal cost of hydro generation need to take into account the generator’s hourly price forecasts.[11]) A related method evaluates data on unplanned plant outages. If a given firm owns plants that are out of service more frequently than is statistically typical for the relevant plant age and type, then this may be considered evidence of market power (again, depending on the accuracy of the estimate).[12]

Does the performance of the actual wholesale market match the predictions of a simulation model with competitive characteristics? Some economists build complex simulation models that model the characteristics of a given wholesale market. The modellers simulate market prices, bids and other output under the assumption that the market is highly competitive. These modelled outputs can then be compared to actual data from the real-world market. This is a useful approach, although it can be time consuming. The results can be difficult for a non-specialist to assess.