Would the Price elasticity of demand for electricity be more elastic over a shorter or longer?

Decisions about the electricity usage, pricing, and infrastructure investment depend on many considerations. Among them, the price elasticity of demand for electricity is especially crucial. The focus herein is on residential demand for electricity in the long run. Estimates of its price elasticity are plentiful and diverse, and reflect both differences in space and time but also in estimation techniques. In a frequently cited contribution, Espey and Espey (2004) carried out a meta-analysis of price and income elasticity estimates from 36 studies published over the period 1947 to 1997. From the 123 estimates that they analysed, short-run price elasticities ran from −2.01 to −0.004 with a mean of −0.35; and 125 estimates of long-run price elasticity fell in the range from −2.25 to −0.04 with a mean of −0.85.1 Differences in econometric techniques may explain some of the variation, but even with the same methodology, a wide range of estimates can be obtained. For example Krishnamurthy and Kristöm (2015), using a common methodology, obtained a range of price elasticities between −0.27 and −1.4 for a set of 11 OECD countries.2

In more recent years, the feasibility of real-time pricing has sparked interest in determining near-immediate price elasticities when consumers have informational feedback. A great deal of this research is based on experiments (see Faruqui and Sergici (2010) and Jessoe and Rapson (2014) for experimental evidence with respect to residential demand). That is in contrast to econometric studies focusing on short-run and long-run price elasticities, which use either time series, cross-sectional, or panel data sets that are typically from surveys rather than from experiments. One exception is the study by Battalio et al. (1979), which dealt with short-run rather than real-time elasticity. Using a system of rebates and information, the researchers conducted a field experiment on a sample of residential customers in College Station, Texas, over a three-month period. By offering cash payments to a subset of customers for each percentage reduction in their electricity consumption compared to a year earlier, they obtained an estimate of short-run price elasticity of demand. While interesting, experiments of that type have severe limitations. The participants know that they are in an experiment and the experiment is for a short period of time, so there is no incentive for them to invest in changing heating and cooling systems or electrical appliances. Thus, such experiments give no insight into long-run price elasticity.

The findings reported in this paper are based on a very rare set of circumstances that yields a long-run elasticity via a natural experiment.3 Among other things, the subjects involved do not perceive that they are in an experiment, nor was an experiment even intended. It is based on a homogeneous area in the Canadian province of Newfoundland and Labrador, where all residential customers initially faced the same price schedule, but then those in a geographic subset were switched to a different price regime. The change in price was abrupt, substantial, and long lasting. Data on electricity consumption for both groups is available and provides insight into consumer adjustment. In particular, the similarity of the two groups allows for a simple differences-in-differences approach to estimating the long-run impact of the price change.

The next section provides the background on the natural experiment. Section 3 illustrates the magnitude of the price shock and how electricity consumption patterns changed in its aftermath. In Section 4, the long-run price elasticity is determined. Section 5 briefly discusses that result and policy implications, the latter of which are substantial if a similar elasticity value applies to residential customers elsewhere in the same province.

The setting

The focus is on residential demand in communities located on the south coast of the Labrador area of the province. That coastal area and the two regions of interest there are identified in the map of northeastern North America in Fig. 1. One of the two regions is L'Anse au Loup (LAL), named for the largest community within it. The other region will be referred to herein as the Isolated Southern Labrador (ISL) one.

Until late in 1996 all of their electricity demand was met by diesel turbines

The price change and consumption patterns

The key change in the price regimes is illustrated in Fig. 2. It shows two time-series. One series, which begins in 1991 and goes to 2016, shows isolated systems’ residential rates per KWh, expressed in 2015 dollars, for consumption in excess of 1000 KWh per month.8 Those rates

Price elasticity of demand

The reaction to the price change by LAL residential consumers can be quantified by determining the price elasticity of demand (η), i.e., the ratio of the percentage change in consumption to the percentage change in price. Since those changes are large, the arc formula, by which a percentage change is measured relative to the mid-point of the start and end values, is used to calculate the percentage changes.

Estimating the percentage change in average consumption of electricity due to the price

Discussion and policy implications

The finding that demand is price-elastic in the LAL region has direct policy implications for that region. As stated in Section 2, the rates for residential customers there must, as a matter of government policy, be equal to the PUB-approved rates for customers on the island's interconnected grid. That means that there is no connection between the price of electricity in the LAL region and the higher unit cost of providing service there. Price-elastic demand has led to a substantial increase in

Cited by (13)

Recommended articles (6)