Uncertainty—an argument for more stringent energy conservation

Abstract

Most energy conservation analyses assume various future paths for energy prices and other parameters and then proceed to calculate the optimal value of conservation investment under the specified assumptions. Sensitivity analysis is used to examine the amount which the calculations change as parameters are varied. In fact, the future is not known with certainty, and it is desirable to include uncertainty explicitly in the analysis. This paper approaches the uncertainty problem in a simple way but one which provides considerable insight. We assume that future energy price growth is characterized by a probability distribution, and calculate the optimal investment strategy for conservation investment given this uncertainty. The results are striking. Introduction of uncertainty leads to the conclusion that more conservation investment is desirable than would be made without uncertainty. The conclusion stems, in essence, from the observation that the upside risk to the consumer resulting from unexpectedly high energy prices is larger than the downside savings which would result from unexpectedly low energy prices. The policy conclusion is straightforward: All else being equal, if you think that future energy prices are uncertain, it pays off (for both the individual and the nation) to err on the side of “too much” rather than “too little” conservation. One set of results is expressed in terms of an effective growth rate of energy prices. An example is an electrically heated house located in Washington, D.C. If the most likely growth rate of energy prices is 5% per year, but there is a normally distributed uncertainty about this price growth of 5%/yr (i.e. it is equally likely that prices will grow by 10%/yr and by 0%/yr), then one should calculate the conservation investment as if the price growth rate is 6.8%/yr. A second price assumption used is a price jump at some time from an initial price to a new, higher, final price. Uncertainty is introduced in the time when the price jump occurs. The inclusion of uncertainty leads to the conclusion that the effective time of the price jump is sooner than the most likely time. This result implies that a higher investment in conservation is warranted than would be optimal absent uncertainty. Our model is simple and maybe applied directly by individual decision makers. From a societal point of view, there are many additional arguments that suggest that uncertainty leads to changes in optimal strategies. The essential results cannot be captured simply by carrying out sensitivity analyses on deterministic models. Changes in model structure are required. We believe the inclusion of uncertainty in energy modeling is an important area, and one which can lead to fruitful results for policy analysis. We also make a prediction about energy conservation investment in buildings: Total energy use is predicted to fall exponentially with total conservation investment.