Establishing the impact of climate change on energy demand requires a measure of heating and cooling requirements. In the United States, this measure is a degree day, which is defined in terms of an absolute difference between average daily temperature and 651F, which is an arbitrary benchmark for household comfort. Commercial heating degree days are incurred when outside temperatures are below 651F, generally during the winter heating season from October through March. Cooling degree days occur when temperatures exceed 651F, often during the summer cooling season from April through September. Degree days are intended to represent the latent demand for heating and air-conditioning services based on temperature departures from the comfort level of 651F.
Heating and cooling degree days are reported by the National Oceanic and Atmospheric Administration at a weekly frequency for 50 cities throughout the United States. State, regional, and national population weighted averages of heating and cooling degree days are also reported and commonly used by the utility industry and energy commodity trading community.
Clearly, a higher number of degree days––heating, cooling, or both––should be associated with greater fuel consumption. Estimating how energy consumption rises for every degree day requires some form of model that links weather conditions to US energy consumption. There are two basic modeling approaches. One approach uses engineering models that estimate energy consumption by multiplying a utilization rate for an energy-consuming durable by a fixed energy efficiency rate. For example, to compute the reduction in space heating energy from a warmer winter, an engineering approach would involve estimating the impact on the utilization rate and then multiplying the fixed unit energy efficiency by this new utilization rate. The main advantage of this approach is its high level of detail, although the accuracy is often in doubt given the limited data available to compute key parameters, such as utilization rates and unit energy efficiencies, particularly for older equipment.
Another approach develops statistical relationships often specified by economists on the basis of behavioral models that assume energy consumers, such as households and businesses, maximize their satisfaction conditional on heating and cooling degree days, relative fuel prices, income or sales levels, and the technical characteristics of energy-consuming durable equipment. Although this approach is more abstract than engineering models, it captures the behavioral dimensions of energy demand. Econometric models of energy demand are able to estimate the separate effects of price, income, weather variations, and other factors affecting energy demand.
The sensitivities of energy demand to price, weather, and other factors are expressed in terms of elasticity’s, defined as the percentage change in energy consumption associated with a given percentage change in one of these causal factors. The term associated is used to convey that the relationship is estimated based on a statistical model, which by definition allows a random error that reflects a number of uncertainties stemming from the inherent vagaries of human behavior. Such elasticities are a critical component in models of economic growth that are used to estimate the net social benefits of policies to mitigate or control greenhouse gas emissions. If a warmer climate is likely and if this results in a net reduction in energy demand, then the net social cost of greenhouse gas pollution would be lower by the energy expenditure savings. Determining the relevance and size of this potential effect has broad implications for climate change policies. Moreover, the feedback of climate change on energy demand also indirectly affects carbon emissions. Over the long term, this feedback effect may be an important element in understanding the global carbon cycle. In addition, future climate change agreements may need to allow some variance in carbon emissions due to normal climate fluctuations.
Many utilities and government agencies recognize the link between weather conditions and energy demand. Detecting the effects of climate trends on weather conditions is not without controversy, and the following section explains some of the nuances of detecting trends in heating and cooling degree days. Next, the formulation and construction of econometric models of energy demand that incorporate weather conditions are described in more detail. Estimates of price, income, and weather elasticities from these models are described along with numerical simulations of the impacts of climate change on U.S. energy demand. Finally, a discussion of the policy implications of the findings is presented.
