Relationship between Energy Consumption and Energy Price

The notion that energy consumption is related to energy price is an intuitive one. All else equal, one would expect higher energy prices to result in lower consumption & vice versa. But just how sensitive is consumption to price in the case of energy? And how strong is the correlation between the two?

To explore these questions, I examined energy consumption and price data made available by user Joseph Lisle on Kaggle.com. Using his dataset, United States Energy, Census, and GDP 2010–2014, I’ll describe the relationship between these two variables using state-by-state energy consumption and price data from 2014.

Per Capita Energy Consumption by State (2014)

Between the highest energy-consuming state (Louisiana) and the lowest energy-consuming state (Arkansas), per capita consumption varies by a factor or nearly 4:1.

Energy Price per Unit, by State (2014)

Compared to energy consumption, state-by-state energy prices vary far less than energy consumption. Throwing out the state with the highest energy prices per unit, Hawaii, the variation between the states with the lowest & highest energy prices per unit is less than 2:1.

Relationship between Energy Consumption and Energy Price (2014)

To describe the relationship between these two variables, I fit an OLS linear regression to the data, as shown further below. The linear best fit does confirm the inverse relationship between energy price and energy consumption. More specifically, for every $1 increase in the price per million BTUs of energy, the regression estimates a decrease of -24.3M BTUs per capita in energy consumption.

The correlation coefficient and r² for this relationship are -0.488 & 0.238, respectively. This indicates that approximately one quarter of the variation in states’ per capita energy consumption can be explained by differences in energy price per unit.

Notably, there are a handful of states where energy consumption is markedly different than estimated by the linear regression. At the upper end of the spectrum of per capita energy consumption, Louisiana, Wyoming, North Dakota & Alaska have rates of consumption that are far higher than indicated by the regression estimate. With the exception of Alaska, these are all states that have relatively low energy prices.

At the other (lower) end of the spectrum, Hawaii has higher rates of consumption than indicated by the regression estimate. Hawaii is uniquely challenged when it comes to providing cheap energy to its residents — Hawaii’s per capita energy consumption may represent a floor with respect to just how little energy must be consumed on a per capita basis, despite very high energy prices per unit.