The Uncertainty of Forecasts

I’ve long known that economic forecasting is as much art as science, and that forecasts should be viewed skeptically. A recent chart I compiled, however, really drove home this point.

I’ve long known that economic forecasting is as much art as science, and that forecasts should be viewed skeptically. A recent chart I compiled, however, really drove home this point.

Data from the U.S. Department of Energy’s Energy Information Administration (EIA) is widely used by government, industry and non-profits. Drawing on voluminous records of historic data, the EIA provides short- and long-term forecasts about fuel prices, demand for different fuels, and regional growth patterns. The forecasts incorporate the results of decades-long policies and try to gauge the impact of newly adopted laws and regulations. They also try to anticipate technology trends and global changes that could affect the U.S. While I sometimes may quibble with the results of the EIA’s forecasts, I don’t envy their task of assimilating so much information and making predictions.

I recently compared data about historic natural gas prices to historic forecasts. In EIA’s 2002 Annual Energy Outlook, the agency predicted that electricity generators would pay a fairly steady, low price for natural gas from 2002 to 2010, in the range of $3 to $4 per thousand cubic feet. What really happened, though, was that from late 2002 to early 2009, generators paid above $6 per thousand cubic feet, and at times paid above $12. The figure below displays the forecasted price and the actual price.

 

The chart also shows the EIA’s latest prediction for the price that electricity generators will pay for natural gas in the next decade. Seeing the current forecast in the context of the 2002-2010 forecast and actual prices reminds me why we shouldn’t put too much faith in such forecasts, whether we agree with them or not. If we recognize and acknowledge the degree of uncertainty in forecasts, we’ll be better able to pursue policies that limit risk and protect us from the costs—financial or otherwise—of unexpected scenarios.

Historic prices paid for natural gas by electricity generators from U.S. Department of Energy, Energy Information Administration, U.S. Natural Gas Prices (spreadsheet), downloaded from www.eia.gov/dnav/ng/ng_pri_sum_dcu_nus_m.htm, release date 28 February 2011. Forecast data for national prices from 2002: U.S. Department of Energy, Energy Information Administration, Annual Energy Outlook 2002, Supplemental Table 96. Natural Gas Delivered Prices by End-Use Sector and Census Division, 21 December 2001. Forecast data for national prices from 2011: U.S. Department of Energy, Energy Information Administration, Annual Energy Outlook 2011, Natural Gas Supply, Disposition and Prices, AEO2011 Reference Case, 16 December 2011. All prices are presented in 2009 dollars. Historic actual prices and the 2002 forecast were converted to 2009 dollars using Consumer Price Index data from the Federal Reserve Bank of Minneapolis, What Is a Dollar Worth?, downloaded from www.minneapolisfed.org/community_education/teacher/calc/, 11 March 2011.

EKRPIC1.jpg
 
 

 

Topics
Authors

Elizabeth Ridlington

Associate Director and Senior Policy Analyst, Frontier Group

Elizabeth Ridlington is associate director and senior policy analyst with Frontier Group. She focuses primarily on global warming, toxics, health care and clean vehicles, and has written dozens of reports on these and other subjects. Elizabeth graduated with honors from Harvard with a degree in government. She joined Frontier Group in 2002. She lives in Northern California with her son.

Find Out More