Forecasting GDP Using the "Climbing Limo" Method

Steve Conover, aka The Skeptical Optimist, has been looking at how the U.S. Bureau of Economic Analysis goes about projecting the growth of the United States' Gross Domestic Product (GDP), and has found a better way to forecast growth in the U.S. economy.

Beyond the graphics included with the post, which rank among the best I've seen in the blogosphere, the Skeptical Optimist's approach to evaluating his new forecasting technique is top notch. After developing several variations of how previous GDP data may be used to forecast future GDP levels, Conover tests each approach against historical data, which allows him to evaluate the effectiveness of each approach, and makes it possible for him to compare it with the BEA's own track record going back to 1948.

The results? Conover's two quarter method for forecasting GDP level three quarters into the future beat the BEA's one quarter method 55% of the time over the entire period considered. Since the first quarter of 1993, Conover reports his method proved more successful at forecasting future GDP levels in 63% of the cases.

Those kind of results deserve a calculator. Enter the seasonalized annual GDP data for the three consecutive quarters of interest below, and click the "Calculate" button for results determined by the Skeptical Optimist's "Climbing Limo" method:

Gross Domestic Product (GDP) Data
Seasonalized Annual GDP	Amount ($USD Billions)
Most Recent Quarter
One Quarter Ago
Two Quarters Ago

GDP and Annualized Real GDP Growth Rate
Estimated Results (Three Quarters from Now)	Values
GDP ($USD Billions)
Annualized Real GDP Growth (%)

The default data for the tool is taken for the period from the third quarter of 2004 (2004Q3) through the first quarter of 2005 (2005Q1).