The Dangers of Looking Backwards While Going Forward

While we often make fun of economic prognosticators, we thought we'd take the opportunity to deflate our preferred method of forecasting future levels of GDP. Consider today's news that economic growth in the U.S. in the fourth quarter of 2006 (2006-Q4) grew at an annual rate of 3.5% to a level of 11,541.6 billion dollars after adjusting for inflation.

Now, this advance estimate is quite a bit off from the 11,625.2 billion dollar level the Climbing Limo method had forecast for the U.S. economy in the fourth quarter of 2006. This 0.7% error was produced using a model that only looked at inflation-adjusted GDP data for the third quarter of 2005 (Real GDP was $11,115.1 billion in 2005-Q3) and the first quarter of 2006 (Real GDP was 11,316.4 billion in 2006-Q1).

So, how did the forecast get so far off from where we've now seen the advance GDP data come in? Simply put, two big things have changed between the U.S. economy of 2006-Q4 and the U.S. economy that existed between 2005-Q3 and 2006-Q1: the housing market has decelerated and motor vehicle production in the U.S. has become a drag upon the economy.

In a nutshell, that's pretty much it. The Reuters article underscores this point with respect to domestic automobile production, noting that:

... motor vehicle output restrained GDP sharply. Excluding motor vehicle production, GDP would have been up at a 4.8 percent rate during the quarter.

Now, all this is just advance data and at this point, we would anticipate that the final numbers that will be released on March 29, 2007 will be revised upward, but not to the levels the Climbing Limo method had forecast.

In the meantime, here's our chart updated with the advance Real GDP data and an extra point showing where the economy could have been if not for Detroit's most recent problems. For added measure, we've included a datapoint showing where the economic consensus poll was forecasting when we got around to posting our prediction:

In reality, all this highlights the key weakness of any forecasting method that only looks at past data. Not only was the Climbing Limo off by a substantial margin, the collective wisdom of economic forecasters was overly pessimistic, even though they presumably were using far more up-to-date data in their own predictions. As they say in the investing industry, "past performance is no guarantee of future results."

Previously on Political Calculations

• Looking Backward While Going Forward