Skip to main content
ABC News
The Messy Truth Behind GDP Data

The economy grew at a 2.6 percent annualized rate in the last three months of 2013, according to data released Thursday by the Bureau of Economic Analysis (BEA). But this was not the first estimate for that quarter’s gross domestic product, which measures the total value of goods and services the U.S. economy produces. Nor will it be the last.

The BEA releases multiple estimates of each quarter’s GDP, one each month beginning the month after each quarter ends. The BEA also issues annual revisions each July and benchmark revisions every few years. Those revisions are often substantial.

Thanks to the Federal Reserve Bank of Philadelphia’s Real-Time Data Research Center, we can look at how GDP estimates typically change. Plotted below is the average change in quarterly GDP growth across the BEA’s two follow-up revisions to the first estimate, as well as the historical revision (which incorporates the annual and benchmark changes). What is measured is the absolute value of the change (whether positive or negative). The data plotted starts in 1975; prior to 1975, the BEA made no distinction between the first and second estimates in most cases.


From the initial estimate to the first revision, the average absolute change is a little over 0.50 percent. From the first to the second revision, it changes on average nearly a quarter percentage point.

What’s most astounding is how much GDP changes from the third “real time” estimate to its historical estimate as refined by annual and benchmark revisions: nearly 1.5 percentage points. To put that in context, the average quarterly growth rate since 1975 is 2.7 percent. So GDP numbers in real time are, at best, a dim reflection of the state of the economy.

Andrew Flowers wrote about economics and sports for FiveThirtyEight.