Visualizing GDP: Dissecting the Third Estimate Upward Revision

March 27th, 2014

by Doug Short

Note from dshort: The charts in this commentary have been updated to include the Q4 2013 Third Estimate.


The chart below is my way to visualize real GDP change since 2007. I've used a stacked column chart to segment the four major components of GDP with a dashed line overlay to show the sum of the four, which is real GDP itself. Here is the latest overview from the Bureau of Labor Statistics:

The increase in real GDP in the fourth quarter primarily reflected positive contributions from PCE, exports, and nonresidential fixed investment that were partly offset by negative contributions from federal government spending and residential fixed investment. Imports, which are a subtraction in the calculation of GDP, increased.

The deceleration in real GDP growth in the fourth quarter reflected a downturn in private inventory investment, a larger decrease in federal government spending, a downturn in residential fixed investment, and a deceleration in state and local government spending that were partly offset by accelerations in PCE and in exports, a deceleration in imports, and an acceleration in nonresidential fixed investment.

Let's take a closer look at the contributions of GDP of the four major subcomponents. My data source for this chart is the Excel file accompanying the BEA's latest GDP news release (see the links in the right column). Specifically, I used Table 2: Contributions to Percent Change in Real Gross Domestic Product.

Over the time frame of this chart, the Personal Consumption Expenditures (PCE) component has shown the most consistent correlation with real GDP itself. When PCE has been positive, GDP has usually been positive, and vice versa. In the latest GDP data, the contribution of PCE came at 2.22 of the 2.63 real GDP. The Q4 contribution from PCE increased significantly from Q3, but the Second Estimate for Q4 showed a decline from the first estimate.


Note : The conventional practice is to round GDP to one decimal place, the latest at 2.6. The 2.63 GDP in the chart above is the real GDP calculated to two decimal places based on the BEA chained 2009 dollar data series.

To help us understand the Q4 revisions, here's a side-by-side look at the Advance, Second, and Third Estimate subcomponents. PCE was significantly trimmed in the Second Estimate but bounced back in the Third estimate. The reverse was the case for Private Investment. Ultimately net exports was the biggest difference maker between the Advance and Third Estimates, going from 1.33 to 0.99.

Here is a look at the contribution changes between over the past four quarters. The difference between the two rightmost columns was addressed in the GDP summary quoted above. I've added arrows to highlight the quarter-over-quarter change for the major components.

As for the role of Personal Consumption Expenditures (PCE) in GDP and how it has increased over time, here is a snapshot of the PCE-to-GDP ratio since the inception of quarterly GDP in 1947. The Q4 2013 ratio is 67.9%, fractionally off the all-time high of 68.6% in Q1 2011. From a theoretical perspective, there is a point at which personal consumption as a percent of GDP can't really go any higher. We may be hovering in that upper range.

Let's close with a look at the inverse behavior of PCE and Gross Private Investment (GPDI) during recessions (note my use of different vertical scales to facilitate the overlay). PCE generally increases as a percent of GDP whereas Private Investment declines. That is not what we're seeing in the current data. I've plotted the two with different vertical axes (PCE on left, GPDI on the right) to highlight the frequent inverse correlation.

With Q4 of 2013 now put to bed (until the July annual revisions), up next will be the first look at the controversial Q1. I say "controversial" primarily because of the debate over the impact of weather on the economy.

Website by the Boston Web Company