Ed Dolan's Econ Blog

How GDP Revisions Change Our View of the Great Recession: The Story in Charts

On July 31, the Bureau of Economic Analysis released revised data for US national income accounts. The revised data give us a new view of the Great Recession that began at the end of 2007. It still merits its name as the most severe economic downturn since the Great Depression of the 1930s, but the contraction now looks a little shallower than previously thought and the recovery a little more robust.

The following chart compares the old and revised real GDP data over the past six years. The old and new data series are not directly comparable. Not only was the old series stated in 2005 dollars and the new in 2009 dollars, but there are numerous statistical and methodological differences as well, as discussed below. For easier comparison, then, the chart displays both the old and new data in the form of an index with the peak of the previous cycle, Q4 2007, equal to 100.

Several features stand out in the chart. First, the contraction from peak to trough was not quite as deep as reported earlier. Instead of falling by 4.7 percent, real output fell by 4.3 percent. Beginning from the trough, which came in Q2 2009 in both series, the expansion is somewhat stronger according to the new data, especially in 2011. From Q1 2011 to Q1 2012, the economy is now seen to have grown by 3.3 percent rather than the previously reported 2.5 percent. By Q1 2013, real GDP was 3.9 percent above the previous peak, rather than just 3 percent, as reported earlier.

Another interesting feature of the chart is a dip of 0.4 percent in real GDP in the first quarter of 2011. At the time, growth for that quarter was reported as a positive 0.1 percent. Perhaps we should be grateful to have been spared the alarm bells that would have rung forth if the dip had been reported when it happened. As it was, the one-quarter downturn did not turn into a double-dip recession, as it would have been labeled, at least informally, if it had continued for two quarters.

Why the revisions?

Why, one might ask, are these revisions necessary in the first place? Why are we just now learning that real GDP five years ago was larger than we thought it was at the time? The answers to these questions reveal some important things about how GDP numbers are assembled and how they are measured.

Although GDP is often informally characterized as the total value of everything the economy produces, it is more accurate to say that it is the total value of all final goods produced. Final goods are those that are used for the purposes of consumption, investment, government consumption and gross investment, and exports, with imports subtracted from the total. GDP from any one sector of the economy, say manufacturing, is equal to the total value added in that sector, that is, the value of the goods in produces minus the value of intermediate goods used up in the process, such as lumber used to make furniture, coal used to make steel, and so on.

In practice, it is easier, on a quarter-by-quarter basis, to get data on total output of each sector than on value added. The detailed input-output studies needed to estimate value added, given total output, are conducted only every five years or so. The new GDP data are based on a new set of input-output benchmarks. But hold your hat—these brand new benchmarks relate to the input-output structure of the economy as it was in 2007, a full six years ago. It is a sobering thought that any structural changes in the economy brought about by the Great Recession itself—and surely there have been many—are not reflected in the new GDP numbers. It will be at least another five years before they show up in the data.

In addition to the rebenchmarking of input-output data, the new GDP numbers incorporate several methodological changes. The largest and most widely discussed is the decision to treat spending on intellectual property like patents and copyrights as investment rather than as current expenses. Additional changes relate to the way the BLS measures banking services, pension contributions, and alternative energy installations.

Sector-by-sector revisions

The GDP revisions not only change our picture of the recession and recovery as a whole, but also of the way various sectors of the economy have contributed to it. The following charts look at each major sector in turn.

First, personal consumption expenditures. Consumption accounts for about 68 percent of total GDP, and since the recovery began in 2009, it has accounted for just about the same share of GDP growth. The revisions do not dramatically change the picture. On average, the contribution of consumption to the contraction from 2007 to 2009 was a little less than previously reported. During the recovery, its contribution has averaged a little higher than previously reported, although that has not been true for the last two quarters.

Next, gross private domestic investment. The revisions show that investment, as a whole, has contributed a bit less to the cycle than previously reported, both during the contraction and the subsequent recovery. Since the recovery began in 2009, the arithmetic average contribution of investment to GDP growth has been about 0.8 percentage points rather than the previously reported 0.9 percentage points. Looking still more closely at the data, fixed investment, including housing, has contributed more to the recovery than previously thought, while inventory investment has contributed a little less. Notice also that investment is much more volatile than consumption. Even though investment makes up only about 15 percent of total GDP, we can see by comparing the charts that its contribution to growth during both the contraction and the recovery has often been greater than that of consumption.

Government consumption expenditure and gross investment is the next sector to consider. This component of GDP, often known to economists as government purchases of goods and services, includes all the goods and services purchased by government from the private sector plus the salaries of all government employees. It excludes transfer payments like unemployment benefits and social security payments, since these do not represent additions to production.

As the next chart shows, from 2007 through 2009, under both the Bush and Obama administrations, government was a net contributor to economic growth. The impact peaked in Q2 2009, as the Obama administration implemented its stimulus program. However, since late 2009, the impact has been largely negative.

Economists refer to a negative impact of government purchases on growth as “fiscal drag.” Fiscal drag has been a consistent factor slowing the pace of recovery, although the revisions show that it is not quite as strong as previously estimated. Fiscal drag has occurred at all levels of government. The latest data indicate that state and local governments have played an even bigger role than the federal government. Total fiscal drag, starting from 2010, when the initial impact of stimulus programs ended, has averaged about -0.4 percentage points. Of that, state and local government have accounted for about two-thirds and the federal government for about one-third. (However, Q2 2013 data, not shown here, suggest that state and local government purchases may be stabilizing, whereas fiscal drag from the federal government continues.)

Finally, we come to a relative bright spot in the picture of the recovery. As this final chart shows, exports contributed positively to growth from the start of the recovery in 2009 through the end of 2012. Their average contribution is little changed by the latest revisions. However, it is clear that as the eurozone has sunk into recession and many emerging market economies have begun to slow, the positive contributions of exports is fading. The revisions erased a reported decline in exports in Q4 2012, but confirmed a slight contraction in Q1 2013, which was the first since the recovery began in mid-2009. The advance report for Q2 2013, not shown in the chart, indicated that exports were growing again, but we will have to wait to see if that number holds up.

 The bottom line

There are two things to take away from this discussion of the revisions to GDP data. First, they confirm that the Great Recession has been deep and the recovery from it slow, even though the picture is a little better than previously thought. Second, we are reminded that measurement of economic activity is an imperfect art. The staff of the BEA are thought by many to produce data that is as good as or better than those of any other country, but still, they is far from perfect. GDP data are inevitably based on information that is not fully up to date and on methodological decisions that are, to some degree, arbitrary. We must always remember that they do not try to do more than measure the value added by producers of goods and services. They were never intended to be a comprehensive barometer of national economic well-being.


8 Responses to “How GDP Revisions Change Our View of the Great Recession: The Story in Charts”

benleetAugust 2nd, 2013 at 1:39 pm

The median household's wealth or net worth dropped by almost 40% 2007 to 2010 and has not recovered. Average household debt to disposable income still hovers around 106.3%. And real wages between 2007 and 2012 for 60% of workers dropped by 3.7% according to the EPI study here:… — Disposable income grew 2007 to 2012 by 2.4% while GDP grew by 4.0% according to the And the employment to population ratio dived precipitously, and has not rebounded at all. Professor Saez says that nearly all of the gains in the recent 3 years have gone to the top 1%. And investment (net of depreciation in existing plant and machinery) is at levels of 1930s, "net investment as a share of GDP has plummeted to its lowest level since the 1930s. This sharp drop in investment comes despite sharply rising profits", according to this article at Dollars and Sense:
So the GDP growth is not a comprehensive barometer. As usual, Ed Dolan writes a very clear and concise analysis.

FrankAugust 2nd, 2013 at 4:24 pm

You mention the fact that pension contributions are to be measured differently. Do you have any comments on this change?

TomAugust 3rd, 2013 at 6:40 pm

I would say that investment and final goods are not at all the same things. Investment is acquisition of capital goods that are intended to be used in production processes over a long period of time. A capital good is really a kind of lasting intermediate good.

This revision brings GDP accounts somewhat closer to business income accounts, but it also increases the importance of taking the next big step, which is to use net domestic product as the main measure of economic progress.

jackstrawAugust 3rd, 2013 at 9:02 pm

again…absolutely top shelf economic work. i like to sum it up this way "it wasn't the Great Depression." but it sure could have been. without a doubt the US economy "was wrecked." also without a doubt the American people….like out of some movie…"found a way to will itself to recovery." obviously this doesn't get pointed out by the "bailout tycoons" and just as obviously there will never be an end to me pointing out this fact and my (and i hope all of yours) wonder at it. who wouldn't want North Dakota to broaden and deepen? nothing "compels" success even after all past effort and sacrifice (and North Dakota has without a doubt.) so who are we NOT to try and emulate it? who are we NOT to extol and want to wonder? is it because "after every downturn there is an upturn"? the Fed certainly allowed for this to happen. but who was there to "understand" that? who pushed "the on button" here? i would argue not only "no one" but that such a thing is impossible. "but by the Grace of God go we."

Aaruhi MishtaAugust 18th, 2013 at 4:27 am

"GDP is inevitably based on information that isn't finally uptodate."Any comments.And moreover DOLAN writes a concise analysis

Most Read | Featured | Popular