The Composition of GDP During the Recession
The big news today is the advance estimate of second quarter GDP, available here. The top line number is that GDP growth was -1.0% at an annual rate in the second quarter, an improvement from the growth rate of -6.4% in the first quarter. I think it is too early to identify the end of the second quarter as a business cycle trough -- we need to see whether GDP growth turns positive in the third quarter and whether there was a pickup in the other recession indicators.
I wanted to take the occasion of this news release to dig a bit into the composition of GDP. The following table (click on the thumbnail for a larger image) shows some data from Tables 1.1.1 and 1.1.10 of the National Income and Product Accounts. The recession was precipitated by a fall in private investment, which includes equipment & software, inventories, and both residential and non-residential structures. The table begins in 2006-I, the quarter in which private investment peaked. The top panel shows the quarterly growth rates of each GDP component in the intervening period. What is most surprising about the table is how little growth there has been in expenditures by state and local governments since the recession began -- in fact, quarterly growth has been negative in half of the quarters of the recession.
 |
The bottom two rows show the composition of GDP in the quarter prior to the peak in private investment and the latest quarter. Investment's share has fallen by 6.4 percentage points of GDP. How has that been reallocated? Consumption is higher by 0.8, imports are lower by 3.7, federal government expenditures are higher by 1.2, and state and local government expenditures are higher by 0.7 percentage points. My hope going into the recession was that government expenditures, preferably on capital projects, would have played more of a role, generating much larger numbers in the last two columns.
private investment
wrote similarly in my blog today econmkts.blogspot.com
Investment
That's a good idea to look at the shares. I would add that the 11.2% share of investment is the lowest ever in the data series going back to 1947 (mainly due to housing, nonresidential fixed investment was 9.9% of GDP, which is lower than average, but not alarmingly so). At 70.6%, the share of consumption is the highest ever - even though consumption has fallen, output has fallen more.
RE Composition of GDP in recession
Never replied to a blog before so I don't know if this will show up in the right place.
Depending on how we measure infrastructure spending, about 70-90% of public infrastructure spending is by state and local gov'ts, not the feds, so we have to look there to see what is going on. Here, we need to think about stimulus, since the stimulus package provides about $54b in direct money for SLG capital spending. But it turns out that at this stage of the cycle, stimulus is having an impact on S&L gov't expenditures, but maybe not for the reasons one might think.
In 2009q2, real GDP fell by about $32 billion (b), while state and local government consumption and gross investment spending rose by $9.3b (Table 1.1.6). Essentially all of the SLG increase was in investment in structures, which rose by $9b (Table 3.9.6).
These numbers are at seasonally adjusted annual rates. Roughly speaking, it suggests actual spending on structures in the quarter rose by one quarter of $9b, or about $2.25b (sorry, I don't have seasonally unadjusted data from BEA, which would allow us to get the actual spending change rather than dividing by 4).
Now, let's look at stimulus spending for SLGs. I don't have Apr-Jun spending readily available (although it can be had), but stimulus spending on infrastructure from the start of stimulus (Feb) through latest week available (July 24) was $1.02b (see various status reports at www.recovery.gov). Almost all of that would have been in Apr-Jun.
About half of that ($510 million) was for highways and $420m was for other transportation, where there is relatively little spending on infrastructure. So the amount spent by SLGs from stimulus on structures was close to zero (rounding to the nearest $ billion) and is not the cause of the increase in SLG spending in the latest GDP report.
What, then, is going on, and did stimulus play a role?
First, there has been a lot of stimulus spending for SLGs - about $36b though July 24. Almost all of that is in the "fiscal relief" provisions of the stimulus package - spending ostensibly to support Medicaid and education via the State Fiscal Stabilization Fund. (As a practical matter, this money is flexible budget relief.)
This certainly would affect GDP, mostly by staving off cuts in spending (might have staved off some tax increases, too, but few could have taken effect in the Apr-Jun quarter). In other words, we might have seen SLG consumption spending decline but for stimulus spending. This is definitely a positive impact of stimulus, but not evidenced through capital expenditures.
Second, even the increase in structures spending could be stimulus related in an indirect sense, but that is purely speculative. There is some evidence in research and data that state and local governments postpone some capital spending during recessions (see, for example, a working paper by McGranahan from the Chicago Fed a few years back). My personal experience as a former gov't financial manager suggests it is true, also. Perhaps SLGs did some of this in the Jan-Mar quarter - in this case, fear-factor driven - and then, with the stimulus package enacted, decided they could go forward even though stimulus money had not yet arrived, raising the Apr-Jun quarter relative to Jan-Mar.
Recap: (a) state and local gov'ts clearly have been less of a drag than we might otherwise expect due to the $36b of federal SLG stimulus outlays so far, and (b) the drag-reduction so far is occurring through general budget relief, not capital spending. (It is simply too early for significant SLG capital spending.)