The U.S. Department of Commerce just released its numbers for the American economy in the third quarter. The media reports, such as this one by Fox Business, explain that inflation-adjusted GDP growth fell modestly from 3.0% in the second quarter to “an annual rate of 2.8% in the third quarter” of this year.
All in all, the media makes this, the last GDP report before the election, look relatively good.
There is just one problem—and it is the same problem that I have pointed to over and over: The 2.8% number is not the growth in GDP from last year’s third quarter. It is the growth in GDP from the second quarter of this year, multiplied by four to make it look like an annual growth rate.
I have long criticized the Department of Commerce’s Bureau of Economic Analysis, BEA, for using these annualized numbers. They conceal reality and add nothing of statistical value. My criticism is once again validated when we look at the actual annual growth in real GDP from the third quarter of last year to this year’s third quarter. We find that number in the BEA’s National Income and Product Accounts (NIPA) Table 8.
According to this stack of GDP data, which is as straight out of reality as it can be, the U.S. economy only grew by an inflation-adjusted annual 2.4%.
This is quite a bit worse than the 2.8%. To make matters worse, the actual number for the second quarter was 3.0%, which is the same as the artificial, annualized number that the BEA produced for the same period. This means that the economy geared down its growth not by a minute 0.2 percentage points, but by a statistically meaningful 0.6 percentage points.
The smaller decline suggests that overall, the economy is doing well, while the larger decline indicates that we are now in the first stage of an economic stagnation, or even a recession.
My criticism of the BEA’s annualization practices—or statistical rituals—is only reinforced by a closer look at the BEA’s press release data. They claim that inflation-adjusted private consumption grew by an annual 3.7% over the third quarter of 2023. But the raw data from NIPA Table 8 place the third-quarter growth rate at a modest 2.9%.
Since both the annualized (ritualized) data and the raw data report 2.8% for the second quarter, the difference between the artificial but official 3.7% and the actual 2.9% again becomes important for any analysis of where the economy is heading:
- The artificial, annualized number suggests strong, resilient consumer spending, which in turn implies optimism just before the election;
- The actual numbers from NIPA Table 8 tell us that consumers are being careful but not pessimistic with their money.
While the artificial number would suggest consumers are happy in the Biden-Harris economy, the actual number is what you would expect if consumers are worried about a looming economic downturn. The latter point is reinforced by details in the GDP numbers, which reveal a more modest growth in spending on services. Since they account for two-thirds of all consumer spending, this is also what one would expect if consumers are slowly getting more worried about a recession.
An even stranger statistical discrepancy shows up under government spending. The official, artificial BEA number tells us that in the third quarter, national defense spending increased by 14.9% year over year. This pushed total federal spending up by 5%, a high number on any day and a very high number compared to 6.4% in the second quarter and -2.5% in the first.
The only problem is that if we disregard what the BEA comes up with using statistical rituals, the real-world increase in defense spending is only 4.4%. The consequence for total federal spending is similarly disappointing: it ‘only’ grew by 3.1%.
Government spending is part of GDP, just like private consumption, business investments, and net exports. The difference between 5% and 3% is big, and when we consider that it conspires with the artificially exaggerated private consumption figure, the resulting GDP number looks like a left-leaning economist’s dream just before an election.
For clarity, I do not believe that the Bureau of Economic Analysis is performing statistical rituals for the deliberate purpose of benefitting one political side over the other. However, with such glaring inconsistencies as I have pointed to here, the good folks at the BEA should probably take a serious look at how it produces its so-called advance estimates of GDP. There has been a debate in the past year—though mostly in social media—over the major revisions that both the BEA and the Bureau of Labor Statistics have been forced to engage in.
Bad GDP data have many consequences, which tend to increase exponentially with the deterioration of the statistical product. One of them is that our politicians do not get an accurate image of the economy when they make decisions on taxes and spending. When, as in this case, the BEA significantly exaggerates GDP growth, Congress is given the false impression that things are good enough to spend more money—and worry less about the budget deficit.
There are also consequences in the private sector. Misinformation about GDP easily leads people to make bad decisions on stock market investment. Businesses can be misled as to where they should allocate their job-creating capital formation. And so on.
Expect major revisions of the GDP figures included in the BEA’s advance estimate. But don’t expect them before the election.
