Both politicians and commentators frequently decry the lack of economic growth in the U.S. since the end of the Great Recession in 2009. Those voices include many responsible ones, such as Seeking Alpha’s John M. Mason, who wrote on the subject on August 30. Mr. Mason, like many others, blames a lack of corporate capital spending—and its substitute, financial engineering through share buybacks and the like—for a significant part of the slowdown in growth.
Frankly, I wonder whether the lack of corporate capital spending is a main cause. And I wonder whether the low growth has really been so low. In many respects, I subscribe to Robert J. Gordon’s headwinds theory (see, e.g., The Rise and Fall of American Growth), and specifically to the demographic aspects of that theory.
Annual real GDP Growth since 1970
Taking a look at annual real GDP growth (the usual measure of growth, though not necessarily my favorite) from 1970 (I use 1970 as the end of the easy post-war growth period), through 2015, the average is 2.8%. If we break down the 46-year period into two components, 1970 to 1999 (30 years) and 2000 to 2015 (16 years), the averages are 3.3% and 1.8%–clearly a significant slowdown, with serious consequences for people’s potential standard of living. (Data from FRED, averages computed by the author)
The two 15-year periods from 1970 through 1984 and 1985 through 1999 exhibit no such wide swings. The first 15 years show average real GDP growth of 3.3%, despite serious recessions in 1973-74 and the milder double-dip recession at the beginning of the 1980s. The second 15-year period was slightly better, with closer to 3.4% average annual growth, in part because it had only one (fairly mild) recession and in part because it included the best growth era since 1970— the productivity leap forward of 1996-1999.
Onset of the slowdown—the year 2000
The slowdown did not begin after the Great Recession. It began much earlier in 2000. The 10-year period 2000 through 2009, which included the boom and the bust, averaged 1.7% growth, significantly lower than the 2.1% growth averaged during the seven-year slow-growth recovery period of 2010 through 2016 that so often is decried. That the watershed was 2000, not the Great Recession or the Great Financial Crisis, tends to be substantiated by Doug Short’s excellent chart of his “Big 4” indicators. Here it is from 1960 to date.
As you can see, the slope of the line changed in 2000, indicating slower growth beginning in that year, not later.
Why has growth slowed since 2000? I do not purport to have all the answers.
One has to begin by noting that GDP growth depends on two components: growth in hours worked and growth in productivity. Both the demographic factors that affect hours worked and productivity growth have been slow in the period since 2000.
The productivity conundrum has been written about a great deal. I last wrote on productivity growth a year ago, on August 11, 2016, and cited a number of papers and analyses at that time. Basically, I said that productivity growth tends to be cyclical, and that, in part because the people who get rehired after a recession are less productive than the people who were not laid off, productivity gains come late in a recovery—at least the way we conventionally measure it. Therefore maybe the productivity numbers are about to take a turn for the better in the near future.
But today I want to discuss the other side of the coin—the demographic factors that influence hours worked.
When the population is growing and a large segment of the population, such as women, is increasing its workforce participation, it is fairly easy for the demographic aspects of hours worked to propel the GDP numbers forward. That was the case from 1970 through 1999. During that period, population growth averaged 1.06% a year. But after 1999 that statistic declined significantly. The period 2000 through 2015 showed annual population growth of only .85%. And since the Great Recession, annual population growth has been only .74%.
Thus the U.S. has experienced a .3% per year slowdown of population growth. That may be good for the environment, but it is not good for GDP growth.
Growth of the working population (age 15 to 64) (again, figures from FRED, computations by the author) has declined more precipitously than general population growth. It has declined from an average of 1.9% per year in the 1970 through 1999 period to .5% in the 2000 through 2015 period. That decline basically would account for the entire decline in measured real GDP growth. (In the 2010 through 2015 period, the growth rate has gone back to a more substantial and normal 1.2%.)
As you can see from the following Doug Short chart of the labor force participation rate, it peaked in early 2000 and began its descent from there.
Of course one can (and should) ask why the working population growth has declined more than the total population growth. Is that the result of a lack of jobs? Is it the result of technological obsolescence of some workers? Is it a result of lifestyle preferences? Is it a result of incarceration or drugs? Are the jobs being taken by old people? Are there just more young people in school and therefore not in the workforce? Or other factors? Again, I wish I had all the answers. But we can do some useful analysis by focusing on where some of the big differences might be. This kind of social science is approximate. One should not expect absolute answers.
Women in the workforce
One potentially influential area is the trajectory of women in the workforce. As you can see from the following chart from the DOL (I have redacted the ethnicity data and left only the overall data), women’s labor participation rate has formed a buttonhook pattern. From about 43% in 1970, women’s labor participation rate grew to about 60% in the late 1990s, then began a slow decline that accelerated during and after the Great Recession, to stand at about 57% in 2015. Thus women’s participation rate drove a great deal of the real GDP growth from 1970 to 1999 and accounts for some (perhaps a material part of) of the reduction in GDP growth rate since 1999.
Older women, however, have been bucking the trend, as the following Doug Short chart indicates. Like men, older women are seeking to remain in the labor force longer than they did before 2000.
The following chart shows the data for older men and women combined.
Thus, there is a possibility that older people remaining in the labor force longer is discouraging younger people from participating because they are finding it hard to compete with the experience and skills of the older workers who choose to remain in the labor force. Personally, I find that a little far-fetched, but it is logically possible.
Is there an incarceration cause?
The potential incarceration story does not hold up under scrutiny. Incarceration may indeed hold the economy back—and hold back in particular those ethnicities that are most often incarcerated. But as the following chart from the Prison Policy Institute shows, incarceration rates rose steeply from the mid-1970s through the late 1990s, but then they leveled off precisely when the GDP growth slowdown began in 2000. No correlation there; so no likely causation.
According to the U.S. Department of Health and Human Services (see here), 12.5 million people misused prescription opioids in 2015, and 2 million had “prescription opioid use disorder” (for the definition of this disorder, as provided by the American Psychiatric Association, see here). My layperson’s reading of the definition suggests that the employability of a person with prescription opioid use disorder would be questionable. Therefore the number of people whose employment status is impaired by opioid abuse probably is at least in the 2 million range—but it may be higher, since there is a lot of room between the 12.5 million people “misusing” and 2 million people with the “disorder”. And it is pretty clear from the following graph from The Economist that the problem has been accelerating since 2000.
Why the year 2000? I have no answer.
Thus the opioid problem plays a potentially major role in reducing the workforce participation rate—probably by 100 basis points (1%) or more. That accounts for at least one-quarter of the 4% reduction in the participation rate since 2000—and probably more—perhaps even half of the reduction. (The workforce is approximately 160 million people — see here.)
College population reducing labor force participation
The college participation rate also is a promising source of the workforce participation rate reduction: “In fall 2017, some 20.4 million students are expected to attend American colleges and universities, constituting an increase of about 5.1 million since fall 2000,” says the National Center for Education Statistics. That increase takes about 3% of the potential working population of 160 million (see here) temporarily out of the workforce, assuming they are not also working at least part time. It therefore potentially accounts for a large part of the 4% decline in the workforce participation rate from 67% in 2000 to 63% in 2017.
And since a majority of that 5.1 million-student increase are women (probably about 57%), it also accounts for a substantial part of the buttonhook picture of women’s labor participation that we saw above.
And the increase in college participation is a good thing. It means that in the future, America will have a more educated workforce.
Is it simply a case of too few jobs?
Creating more good jobs is always a big deal. It is always a primary goal of government to foster more jobs. I wrote about that fairly recently here. But it is hard to tie the “too few jobs” story to 2000, when employment was (at least ostensibly) good in the mid-2000s. We can almost tie the beginning of the “China did it” story to 2000, since China entered the WTO in 2001. But the recent inability of employers to fill jobs (we have a very large unfilled jobs backlog—over 6 million—and a low unemployment rate) suggests that the jobs issue is much deeper than just a number of jobs. At least to some extent, there is a mismatch between what employers need and what job seekers have to offer.
This section easily could be a series of articles by itself—and many people are writing about the subject. For now, I will leave it that the slow GDP growth can be explained by factors not related to a general paucity of job opportunities.
The slow growth period dates to 2000, not to 2010
So what is the bottom line here, after all this sashaying hither and yon?
First, the slow growth period dates to 2000, not to 2010. If we want to know what happened, it would be good to look in the right time period.
Demography is likely a material cause of the slow-growth period
Second, are there demographic likely causes for a substantial part of the reduction in the real GDP growth rate since 2000? The answer is “yes.” We have found that there is less population growth and more student population growth, which, combined, could account for a reduction of 1% in the annual real GDP growth rate, which is approximately the decline since 2000. The opioid problem might account for a substantial part of the workforce participation rate, perhaps thereby accounting for the entire slowdown with demographic causes. Indeed, if we just add up the reasonable demographic possible causes, we could conclude that the economy is doing amazingly well despite such strong headwinds.
That does not tell us “why” the year 2000. Many data sets begin to look different at some time during that year. And that year followed four excellent years for GDP growth and productivity growth. The good might be blamed on Y2K expenditures or something like that. But why did the negatives begin in 2000?
My guess is that something cultural happened that is not yet well understood. But I am sure that you gumshoes out there in seekingalphaland are going to educate me on that.
But there may be a third lesson
If demography is a material part of the growth slowdown causes, then maybe all the hand-wringing and policy-wonking about what to do about it is off base. Maybe the economy is really pretty good. No recession in 8 years, perhaps none in prospect if some global political or military event doesn’t blow up the relative prosperity. Yes, there have been many people left behind. And our society should do things to help them get back into the workforce and the economy. But perhaps the perfect is, as so often, the enemy of the good.
(Look at that! I wrote a long article about the economy without mentioning the Fed, Congress or the ubiquitous President.
Aww, now you ruined it!.)
Disclosure: I/we have no positions in any stocks mentioned, and no plans to initiate any positions within the next 72 hours.
I wrote this article myself, and it expresses my own opinions. I am not receiving compensation for it (other than from Seeking Alpha). I have no business relationship with any company whose stock is mentioned in this article.