Editor's Note: Below is a Hedgeye Guest Contributor research note written by Dr. Daniel Thornton. During his 33-year career at the St. Louis Fed, Thornton served as vice president and economic advisor. He currently runs D.L. Thornton Economics, an economic research consultancy.
On the op ed. page of the Wall Street Journal (WSJ, August 11), John Cochrane suggested that “America’s foremosteconomic problem is sclerotic growth,” saying that “if the economy continues to expand at only a 1% or 2% a year, instead of the historical 3% to 4%, then current economic and political problems will become crises.” He went on to say why Clinton’s economic programs for stimulating economic growth wouldn’t work and suggested how growth could be spurred. This essay is the first of two motivated by Cochrane’s opinion piece. This essay addresses the question of whether 3% to 4% economic growth is feasible. The second essay will consider whether Clinton’s or Cochrane’s proposals will succeed in increasing economic growth.
Figure 1 shows the annual growth rate of real GDP quarterly from 1948 Q1 to 2016 Q2. The figure suggests that economic growth has been trending down over the entire period. The peaks in output growth have tended to decline over time while the troughs tend to be more constant. The figure shows two nearly identical trend lines: The dashed line is estimated using data from 1948 Q1 to 1994 Q4 and extrapolated to 2016 Q2. The solid blueline is estimated using all of the data. The implication: If the dashed line represents the trend in output through 1994 Q4 and growth remained on trend, which it appears it has, then economic growth is back to trend.
The conclusion that output growth has trended down is supported by Figure 2, which I used in Slow Growth. Figure 2 shows a 6-year moving average of output growth and a trend line estimated using data through 1994 Q4 (the data are plotting on the last observation in the moving average). Consistent with Figure 1, Figure 2 suggests that output growth trended down through the mid-1990s. In this case, however, output growth has not returned to its 6-year average trend rate.
Figures 1 & 2 may be deceptive. For example, Figure 3, which shows a 10-year moving average growth rate of output, suggests a very different picture; one that tends to support Cochrane’s suggestion that the economy could return to 3% to 4% growth. Specifically, it suggests that output grew at about a 4% rate from 1948 Q1 until about mid-1973, then declined 1 percentage point, permanently and rapidly, i.e., in about a decade. Long-term output growth then remained in a range of 3% to 3.5% until early 2006, when it declined rapidly with the onset of the 2007-2009 recession and has remained low since.
Rather than trending down as Figures 1 & 2 suggest, Figure 3 suggests that output growth fell quickly to a new long-run level. Hence, it might be reasonable to expect that output growth could once again return to something in the 3% or higher range. Of course, there are a number of unanswered questions: Is the decline after 2006 temporary or permanent? If it’s temporary, why has growth remained so slow more than seven years after the recession? If it’s permanent, what caused it? Furthermore, what caused the apparent permanent decline following the 1973-1975 recession?
But as I noted previously, moving averages can be deceiving. Figure 4 shows a 20-year moving average of output growth over the 1948-2016 period and a trend line based on the moving average data through 1994 Q4. The figure indicates that output growth declined steadily through 1994 Q4 from about 4.3% to about 3%. The figure indicates a steady decline rather than a precipitous drop. However, output growth begins to rise again inthe mid-1990s to a peak of 3.5% before declining rapidly in the wake of the 2007-2009 recession. 20-year average output growth remains significantly above the trend rate of 2% in 2016 Q2 (2% is also the trend growth rate for 2016 Q2 in Figure 1). Figure 4 suggests output growth declined steadily until the mid-1990s when something happened that pushed output growth significantly above trend and keep it there until the recession.
It is important to note that this conclusion is consistent with the behavior of output over that period. Figure 5 shows the real GDP and the Congressional Budget Office’s (CBO’s) estimate of potential along with aquadratic trend-line estimate of potential based on real GDP over the period 1949 Q1 through 1994 Q4 (I used this figure previously, Failure, Slow Growth, and Normalize).
As with Figure 4, Figure 5 suggests the economic growth was unusual from the mid-1990s to the onset of the 2007-2009 recession. Furthermore, as is the case with Figure 1, which indicates that output growth returned to trend when the recession ended, Figure 4 suggests that output returned to the trend when the recession ended.
What can account for the high level and growth rate of output during the 1995-2007 period? As I have argued, Failure, a reasonable explanation for the uncharacteristic rise in output in Figure 5 (and the uncharacteristic rise in output growth in Figure 4) is that it was caused by a confluence of several events: Advances in technology and technological innovations, the launching of the World Wide Web (1991), the widespread use of securitization and other financial innovations, the growth of finance generally, making home ownership a national priority, a marked decline in lending standards (especially for residential real estate lending), and lax oversight of the mortgage market and other financial markets. The last three were responsible for a massive over production of residential homes and supporting infrastructure.
This explanation is supported by the FACT that the period was characterized by two price bubbles: The dot com bubble drove the NASDAQ and other equity prices indexes to new heights only to burst in 2000 — the NASDAQ peaked in February 2000. The result was the 2001 recession, which was mild and short. The house price bubble drove home prices to unsustainable heights and peaked in late 2006. The result was the 2007-2009 recession and its aftermath; TARP, a massive increase in the federal debt, unconventional monetary policies (QE, forward guidance, maturity extension) and, in Europe and elsewhere, negative interest rate policies.
Did output growth fall quickly to a new permanent level after the 1973-75 recession as suggested by Figure 3, or trend down from 1948 through the mid-1990s as Figures 1, 2, and 4 suggest? It is difficult to know for sure, but the figures suggest and it seems more reasonable that it trended down rather than fell quickly. Indeed, it is difficult to think of a reasonable explanation for a rapid and permanent fall in output growth. The oil-price crisis is the most obvious candidate, but by the mid-1980s the real price of oil was significantly below its pre-oilprice-crisis level. Furthermore, firms make adjustments to such changes so it is hard to see how such a change could have a large permanent effect on output growth.
Also, as I noted in Normalize, output is determined by the quantities of capital and labor for a given level of technology. Capital is very difficult to measure, but we have reasonably good measures of labor. If the growth rate of output trended down, it is reasonable to expect that employment growth also trended down. Figure 6 shows the 20-year moving average of total non-farm payroll employment since 1948Q1.
The figure shows that employment growth has trended down steadily since the late 1960s and then fell swiftly after the 2001 recession (the 10-year moving average looks very similar). The marked decline in employment growth after the 2001 recession was, in part, due to an historically unprecedented decline in the labor force participation rate from 67.1% in February 2001 to 62.4% on September 2015, see Good Idea?. Hence, downward trend in output growth is generally consistent with the downward trend in employment growth. Conclusion: Output growth has been declining over time.
Why has output growth trended down? I don’t know. But I strongly suspect there is no single cause. Here are some possible causes. First, the economy has been transitioning from a manufacturing economy to a service economy during the entire post-war period; manufacturing currently accounts for only about 12% of GDP. While the declining trend in manufacturing does not necessarily mean a decline in output growth, it is not unreasonable to believe that this transition was accompanied by downward trend in productivity and, hence, economic growth.
Productivity growth is difficult to measure. But, one commonly used measure is the growth rate of output per hour, shown in Figure 7. Consistent with the previous analysis, the figure shows that, with the exception of the brief period from the mid-1990s to the mid-2000s (Alan Greenspan was one of the first to recognize that productivity had been increasing), productive growth trended down over the period and is currently at trend.
This analysis strongly suggests that output growth has been falling over time but increased significantly above trend during the period from the mid-1990s to the bursting of the house price bubble in late 2006. This period was characterized by unusual events and a dramatic, albeit temporary, rise in productivity. The analysis also suggests that both output and output growth have returned to something close to trend. Indeed, it appears that this essentially happened when the recession ended in June 2009.
There are a number of important implications of this analysis. First and foremost, the analysis implies that it is extremely unlikely that output growth will return to the 3% to 4% rate that Cochrane and many policymakers believe reflects the economy’s potential. The economy’s potential for long-run growth would seem to be something in the neighborhood of 2% to perhaps 2.5%.
Second, the analysis suggests that a higher level of long-run output growth can be achieved only if the factors that have caused output growth to fall persistently are reversed. But this can only happen if these factors are correctly identified. It is impossible to take the corrective steps needed to reverse the downward growth tend if the causal factors are unknown. Moreover, it seems likely that these factors are deeply structural, so significant structural changes will need to be made.
Third, the implication that output and output growth have returned close to trend when the recession ended means that the excessive actions taken by the Federal Open Market Committee (FOMC) in early 2009 and thereafter were unnecessary and, by implication, counterproductive. The analysis strongly reinforces my argument, Normalize, that the FOMC needs to move quickly to normalize its balance sheet, and my argument that the FOMC should never have engaged in QE, What It Should Have Done.
Moreover, slow growth is not a reason to continue the FOMC’s low interest rate policy as Bullard suggests. Rather, it is a reason to move as quickly as possible to return banks’ holdings of excess reserves to their pre-September 2008, i.e., to a level that will enable the FOMC’s to affect the funds rate without paying interest on excess reserves. Moreover, the FOMC needs to abandon its low interest rate policy, which is doing more damage than it is good, see Insanity and Unintended.