Nice missive by Steven Roach, it's a bit early to declare that the US economy can weather any downturn:
Global: The Metrics of Truth
Stephen Roach (New York) 2-14-2001
Alan Greenspan has upped the ante. In his bi-annual policy report to the US Congress, he went further than he ever has in extolling the virtues of America’s productivity-led economy. He left little doubt of
the critical role that rapid productivity growth now plays in the formulation of monetary and fiscal policies. But he also suggested that the productivity resilience of the past six months was the smoking gun
of the New Economy. In his words, it was the "true test" of a Teflon-like US economy that could withstand a sharp cyclical blow without toppling into recession. How valid is this profound assertion?
Like biblical miracles, productivity breakthroughs do not lend themselves neatly to empirical validation. The productivity experts in the economics profession have long cautioned against making too much
out of productivity acceleration in the upleg of a business cycle expansion. It is uncertain as to how much of that pickup is traceable to temporary forces and how much stems from more enduring efficiency
improvements. The answer can only be known with clarity in the downleg of the business cycle, after a weaker economy unmasks the cyclical portion of the productivity trend. Since the 2.8% average
productivity gains of the past five years may have been simply nothing more than a cyclical "catch-up" from the anemic 1.5% trend of the preceding five years -- to say nothing of the equally weak
performance of the 1970s and 1980s -- this point is hardly idle conjecture. Only a downturn can provide the long-awaited metric of truth.
Yet Greenspan was emphatic in concluding that this "true test" has now occurred. With all due respect to the Fed Chairman, I take great issue with this assertion. The final six months of 2000 hardly
qualifies as a downturn in the real economy. Looking through the volatility of quarterly gyrations in economic activity -- something that any productivity expert would do -- real GDP growth "slowed" to
3.5% (y-o-y) in the four quarters ending 4Q00. That’s a relatively minor shortfall from the 4.3% trend of the past five years -- far from a sharp cyclical weakening in the economy that would provide an
ironclad test of the productivity acceleration thesis. To say that productivity held up well under current circumstances is stretching the point. It turns out that productivity growth did, in fact, downshift by
1.9-percentage points in the final six months of 2000, slowing from peak annualized gains of 5.3% in 2Q00 to 3.4% in 4Q00.
As long as I’m being picky, I must confess to being highly suspicious of the preliminary productivity results that were just published for the final period of 2000 -- a 2.4% annualized rate of increase relative
to the third period. Productivity, of course is output per unit of labor input, and it’s the latter component of the ratio -- labor input -- that appears especially dubious to me. Curiously, hours worked in
nonfarm business establishments -- the broadest proxy for labor input -- were actually unchanged in 4Q00. Yet the hours worked embedded in the productivity estimates showed a 1.1% annualized rate of
decline -- a contraction in labor input that accounted for fully 45% of the estimated gain in nonfarm business productivity in the final period of 2000.
Forgive me if I get a bit technical at this point. There are several factors that can account for this discrepancy -- the most notable being shifts in the hours worked by self-employed individuals. These
workers, who account for about 6% of the total US workforce, do not appear on the payrolls of the nonfarm business establishments that drive the bulk of the economy’s productivity. Consequently, a full
accounting of economy-wide productivity should probably include these individuals, even though they remain largely outside the sphere of mainstream business activity. It turns out that the number of
self-employed workers fell at a 5% annual rate in 4Q00 -- a drop that seems to justify the government’s surprisingly cautious assessment of overall hours worked in the broader economy. But it also turns
out that this same self-employed headcount has fallen quite sharply in the fourth quarter in five of the past seven years. In other words, the purported plunge in self-employed hours worked could be more
of a recurring statistical anomaly rather than a by-product of newfound business efficiencies.
Rest assured, the real test is yet to come. In that vein, the productivity implications of the current quarter are shaping up to be especially interesting. Hours worked rebounded sharply in January to a level
that stood 1.6% (annualized) above the 4Q00 average. Under the presumption that the economy seems to be stabilizing, January’s gain in labor input could well hold for the quarter as a whole -- an
increase that stands in sharp contrast to our estimate of a -0.6% decline in real GDP. Putting the two trends together -- falling output in conjunction with rebounding labor input -- and productivity could
well be declining in the current quarter. That, of course, would mirror the pattern that has occurred in all recessions of the past -- drawing into sharp question the Greenspan claim that the New Economy
has passed the "true test" of productivity resilience.
Alas, there is no absolute truth in the murky realm of productivity analysis. Never really has been. And the integrity of the productivity calculus is far worse in the services-based New Economy than it ever
was in the goods-producing Old Economy. Output measures are especially elusive in the services sector. Moreover, output is seriously distorted by the fantasy of quality-adjusted ("hedonic") price indexes
for new information technologies; back-of-the-envelope calculations show that an implied 6% annualized IT deflation has boosted aggregate output -- and productivity -- growth by as much as 0.5%
annually since 1993. Then there’s the numerator, where unmeasured work time -- a natural by-product of the IT-driven portability of information workers -- has probably biased productivity growth to the
upside for many years. In short, the sad fact is that America’s productivity metrics are woefully deficient by any standards of statistical integrity.
When I think of productivity miracles, my mind always drifts back to the 1960s. That, of course, was the mother of all productivity booms in the United States -- average gains of 3.6% over the 1961-68
interval. Yet when that cycle started to fade, there was talk of the cushion of productivity resilience, as well. The numbers seemed to bear that out: Real GDP growth slowed from its peak-of-cycle growth
rate of 8.5% (y-o-y) in 1Q66 to 3.5% two years later; yet productivity was still cruising at 3.4% in early 1968. In the end, of course, this resilience turned out to be short lived. America was on the brink of
nearly two decades of its worst productivity performance in the modern era.
Alan Greenspan has correctly singled out productivity growth as the single most important feature of any economy. It’s hard to argue that point. But it’s another thing altogether to leap from theory to
verifiable truth. The productivity resilience of today’s US economy has yet to be tested. Moreover, if you want to believe the numbers -- and I don’t -- the next tick on the productivity clock could well be
a disappointing one at that. It is reckless to hang the future of the US economy on dubious productivity statistics. The same can be said of the long-term growth estimates that are now shaping the great
budget debate. The jury remains out on the true test -- if it ever really occurs. |
