Productivity growth, the most mysterious of economic statistics, was announced Thursday for the first quarter of 2009 — revised upwards from 0.8% to 1.6%. After a quarter century of stellar growth from 1948 to 1973 productivity growth suddenly collapsed, and remained low for the next decade. Then after 1982 it recovered somewhat, accelerating further slightly in the middle 1990s, although still not to its 1948-73 level. In 1997, Federal Reserve Chairman Alan Greenspan famously used a mysterious acceleration in productivity (later almost, but not quite wiped away by statistical revisions) to justify growing money supply considerably faster than would previously have been thought prudent. It’s thus of great interest to see what might have happened to productivity growth following last September’s collapse – the market took Thursday’s figures as a positive signal. There is however reason to believe it was a false one.
It cannot be overemphasized that the productivity performance of the US economy since 1990 has been solid but not stellar. From 1990 to 2008, according to the Conference Board’s Total Economy Database, labor productivity has increased by 1.80% (2.05% in the bubble years since 1995). That compares with 1.27% in the sluggish 1973-90 period (1.00% in 1973-82) and 2.57% in the halcyon years of 1950-73. However even slumping and despised Japan has done better since its bubble burst, with 1990-2008 productivity increasing by 1.94% per annum. Moreover Britain has done much better, at 2.36% per annum. On the other hand France and Germany, at just over 1.5% annually, have done worse, as has Italy (0.93%) and Canada (1.36%). US productivity has since the 1990s put on an adequate performance, in other words, but no “miracle.”
Examining multifactor productivity statistics gives a better feel of the causes of productivity growth. Multifactor productivity had quite a good year in 2008, growing 1.1%, slightly above the average of the 1987-2008 period. However even more notable is the contribution of capital intensity to labor productivity growth; whereas over the 1987-2008 period it contributed 0.9% annually, in 2008 it contributed fully 1.6% to reported labor productivity growth of 2.8%. Capital inputs to the US economy have almost doubled as a share of GDP since 1990. The productivity of capital usage has been in steady decline since the 1960s, and has declined by a further 11% since monetary policy was changed to permanent expansion at the beginning of 1995.
The last 15 years therefore have caused a modest increase in US labor productivity growth, albeit nowhere near the level of 1948-73, which has been caused by the surge in money creation and a substantial increase in the US economy’s capital intensity. In the late 1990s, most of the new capital came in the form of Internet services and fiber-optic cable capacity, much of which has served us increasingly well in the years since 2000; since the turn of the millennium most of the increase has come in the form of housing, of much more questionable long term utility. Nevertheless, even in the last six months the exceptionally low interest rates have caused a further surge in the capital intensity of the US economy, which has had a further effect in increasing labor productivity (albeit mostly in the unpleasant form of spiraling unemployment, since output has declined.)
Going forward, this trend will have to reverse. The cost of capital is clearly in the process of increasing sharply, both through higher interest rates and through the lower ebullience of equity and housing markets. This means that the output from capital usage will also have to increase – the capital intensity of the US economy can no longer increase, and must start to decline. From a purely arithmetical viewpoint, therefore, it is most unlikely that labor productivity can maintain the relatively satisfactory growth rates of the last decades, and much more likely that it will suffer a reverse similar to that of 1973-82, or possibly worse.
Tackling the problem now from the other end, there are a number of recent changes which make me think that productivity growth will be much lower going forward. For one thing, much of the growth of the last generation was in financial services, which at their 2006 peak represented 40% of the earnings of the Standard and Poor’s 500 share index. By definition, if the financial services business is capable of paying all those billions in bonuses to its practitioners, it must be very productive indeed, in order to generate the necessary revenues.
Of course, we now know that much of the income on which the financial services boom was based was illusory, that many of the services the industry provided were detrimental to the nation’s wellbeing and should not have been counted in output, and that in reality we will be considerably better off with the industry downsized and made to respect its betters. Still that’s a lot of high-productivity jobs gone, a lot of high-productivity (if spurious) output made to disappear. The downsizing of financial services alone is likely to have a significant dampening effect on productivity growth going forward.
Another high productivity business that will be much smaller at least for several years to come is housing and construction in general. Modern construction employs huge amounts of capital as well as labor, so heavy construction activity tends to increase labor productivity, even though construction workers themselves have only modest value added. With an overhang of properties in most residential markets, and an overhang of office space in many city centers, it seems likely to be several years before construction gets fully back on its feet, and both productivity and the capital intensity of the economy will be lower while the industry struggles.
Conversely, the principal gainer from the turmoil of the last couple of years is undoubtedly government. One does not need to be an ardent believer in the superiority of the private sector (as I am) to note that the output of government services is generally entered in the national accounts only at their cost, without regard as to whether their value may exceed that cost. Thus government’s contribution is generally low, since output is increased only by the cost of its inputs. Furthermore, there is very little scope to improve government’s productivity, even if that productivity improvement could be measured other than through decreased costs, which would depress reported output.
An increase of 5-8% in the proportion of the economy taken up by government, as seems likely, besides the depressing effect that it will have on the rest of the economy, will thus depress productivity growth directly. I have suggested in the past that economic analysts would do well to examine “Gross Private Product” — the proportion of output produced by the private sector and therefore valued by market forces. In the coming years, with state healthcare scheme and a “cap and trade” market-distorting carbon emission permits scheme both in view, it seems overwhelmingly likely that GPP will show weak if any growth.
Finally, there are the direct effects of recent changes in the US economy, which appear likely to involve higher real interest rates than we have been accustomed to in the last 15 years and will certainly involve a lower degree of leverage, in the household sector, the financial sector and the economy as a whole.
Lower leverage increases the amount of capital necessary to support a given volume of activity. It thus decreases reported labor productivity either by increasing capital inputs or depressing output.
Higher real interest rates, which seem likely in order to ward off inflation, have a more mixed effect on productivity. By lowering output in relation to labor inputs, they depress it. On the other hand, by increasing the rewards to capital, they ensure it is allocated more effectively, thereby increasing productivity in the long run. The empty McMansions produced by the housing boom, for example, may have looked very good in 2005’s productivity statistics as they were being constructed but are depressing productivity now since they add only costs and no rental or other revenues, while consuming capital inputs through their depreciation. Long-term productivity growth requires the minimization of unproductive investments made during bubble periods of one kind or another.
Thus the qualitative consideration of recent developments in various sectors tallies with the quantitative argument from examination of past productivity developments. Take both arguments together, and you can see why reported growth in productivity is likely to be far lower in the next few years than in the recent past. Like the downtrend of 1973-82, the new downtrend will be difficult to explain and may be prolonged. The likelihood of a prolonged period of sluggish productivity growth is thus yet another reason to expect the 2010s to be economically a miserable period.
-0-
(The Bear’s Lair is a weekly column that is intended to appear each Monday, an appropriately gloomy day of the week. Its rationale is that the proportion of “sell” recommendations put out by Wall Street houses remains far below that of “buy” recommendations. Accordingly, investors have an excess of positive information and very little negative information. The column thus takes the ursine view of life and the market, in the hope that it may be usefully different from what investors see elsewhere.)