Mike Mandel on productivity and Tyler Cowen on stagnation
Last week Mike Mandel published a very interesting chart as his nominee for the chart of the year.http://innovationandgrowth.wordpress.com/
He is using the net investment data to argue that capital spending is much smaller than generally assumed.
Standard thinking is that since the early 1980s capital spending has been booming and despite cyclical swings real business fixed investment has moved up to record levels as this chart of real business fixed investment as a share of GDP shows..
information tech ( IT) and all other, or more traditional capital goods it demonstrates that IT has accounted for virtually all the growth in real investment since 1980 and that it share of total capital spending has been steadily increasing — it now accounts for 45% of the total.
But IT equipment has a very different life span than more traditional capital equipment. Business computers have a life span of only about 2 years and communication equipment has a similar life span. In contrast, traditional capital equipment has a much longer life span. An office building, a ware house or something like a blast furnace can be expected to have a useful life of decades.
Even more rapidly depreciating equipment like trucks now last some 10 to 20 years.
If you are depreciating your equipment over only a couple of years it means that you have to run ever faster and faster to keep even. For example if in year one you buy one computer and in year two you buy two computers and year three it is three computers, etc,.etc., you net addition to your capital stock is much less than your gross purchases of computers. In year three, one of the three computers you buy goes to replace the computer you bought in the first year and in year five, three of the five new computers only replace the three you bought in year three, etc., etc… So you net purchases is three, not five. This is the primary reason Mandel is making a very important point with his analysis of net capital spending.
For some time Mandel has been arguing strongly that the official productivity data significantly over states productivity growth. Economist debate what drives productivity growth, but no one denies that growth in the capital stock is an important way new technology becomes embedded and is an important determinate of productivity growth and real standards of living. Mandel is correct that the growth in the net private capital stock has slowed sharply since the early 1980s.
Moreover, if you look at the growth in real net capital stock per employee it shows a clear break in trend in the late 1970s. From 1945 to 1985 the trend growth rate for this series was 1.6%. But since 1974 the trend growth rate has fallen to about 1.0%, a much slower growth rate. This appears to provide solid support to Mandel’s argument that productivity growth is slower and Tyler Cowen’s great stagnation thesis.
I suspect that these two economists are on to something, but I am not quite ready to fully accept their arguments. I would like to see some other analysis independently support their arguments.
I always look for independent conformation of economic trends. One reason I question their argument is that a dominant determinate of corporate profits is the spread between unit labor cost and business prices. If productivity growth were significantly slower, the tight relationship between profits growth and the growth of the spread between unit labor cost and prices should be breaking down. But as the chart below shows, if anything, the relationship appears tighter in recent years than it was in an earlier era. So while I am not quite ready to fully accept either Mandel’s weaker productivity thesis or Tyler Cowen’s great stagnation thesis, but I am not ready to reject them either.
There is a bit of math which shows that any increase in investment as a share of GDP leads initially to more rapid GDP growth (all else equal), but that over time, the pace of growth slows back to the prior trend. The reason for the pattern is that depreciation begins to take up an increased share of the increase in investment, leaving less for an increase in the capital stock, up to the point that you hit a stable turn-over in capital.
What does that have to do with the issue spencer has raised? More rapid depreciation of IT equipment means more rapid deterioration in output growth. The math of eroding growth – shifting capital spending increasingly toward replacement – works faster with faster depreciation. Now, if there is no increase in the share of GDP going to capital, but there is a shift in capital spending toward capital goods that depreciate more rapidly, then instead of gorwth coming back to trend after a one time lift, the growth trend should slow.
Just as a footnote, Larry Lindsey made the point about the one-time nature of a lift to growth from an increased share of GD going to investment. His point was that a cut in the marginal tax rate for investment income would induce a one-time lift to growth, not a change in the long-term trajectory. This from one of the early (and honest) supply-side champiions.
Thought you might find interesting: a longer-term view, with net, gross and cap consumption broken out:
http://www.asymptosis.com/wp-content/uploads/2011/12/image007-480×328.jpg
While gross business fixed-capital investment grew rapidly from the 30s to the 80s, the rate of capital consumption grew even faster. Result: Net business fixed-capital investment/GDP in the 00s was 40% below the 70s.
And yes: business investment has skewed hugely toward more-rapidly-consumed hardware and software, and away from structures, ever since the 30s — but especially since the 80s.
http://www.asymptosis.com/wp-content/uploads/2012/01/hwswvsstructures.png
Who needs capital stock when you are farming everything out and collecting the royalties?
Isn’t net investment just a bull shit concept?
Investment is real money layed out for capital goods and physical plant.
Depreciation is an arbitrary accounting device to amortize investment from the past.
If businesses were allowed to treat capital expenses as accounting expenses in year 0, the depreciation concept would not even exist.
Just because they are both denomonated in local curency doesn’t mean that doing simple math on them actually means anything.
I’m perfectly willing to be wrong on this if somebody can weave a convincing narrative that subracting arbitrary fractions of past invest from the reality of current investment is a) meaningful in any sense, b) specifically meaningful in explaining business climate and decisions, and c) helps us learn anything about the economy in general.
Cheers!
JzB
Jazz, I’m with you babe. I guess it’s kind of useful to have some handle on the existing stock of productive real assets (and an estimate of depreciation is necessary to achieve that), but even that’s a real pipe dream. They can tally up hardware, software, and structures with some estimation accuracy, but what about skills, ideas, knowledge, “organizational capital,” and a zillion other intangible real assets that swamp the quantity of fixed assets out there?