clarity

[Doug Henwood]

Speaking of clarity, Brad, how about this passage from J. Bradford De Long
and Lawrence H. Summers, "Equipment Investment and Economic Growth: How
Strong Is the Nexus," Brookings Papers on Economic Activity 2 (1992), pp.
157-211? Though I don't have "training" in economics beyond the undergrad
level, after years of reading this sort of thing the passage seems quite
clear to me - and no doubt to you, as its lead author. But how many folks
unexposed to the literature would find it a model of clarity?

<quote>
Estimating Total Factor Productivity

Neoclassical growth theory can be viewed as either an organizing framework
for thinking about growth or as a substantive theory. To the extent that it
is a substantive theory, one of its most basic predictions must be that TFP
growth is not associated with the principal dynamic variables-investment,
depreciation, and population growth rates- about which neoclassical growth
theory makes predictions. In this subsection, we test and reject the null
hypothesis that TFP growth is uncorrelated with equipment investment.

It should come as no surprise that the very strong association of output
per worker growth and equipment investment documented above is, in large
part, also a strong association between equipment investment and total
factor productivity growth. Given the limitations of our database, the
calculation of total factor productivity estimates is not straightforward.
We require estimates of the average share accruing to factors of
production, and estimates not of gross, but of net investment rates. Thus
total factor productivity estimates require estimates of initial capital
stocks. Because such initial capital stock estimates are crude, they
introduce a potential source of noise into TFP growth calculations.

We have estimated 1960-85 TFP growth rates for 31 of the economies in our
high-productivity sample. For these 31 economies, we have year-by-year
estimates of nominal investment in different types of assets and of price
structures in the 1950s. Along with an assumption about pre1950 investment,
we can construct 1960 estimates of capital stocks that can then be used to
calculate total factor productivity growth from 1960 to 1985. The
restriction of our total factor productivity growth estimates to 31
high-productivity economies limits us to a sample that does not show the
growth-equipment nexus as strongly as some of our other samples. For
equations such as those in table 1, the equipment investment coefficient
over the 1960-85 period is 0.198 for this particular sample, toward the low
end of the range found in our later regressions.
</quote>

Doug