A DIFFICULTY IN MEASURING PRODUCTIVITY WITH A PERPETUAL INVENTORY CAPITAL STOCK MEASURE
| Date | 01 August 1983 |
| Published date | 01 August 1983 |
| Author | Edward M. Miller |
| DOI | http://doi.org/10.1111/j.1468-0084.1983.mp45003005.x |
A DIFFICULTY IN MEASURING PRODUCTIVITY
WITH A PERPETUAL INVENTORY CAPITAL STOCK
MEASURE
Edward M. Miller
Attempts have recently been made to measure year-to-year changes in
total factor productivity (Denison, 1974, 1979; Christensen and Jorgen-
son, 1973; Jorgenson, 1980; Gollop and Jorgenson, 1980; Christensen
et al., 1980; Kendrick, 1973; Kendrick and Grossman, 1980; American
Productivity Center, 1980).' These are argued to be a significant
improvement over the measures of labour productivity that have tradi-
tionally been used in that they incorporate the productivity of capital
as well as labour. However, the capital stock data used to compute
these measures are seriously flawed. The problem described here seems
to have been missed in the extensive literature on measuring capital
for productivity measurement. (For examples, see Creamer, 1972; Daly,
1972; Nadiri, 1970; Lave., 1966; National Research Council, 1979;
Ruggles, 1961; Usher, 1980; Diewert, 1980.) It will be shown that the
true capital stock and that used as an input into the productivity esti-
mates typically move in opposite directions if there is a change in the
rate of capital-embodied technical advance. A perpetual stock measure
of the capital stock logically requires that one knows how the produc-
tivity of capital has changed, which is precisely what one is seeking to
determine. Thus, there is a fatal circularity in the current methodology.
I. THE PERPETUAL INVENTORY METHOD
Most current capital stock measurements use a perpetual inventory
method. (See Young and Musgrave (1980) and Musgrave (1976) for an
account of the perpetual inventory method used by the US Bureau of
Economic Analysis in its capital series, and US Bureau of Labor Statis-
tics (1979) for an account of theirs.) The perpetual inventory method is
quite simple. A running total of the capital stock is kept. Each year, the
new investments are added to the stock, and the estimated retirements
and the appropriate depreciation are subtracted.
If there were actual data on retirements, the perpetual inventory
method for calculating the gross capital stock would be simple. Each
year's investment would be added to the capital stock, and the actual
retirements would be subtracted. However, there are virtually no data
See also the new Official US Government Series, US Bureau of Labor Statistics, 1983.
297
To continue reading
Request your trial