Re: costs of unproductive labor

Duncan K. Foley (
Fri, 6 Feb 1998 22:47:31 -0500

A comment on Juriaan's Costs of Unproductive Labor:

Juriaan, in discussing the treatment of unproductive labor, raises some

>The dispute concerning expenditure on unproductive labour concerns the
>question of whether this expenditure should be accounted for in the value
>of the product (c+v+s) as a component of new surplus-value created or as a
>component of constant capital used up (transferred value). For the purpose
>of calculating the true general profit rate, we need to estimate the stock
>of circulating capital, in its variable and constant components (and in
>constant components I suggested we include expenditure on unproductive
>labour). The difficulty is, although some information is available on
>inventories of raw materials etc., how do you estimate "stock levels" of
>the capital tied up in wages, what sorts of assumptions do you make, for
>example about production periods, stock rotations and turnovers.

Actually, the current business and national income accounting practice does
pretty well on these matters. Inventories are separated into three
"stages": Stage I inventories are "raw materials", that is, inputs to
production; Stage II inventories are "goods in process", on which some work
has been done; Stage III inventories are "finished goods awaiting sale".
Wages paid are accounted for in the value of Stage II inventories, so the
variable capital tied up is correctly accounted for. Inventory accounting
is on a "first-in first-out" basis, so you can actually recover the time
lag involved in the inventories by cumulating the flows of past capital
outlays. For example, if the inventory is 100 and the past periods flows
were 30, 20, 40, 20, 10, the time lag would be 3.5 periods, since 100 = 30
+ 20 + 40 + .5(20). Piruz Alemi has made these calculations for the U.S.
Manufacturing and Non-financial business sectors for 1947--1993 in his
thesis, and he and I have a paper reporting these results, which I will
post on the web when we solve some computer glitches in getting the graphs
to print along with the text.

We know
>for instance that generally speaking employers do not tie up (hold in
>reserve) the entire annual expenditure of wages, this expenditure is
>partially recouped from current income. We can also get an idea of
>turnovers (stock rotations) by dividing intermediate consumption by the
>average stock level during the year. Most writers avoid the problem
>altogether though and simply calculate profit rates as the ratio of surplus
>value to the value of fixed capital assets (given that normally the outlay
>on fixed capital dwarfs the outlay on wages or inventories).

Actually the situation with regard to fixed and circulating capital is more
complicated. The values of the stocks of inventories and fixed capital in
U.S. manufacturing, for example, are about the same magnitude. But the
flows of value corresponding to the stocks are very different, because
fixed capital turns over so much more slowly (on the order of 10 years)
than circulating capital (on the order of a few months). Thus circulating
capital flows are much larger than depreciation or gross investment.

Dumenil and Levy have an interesting discussion of the various bases for
computing profit rates in their book "The Economics of the Profit Rate".


Duncan K. Foley
Department of Economics
Barnard College
New York, NY 10027
fax: (212)-854-8947