[OPE-L] models with unequal turnover periods

From: Jurriaan Bendien (adsl675281@TISCALI.NL)
Date: Sun Sep 09 2007 - 10:40:25 EDT

I studied this a bit empirically 20 years ago using New Zealand
manufacturing data, economic census data, national accounts data and
unpublished data on inventory valuations. More recently I also looked at the
NIPAs.  A few observations:

In rich countries such as the US, nowadays about half of all intermediate
consumption (flow of expenditure on inputs used up in production, i.e. Cc)
consist not of goods purchased/rented, but services purchased/rented.
Because this is so, there is in fact no fixed relationship between Cf and
Cc. The proportions of services purchased versus services produced in-house
can fluctuate, and I/O tables show that over time the amount of services
purchased can vary significantly across time, in different sectors.

If you take the goods-producing industries, you can strike a ratio between
the average inventory level during the year, and the total intermediate
expenditure during the year, to obtain a rough indication of the number of
stock rotations (inventory cycles). You can do this for about 10 or 20
years, and you will see that inventory levels and the inventory
turnover-times can change quite a bit even from year to year. The historical
trend is one of increasing throughput.

You can also establish that for different branches of industry, Cc/Cf and
the Cc turnover speeds are quite different.

In times of of high inflation (e.g. first half of the 1970s) the inventory
valuation adjustment can be very great, i.e. more than 1% of GDP.

In national accounts, intermediate consumption includes only operating
expenses which are considered to be directly related to production. Thus,
total operating expenses of producing entities can be larger than included
in official intermediate consumption, because they include expenses
considered to be not directly related to production. They are used up in a
"financial" sense, but not in the official "production" sense.

If businesses decide for economic reasons to rent more physical assets, or
alternatively buy more physical assets, this can independently affect the
size of GDP components and the size of intermediate consumption. If they
buy, this boosts GDP; if they rent or lease, this lowers GDP.

The true value of fixed capital assets is empirically very difficult to
tell, because of lack of uniformity and reliability in valuation methods
(historic cost, current replacement value, current sale value, book value,
scrap value etc.) and because of the divergence between economic
depreciation, actual depreciation charged, and tax-assessed depreciation.
Additionally, national accounts often include costs in the depreciation
aggregate which are not strictly related to depreciation itself, but to the
maintenance, installation or insurance of fixed assets. Consequently,
inferences made about the relationship between depreciation and other
aggregates should keep this in mind; the economic depreciation rate diverges
from actual write-offs.

Empirically, I never found any fixed, permanent relationship between Cc and
Cf, and if there seemed to be a fairly stable historic proportionality in
given cases for some time, that could be attributable to an artifact of
statistical estimation, i.e. a result of interpolations, extrapolations and
imputations (among other things, estimations are "smoothed" by projections
from the past; plus, statisticians typically distrust the reality of
unusual, large fluctuations in the data).

Modern corporations these days will often buy assets, including fixed
assets, without using them for production - these are more speculative
investments, resulting in property income. They may be treated and
depreciated as production capital, but in reality do not enter into


This archive was generated by hypermail 2.1.5 : Sun Sep 30 2007 - 00:00:05 EDT