===== PAGE 1 =====

THE JOURNAL OF
POLITICAL ECONOMY

Volume 43 OCTOBER 1935 Number 5

PROFESSOR KNIGHT AND THE
“PERIOD OF PRODUCTION”

FRITZ MACHLUP
University of Buffalo

HE past few years have been prolific of contributions about
“capital and the period of production.” Any increase in
the output of such contributions certainly takes place un-
der increasing cost (because it is necessary for the writer to study
all his predecessors’ products), and, probably, under decreasing
return. Nevertheless, there seems to be good reason for submit-

! Frank H. Knight, “Capitalistic Production, Time, and the Rate of Return,”
Economic Essays in Honour of Gustav Cassel (London, 1933); “Capital, Time, and
the Interest Rate,” Economica, August, 1934; “Professor Hayek and the Theory of
Investment,” Economic Journal, March, 1935; F. A. von Hayek, “Capital and In-
dustrial Fluctuations,” Econometrica, April, 1934; “On the Relationship between
Investment and Output,” Economic Journal, June, 1934; Martin Hill, “The Period
of Production and Industrial Fluctuations,” ibid., December, 1933; C. H. P. Gifford,
“The Concept of the Length of the Period of Production,” ibid.; “The Period of
Production under Continuous Input and Point Output in an Unprogressive Com-
munity,” Econometrica, April, 1935; J. Marschak, “A Note on the Period of Pro-
duction,” Economic Journal, March, 1934; K. E. Boulding, “The Application of the
Pure Theory of Population Change to the Theory of Capital,” Quarterly Journal of
Economics, August, 1934; Wassily Leontief, “Interest on Capital and Distribution:
A Problem in the Theory of Marginal Productivity,” Quarterly Journal of Economics,
November, 1934; R. I. Fowler, The Depreciation of Capital, Analytically Considered
(London, 1934); Erich Schiff, Kapitalbildung und Kapitalaufzehrung im Konjunk-
turverlauf (Vienna, 1933); Richard von Strigl, “Lohnfonds und Geldkapital,” Zeit-
schrift fiir Nationalokonomie, January, 1934; Kapital und Produktion (Vienna, 1934);
Walter Eucken, Kapitaltheoretische Untersuchungen (Jena, 1934); Karl H. Stephans,

577


===== PAGE 2 =====

578 FRITZ MACHLUP

ting another unit of product. In a series of ingenious articles?
Professor Knight has proposed to discard as worthless some tools
of economic analysis which I consider indispensable for successful
handling of certain problems. The tools he wants to scrap are
those concepts which have been used in capital theory under dif-
ferent names, such as “length of the production process,” “‘round-
aboutness of production,” “period of waiting,” “period of produc-
tion,” “period of investment,” “time structure of production,”
“period of maturing,” and similar terms. Professor Knight's opin-
ion has great weight with every serious student of economics; if
the disputed concepts are to be rehabilitated, Professor Knight's
criticism must be answered. This will be attempted here, partly
with explicit reference to the contentions of Professor Knight,
partly by way of general discussion of the concepts and the prob-
lems involved. An effort will be made, upon this occasion, to
clear up a number of obscurities and ambiguities connected with
the theory of capital “promulgated by Bohm-Bawerk and his fol-
lowers and generally accepted and taught in the past genera-
tion.”3
I. IS CAPITAL PERPETUAL?

1. Stationary, growing, or retrograde economy.—The usefulness,
or even necessity, for economic theory of the conception of a “sta-
tionary economy” cannot be seriously questioned. This concep-

“Zur neueren Kapitaltheorie,” Weltwirtschaftliches Archiv, January, 1935. The fol-
lowing articles come to my attention too late to be taken account of in the present
analysis: Howard S. Ellis, “Die Bedeutung der Produktionsperiode fiir die Krisen-
theorie,” Zeitschrift fiir Nationalékonomie, June, 1935; M. Joseph and K. Bode,
“Bemerkungen zur Kapital- und Zinstheorie,” Zeitschrift fiir Nationalokonomie,
June, 1935; Oskar Morgenstern, ‘Zur Theorie der Produktionsperiode,” Zeitschrift
fiir Nationalokonomie, June, 1935; Richard von Strigl, “Zeit und Produktion,”
Zeitschrift fiir Nationalékonomie, June, 1935; Ragnar Nurkse, “The Schematic Rep-
resentation of the Structure of Production,” Review of Economic Studies, June, 1935.
Besides the published material, manuscripts by FF. A. von Hayek and an unpublished
discussion in correspondence between G. von Haberler, F. A. von Hayek, J. Mar-
schak, L. von Mises, and myself were drawn upon in writing this article.

2 Besides the three articles by Professor Knight listed in the last footnote, see also
his paper, “The Ricardian Theory of Production and Distribution,” in the
Canadian Journal of Economics and Political Science, February, 1935.

3 Knight, Economic Journal, XLV (March, 1933), 79.


===== PAGE 3 =====

THE “PERIOD OF PRODUCTION” 579

tion is inseparably bound up with the application of the method
of variation in economics. One cannot arrive at any conclusion
about the effects of any change if the “other” (independent)
things (or conditions) are not set “invariant” or assumed to be
unchanged. Thus, in a discussion of the effects of a certain “shift
in demand,” or of a changed supply of money, or of better crops,
or of any other “substantive” change, one conveniently starts out
with the assumption that the supply of “capital” does not change
independently at the same time. One of the greatest difficulties in
our theory, the problem of “maintaining capital intact,” is, of
course, evaded by assuming a stationary economy, but this can be
considered permissible for certain problems. There is, however,
one problem in the discussion of which the simple assumption of
the stationary state is not permissible, namely, the problem of the
capital stock or capital structure itself. To assume perfect main-
tenance of capital where one should discuss the conditions of such
maintenance is a case of begging the question, or, at least, of
avoiding the problem which is to be solved.

A statement such as “all capital is inherently perpetual’4 is of
doubtful value. If it is limited to the case of an economy which is
assumed to maintain, or to add to, its capital, it is but a repetition
of the assumption. Ifit is stated as a fact, it is wrong. If we define
a stationary economy as one which (perhaps among other things)
maintains its capital, and if we call capital “perpetual” when it is
(at least) maintained, then the assertion that for a stationary
economy all capital is inherently perpetual obviously means no
more than that an economy which maintains its capital maintains
its capital. Much more important is the question whether econ-
omies always maintain their capital, and under what conditions
they do so.

Professor Knight thinks he is warranted in asserting ‘“‘the fact
of perpetuity.”s He adds, however: “Unless society as a whole
becomes decadent with respect to its stock of capital, no increment
of capital viewed as a quantity ever is disinvested.”® This ‘“un-

4 Knight, Economica, August, 1934, p. 204.

$ Canadian Journal, February, 1933, p. 15 (italics are his).

6 Ibid., p. 15.


===== PAGE 4 =====

580 FRITZ MACHLUP

less” qualification reduces the “fact” again to tautology. (Unless
society acts differently it acts in the asserted way.) In other
places the “fact of perpetuity” is stated in a less safe—because
less tautological—formulation.” There was and is always the
choice between maintaining, increasing, or consuming capital.
And past and “present” experience tells us that the decision in
favor of consumption of capital is far from being impossible or
improbable.® Capital is not necessarily perpetual.

2. Individual disinvestment, social disinvestment.—In a ‘“‘capi-
talistic society,” in which property is owned privately, the process
of capital disinvestment is the result of action or inaction of indi-
viduals, under the influence of their personal preferences or of
certain institutional forces or of state interference.’ It must be
clear that at any moment of time the alternatives are open to
change in either direction, or to leave unchanged the relation be-
tween the current stream of consumption income and the future
stream of consumption income, the latter partly being represented
in the existing capital (which is the “capitalized” value of future
income). The conditions under which such alternatives can be
realized are dependent on the given production structure of the
whole community and on the simultaneous choices on the part of
the other individuals within the community. Dissaving and dis-
investing by one individual can be offset by saving and investing
by another individual, in which case the individual disinvest-
ment does not become social disinvestment. That “in a station-
ary or growing society disinvestment by an individual owner in
no wise involves actual reconversion of ‘capital’ into income’ 1s
again nothing but a tautology. It is just the relation of the total
amounts of capital individually disinvested to the total amounts

7 “Capital is perpetual in so far as economic principles obtain and economic
reasoning is applicable” (Economica, August, 1934, p. 277); or, “in a society which
is not planning for the end of all things, all property income is perpetual” (ibid.,
p. 268). Neither of the two contentions seems to be correct. Economic reasoning

can be perfectly well applied to an economy which consumes parts of its capital.
Nor is it necessary that such a society plan for “the end of the world.”

8 For an illustration of a case of capital consumption see my article, “The Con-
sumption of Capital in Austria,” Review of Economic Statistics, January, 1935.

9 Ibid. 10 Knight, Economica, August, 1934, p. 273.


===== PAGE 5 =====

THE “PERIOD OF PRODUCTION” 581

of capital individually invested at any moment of time which
makes society stationary, growing, or retrograde.

The problem of social “net” investment or “net” disinvest-
ment 1s important, and its discussion calls for consideration of
certain “time relations.” Since investment is a conversion of
present income into future income (i.e., capital) by abstaining
“for a certain interval” from a “strip” of the present consumption
stream; and since disinvestment is a ‘“‘reconversion of ‘capital,’
i.e., of [future] income, into income for a short period at a corre-
spondingly higher time rate,” it is an economic problem of major
importance to find out about the time distance between those
present and future incomes, about the mentioned “interval,” and
about the mentioned “period.” These “periods” are important
for the main problems of capital theory which is the theory of
actual or potential changes in the quantity relationship between
present and future services—or, in other words, the theory of dis-
tribution of services over time.

3. Are production and consumption simultaneous?—The as-
sertion that all capital (in terms of value of anticipated future in-
come) is always maintained and, therefore, “perpetual” was made
to support the view that the notion of a period between produc-
tion and consumption is meaningless. The assertion to the con-
trary that capital need not, and, in fact, is not, always maintained
ought to support the view that the notion of the mentioned period
has meaning and significance. Professor Knight is prepared to ad-
mit that “the notion of a lapse of time between production and
consumption has practical meaning where society has to meet
unanticipated changes in conditions.” Apart from the theoretical
and practical importance of unanticipated changes in conditions
—it is not far from correct to call reality a continuous series of
unanticipated changes—it is clear to me and many others that the
“potentiality” of change is always present and that the choice be-
tween alternatives (only one of which is to abstain from changing,
and hence to repeat, former choices) is a category in economics.

For the picture of a stationary state, from which change is ab-

1 Ibid. In the original, “perpetual” income stands for “future” income.
2 [bid., p. 276.


===== PAGE 6 =====

582 FRITZ MACHLUP

sent, simultaneity of production and consumption can be a useful
fiction. With the aid of a “cross-section” view of production and
with the element of time abstracted, certain relations in the eco-
nomic system become more readily perceptible. Hence, proposi-
tions such as “production and consumption are simultaneous,’
or “the production period for consumed services . . . . is zero,"
are useful and admissible (fictitious) assumptions for the discus-
sion of certain problems of the stationary economy. But the “as
if’ character of such heuristic suppositions must never be forgot-
ten. And for discussion of problems other than those of an ab-
solutely stationary state, especially of problems of capital, of in-
vestment,™s of changes in demand, and similar ones, the assump-
tion of ‘“simultaneity of production and consumption’ is mis-
leading and inadmissible and has to be dropped. Only for prob-
lems of the first type may maintenance of capital, including re-
placement of particular items of “plant,” be conceived as a part
of the production of the output consumed at the same time.”® But
any theory of economic change and any theory of capital has to
regard the time element as its integrant part. The production of
a definite quantity of output can be done with or without full
maintenance of the instruments necessary for its production.
Therefore, the output of consumable services is not dependent
upon the simultaneous input of productive services used for
maintenance or replacement of plant; the productive services
used for maintenance or replacement of plant are not a part of
the production of services consumable at the same time, but at
later moments of time; there is a time interval between the input
of services and the “dependent” output of services,” and this

13 Ibid., p. 275.

14 Knight, Economic Journal, March, 193s, p. 88.

15 Note the following concession of Professor Knight's: “In the act of growth,
‘waiting’ can be said to be involved, both during the interval of construction itself
and at least for such time thereafter as is required for the new capital to yield a total
of consumption equal to that which was sacrificed in creating it” (Economic Essays
in Honour of Gustav Cassel, p. 339).

16 Knight, Economic Journal, March, 1935, pp. 84-85.

17 It is absolutely incomprehensible to me that Professor Knight, as one of the
rare economists who understands, holds, and teaches “the only theory which make:


===== PAGE 7 =====

THE “PERIOD OF PRODUCTION” 583

interval may conveniently be called “period of production” or
“period of investment.”

4. Capitalization of perpetual or of time-limited income.— That
the cost of production of any “capital good” is equal to its capital-
ized future return is a well-established condition of equilibrium.
How the rate of interest enters into both sides of the equation was
shown by Professor Knight with utmost clarity.” On the cost-of-
production side of the equation interest is calculated (as “carrying
charges”) on the services successively invested during the “period
of construction.” On the capitalized-income side of the equation
interest is calculated on the series of future yields to reduce it to
the “present value.” There are two methods of capitalization.
The one simply capitalizes the series of gross returns expected
during the service life of the capital good; the other method re-
duces first the time-limited series of gross returns by way of de-
ducting all maintenance and replacement costs to a perpetual
series of net returns and then capitalizes this perpetual series.
However, “it happens to be somewhat simpler algebraically not
to make this conversion formally, but to express the present
worth of the time-limited income stream directly.”

In practical life capitalization of a limited series of returns is
more usual than that of an infinite series, not only on account of
mathematical simplicity but on account of the character of the
capital items. Cases of perpetual annuities are rare. The valua-
tion of plants, buildings, machines, etc., fixes attention upon a
limited service life of the particular object valued. The valuation
of bonds takes into account the maturity of all payments for in-
terest as well as for the “principal.” How complicated capitaliza-
tion would be if the “perpetual income method’ were to be used

sense at all” (ibid., p. 82), namely, the theory which explains “cost” as “displaced
alternative,” should not see this point. Consumable services available at different,
more or less distant, future points of time are the “alternative” uses of the produc-
tive services available at any one moment of time. The simultaneous output is not—
or only for an infinitesimally small part—among the alternatives.

8 Economica, August, 1934, pp. 265-66; Canadian Journal, February, 1935, pp.
14-15.

v Canadian Journal, February, 1935, p. 14.


===== PAGE 8 =====

584 FRITZ MACHLUP

is easily seen when one again faces the fact that we live in a chang-
ing world. With “bygones forever bygones,” the cost of produc-
tion of existing items does not count for their valuation. There is
now no “actual” capital value to start with, and replacement
quotas could be known only with respect to the ‘historical value”
but not to the “present value.” With replacement quotas un-
known there is no way of knowing the net yield. There is, then,
nothing but a limited series of gross returns, and this is what is
capitalized. Only afterward, when the calculation of the present
value has been finished, could a ‘“perpetual-income series” be
constructed. Take the case of finding the capital value of a plant
in an industry which is experiencing bad times. Maintenance of
the plant?® would mean negative net returns. But this does not at
all mean that the value of the plant is negative. If there are posi-
tive gross returns, for a limited period of time, of course, there
will be a positive capital value.

These few illustrations, I hope, make it sufficiently clear that
practice does not make things obscure by supposing “perpetuity”
of income or of capital. Theory should not do so either.

II. THE LENGTH OF THE PRODUCTION PERIOD

5. The concept and ils name.—It has been stated, in the pre-
ceding discussion, that the time interval between the input of
(productive) services and the “dependent” output of (consum-
able) services is significant for a number of economic problems.
This time interval was called, by Bohm-Bawerk, the “degree of
roundaboutness of production,” or the “length of the production
process,” or the ‘period of production.” The last two names,
which Bohm and most of his followers continued to use in spite of
the protests of Wicksell and others, contributed much to the con-
fusion about the concept. The words convey a meaning which is
not that of the concept: one thinks of the duration of the process
of production proper, which is by some writers called “direct pro-
duction period” (as distinguished from the duration of indirect

20 Tt is fair to note that Professor Knight does not assume maintenance of par-
ticular capital items, but only of their actual value.


===== PAGE 9 =====

THE “PERIOD OF PRODUCTION” 585

production) and by others “duration of the technical process” (as
distinguished from direct and indirect durability).

A number of different sorts of durations or periods are em-
braced by Bshm’s concept. If one had to find out what time
would elapse from the moments at which certain services, say a
number of labor hours employed in a machine shop, are being
“put in” (invested) until the moments at which the dependent
consumable services mature, the following durations would have
to be considered: (a) the duration of producing the machine; (b)
the duration of the production of the goods made with the help of
the machine, and the duration of the production of the succeeding
products made with the help of the mentioned goods, etc., all the
way up to the final stage; (¢) the durability of the machine, (d)
the durability of the goods made with the help of the machine,
and the durability of the products made with the help of the men-
tioned goods, etc., all the way up to the final services which are to
be had. All these four sorts of durations have to be considered,
but this does not at all mean that they have to be added. We
shall speak about this presently. First it should frankly be stated
that all objections to “period of production” as a name for that
composite period meant by Béhm’s concept are fully justified.
Terminological magnanimity may tolerantly pass over the mis-
nomer and continue to use it. But it certainly would be better to
return, according to Professor Hayck's recent suggestion, to
Jevons' and Wicksell's practice and speak of the “period of invest-
ment.’

6. Absolute versus average length of the period.—We warned
against the fallacy of adding up all the “durations” of direct and
indirect production and direct and indirect service life. It is easily
understood that, under the technique of our times, such adding-up
would arrive at an infinite period. Production is continuous—
not only in the sense that at any moment of time “past services”
are consumed and “future services” provided, but also in the

2 Professor Eucken, in his Kapitaltheoretische Untersuchungen, coined the word
Ausreifungsseit, i.e., “period of maturing.” The choice of the name is, however, a
matter of indifference compared with the importance of using the concept in an ap-
propriate way. In the rest of the discussion the terms period of production and
period of investment will be used interchangeably.


===== PAGE 10 =====

586 FRITZ MACHLUP

sense that durable instruments constructed in the past are used in
the construction of durable instruments for future use. In this
sense ‘‘the production period has no beginning and no end unless
. . . . the date of the end of the world is known from the begin-
ning.””? Or, more correctly, only ‘“‘the beginning and end of social
economic life’? would mark the boundaries of that “period of
production.” And “it would never be possible to give any sensible
answer to the question” when (to repeat Professor Knight's
examples) the production of a certain glass of milk “began,” or
when the process of consuming the result of productive activity
such as feeding a cow would “end.”

Has the naiveté or absurdity of the whole conception employed
by Bohm-Bawerk now been revealed and sufficiently ridiculed?
Let Bohm himself give the answer:

Where I have spoken [above] of extension or prolongation of the round-
about process of production . ... I must be understood in the sense [just]
explained. The length or the shortness of the process, its extension or its
curtailment, is not measured by the absolute duration of the period that lies
between the expenditure of the first atom of labor and the last—otherwise
the cracking of nuts with a hammer which might chance to be made of iron
brought from a mine opened by the Romans would perhaps be the most
“capitalistic” kind of production.

Can one warn more expressly against the pitfall of taking the ab-
solute period of production for the significant time interval?
Bohm’s concept was “the average period which lies between the
successive expenditure in labor and uses of land and the obtain-
ing of the final good.”? The fact that some of the services reach
into the indefinite future represents no difficulty for averaging the
time dimensions. The bulk of the services mature in the near fu-
ture and fewer in the more distant future, while the number of
services maturing after infinite intervals is infinitesimally small.

In the well-known diagrams used by Jevons, Bohm, Hayek, and
others be it the ring diagram or the triangle diagram—the ab-

22 Knight, Economic Essays in Honour of Gustav Cassel, p. 338.

2 Knight, Economica, August, 1934, p. 2753.

24 Ibid.

xs The Positive Theory of Capital (London, 1891), p. go.

20 Ibid. On the limitation to “labor” and “land,” see below (sec. 14).


===== PAGE 11 =====

THE “PERIOD OF PRODUCTION” 587

solute period of investment is assumed to be finite. This assump-
tion would imply that there is one stage of production at which no
tools and no instruments, but only the “original” services of land
and labor, are employed. If to this entirely unrealistic assump-
tion another one is added, namely, that the further input of ‘““orig-
inal” services takes place continuously at a uniform rate during
the whole absolute period, then the average period of investment
is half of the absolute.” The fiction of a finite absolute production
period facilitates the exposition of certain relations but has to be
dropped in the course of any further analysis. It is to be regretted
that it has contributed to so many serious misunderstandings.

7. Average period versus shape of the investment function.—It
might be well to familiarize ourselves somewhat further with the
meaning of the average investment period, before we deal with a
more complicated concept. Imagine for a moment that labor is
the only kind of productive service and is homogeneous in quality.
At any point of time all labor being invested at that instant is
planned and expected to yield (to contribute to) consumable
services at some future time. (Most of the input of labor of that
instant is planned to be combined with additional labor which is
to be invested at successive moments of time, and the final prod-
uct will be, of course, a joint product of services invested at differ-
ent points of time.) The labor invested today includes a certain
amount which is to yield services consumable today, another
amount which is to yield services consumable tomorrow, another
the day after tomorrow, etc. Smaller amounts of labor invested
today in some durable instrument of production will get very high
“futurity indices” and an infinitesimal amount will get the futu-
rity-index (or time-dimension) infinity. The average of all these
indices, weighted according to the respective number of labor
units, is the average period of investment.

27 There is another concept which is sometimes confused with the average period
of production, namely, the “average period of waiting,” which is dependent, apart
from the other time dimensions, upon whether production is arranged in simul-
taneously operating stages, i.e., upon whether output becomes available continu-
ously or with intervals (e.g., crop output). See Bchm-Bawerk, op. cit., pp. 327-28.
The subsistence-fund theory refers directly to the “waiting period” and through it
to the production period.


===== PAGE 12 =====

588 FRITZ MACHLUP

The “average period” as an expression for the distribution of
services over time is serviceable only for certain purposes and only
if this distribution scheme is rather simple. Imagine changing our
plans, one day, simply by switching some units of labor to activ-
ities of later maturing yield. This switching may be expressed by
saying that the average period of investment has been lengthened.
But take another type of rearrangement, where some productive
services are transferred from uses with quick maturity to uses
with later maturity, but also some other productive services are
transferred from uses with very late maturity to uses with me-
dium-high futurity indices.?® Such a change in the distribution of
services over time would be possible without any change in the
“average.” This shows that not only the average of all time di-
mensions but also their dispersion may be significant. The “shape
of the investment function’ would tell what the simple “average
investment period” conceals.

There are other reasons for not accepting the “average period
of investment” as an adequate expression for, or measure of, the
time structure of investment. Professor Hayek reminds us that
the time intervals of waiting are, as to their economic significance,
not homogeneous magnitudes which can be averaged.? Another
argument advanced by Professor Hayek is concerned with the
relationship between the interest rate and the investment func-
tion. The investment function, he explains, is rather complicated
because of the “investment in compound interest’; this makes it
not only impracticable to represent it in terms of averages but also
to represent it in one-dimensional diagrams.3 But this need not

28 Changes of this type play a réle in the business-cycle analysis of Professor
Strigl (see Kapital und Produktion [Vienna, 1934}, pp. 192-95).

29 Hayek, unpublished manuscript.

3» Hayek, “On the Relationship between Investment and Output,” Economic
Journal, June, 1934, p. 217. Three principal kinds of investment functions should
be distinguished. The most usual may be called the “cumulative investment func-
tion.” It represents either the cumulative amount of services applied at any succes-
sive moment of time in the course of the process of producing the output finished at
one moment of time or the cumulative amount of services (simultaneously) applied
at successive stages (i.e., consumption distances) of production up to the final out-
put. A second kind may be called the “service-application function,” since it shows
the application of services in the process of production at successive moments of


===== PAGE 13 =====

THE “PERIOD OF PRODUCTION” 580

frighten us too much. For most analytical purposes it will be
quite sufficient to bear in mind that one cannot do everything
with the average period without reference to the “shape of the
investment function.” But one can use it for first approximations
to the solution of a number of important problems in the theory
of capital and, I believe, in the theory of fluctuations.

8. “Time” per unit of output versus “time” per unit of input.—
Two more sources of confusion must be disclosed and blocked.
The one originated from a careless use of the word “time” to meas-
ure both the input of productive services (labor hours, service
hours, land-use years) and the interval elapsing between input and
the dependent final output. The second confusion came from
erroneous averaging of those intervals per unit of output rather
than per unit of input.

A simple numerical example (Table I) will illustrate clearly this
type of error. Under unchanged technical knowledge, unchanged
length of the working day, homogeneous labor, and homogeneous
product, three cases will be considered in one of which the average
period of investment will be lengthened (Case II), and in another
the number of workers will be increased (Case III). Both varia-
tions (support of workers through longer period, and support of
more workers through unchanged period, respectively) are as-
sumed to be rendered possible through the provisions of more
“subsistence fund.”

Comparing Case I with Case IT we see that the increase in

time or successive stages of production. It is simply the slope (first derivative) of
the cumulative investment function. The third kind, investment function proper,
represents the distribution of all productive services rendered at a moment (small
interval) of time among the uses of different consumption distances. It is but the
inverted service-application function. Services applied to the earliest stage of pro-
duction are, for the inverted figure, services applied for uses of greatest consumption
distance. Services applied to latest stages of production are, for the inverted figure,
services applied for uses of smallest consumption distances.

Professor Hayek drew a figure of the cumulative investment function and in-
verted it instead of its first derivative (ibid., p. 210). Unfortunately, he interpreted
the inverted figure as though it were the inverted service-application function. I am
afraid that this error may have confused some readers. It should be noted that for
changing conditions only the third sort of investment function is significant,
though it is for certain purposes too sensitive with regard to discontinuous outputs.


===== PAGE 14 =====

500 FRITZ MACHLUP

length of the investment period leads to a less than proportional
increase in output per labor hour. Now “critics” would say: “The
production of one unit of output needs in Case II not more but
less time than in Case I; namely, one-eleventh rather than one-
tenth hour.” The “time” of which they speak obviously means
“labor,” which is, of course, measured by time. The saving of
“time” in this sense, i.e., the saving of labor services per unit of

TABLE I
Case II Case III
(Labor Increased;
(Labor Constant; .
Case I . . Subsistence Fund
Subsistence Fund
Increased) Increased
S Proportionately)
Number of workers. ........ 100 men 100 men 120 men
Length of working day...... 10 hours 10 hours 10 hours
Average period of invest-
ment*.................. 100 days 120 days 100 days

Labor input during this period] 100,000 hours| 120,000 hours| 120,000 hours
Output during this period. . . 1,000,000 units [1,320,000 unitsf|1, 200,000 units

Labor input per day. ....... 1,000 hours 1,000 hours 1,200 hours
Output perday............ 10,000 units 11,000 units 12,000 units
Output per hour. .......... I,000 units 1,100 units 1,200 units
Output per man hour (i.e., per

unit of labor input)....... 10 units IT units 10 units
Labor input per unit of output ils hour +1 hour + hour

* The length of the investment period is assumed (not calculated from the other data). In Case I1
production is assumed to be more roundabout through the use of an intermediate tool which is entirely
used up at the end of the 120 days. In Case III the same production method asin Case I is applied.

t This figure does not show increasing returns. It has to be more than in proportion to the increase
of the average period of investment, i.e., 100: 120, even to show decreasing returns. A figure of less than
1,200,000 units of output over the longer period would imply negative return to the added 20 days of
investment. The total “waiting’’ is increased not simply by the lapse of time, but by time multiplied
by the amount invested at any moment.

output, is the very objective of the lengthening of the investment
period. The increment in product per labor hour constitutes the
“productivity” of the extension of the investment period. The
measurement of the amount of services invested in terms of time
units must not be confused with the “time dimension” of the in-
vestment in the sense of its “time distance’ from consumption.

But other “critics” would say: “Not only ‘labor time’ but also
‘waiting time’ per unit of output is less in Case II. In Case I we
had to wait 100 days for 1,000,000 units of output; in Case II we
have to wait 126 days for 1,320,000 units of output. Therefore,


===== PAGE 15 =====

THE “PERIOD OF PRODUCTION” 501

average unit waiting time in Case I was one-ten-thousandth of a
day; in Case II only one-eleven-thousandth of a day.” This
“waiting time per unit of output” is an embarrassing confusion.
The roo and 120 days, respectively, were already averages per
unit of labor input, and there is absolutely no sense in dividing
these figures by the amount of output. That it does not make
sense 1s shown by a comparison between Cases I and III, both of
which include an equal and unchanged production period. In
Case III, 1,200,000 units are produced in 100 days. The absolute
increase in output, due to the increase in men and the propor-
tional increase in ‘“‘subsistence fund,” with the unchanged period
of production would give a shorter “period of production per unit
of output’3* if this had any meaning. The production period is
the average number of waiting hours applied to the labor hour; it
is an average period only in terms of input (of services other than
waiting). If the average investment period is to be set in rela-
lation to output, not the absolute amount of output but average
output per unit of labor input is the adequate magnitude for
comparison. An ‘“‘average period in terms of total output” is
meaningless.

9. Construction period and utilization period.—The misconcep-
tions dealt with in the last sections are of very old acquaintance—
Bohm-Bawerk himself was concerned with them in his “excur-
sus.”% A novel method of arriving at erroneous averages is ap-
plied by Professor Knight. He divides the sum of the periods of
construction and utilization of different plants by the number or
value of plants, and states then, correctly of course, that an addi-
tion to the existing equipment may well “involve an increase in

3t The erroneous reasoning would be: “We have to wait 100 days for 1,200,000
units; average waiting time per unit of output is therefore one-twelve-thousandth of
a day; this is less than the one-ten-thousandth in Case 1.” It should be noted that
in Case IIT more workers are employed in an unchanged production period, hence
with a constant amount of “capital per head.” This will be discussed later. It was
on purpose that I assumed here the subsistence fund to be increased proportionately
with the number of men at work, in order to avoid the complications of shortening
the investment period necessary when less “capital per head” is available. This is
also reserved for later discussion.

32 See the third or fourth German edition of his Positive Theorie des Kapitales.


===== PAGE 16 =====

502 FRITZ MACHLUP

the capital, while shortening the average cycle.”33 He thinks of the
replacement cycle of durable capital goods. This cycle consists of
the period of construction of the equipment plus its durability.
Imagine that there are three plants, one with a period of construc-
tion plus utilization of 5 years, another with such a period of 10
years, and a third with one of 15 years. Now, additional invest-
ment is made by creating a fourth plant with a 6-year period.
(The productive services engaged in constructing and maintaining
the fourth plant must have been employed formerly in other
lines, say, in farming.) In the case of three plants Professor
Knight's “average period” was 10 years (5+ 10+ 15 divided by 3),
while in the case of four plants (more capital!) the “average peri-
od’ is only g years (5+6+10+15 divided by 4). This “average
per unit of plant,” or, weighted with the value of plant, the “aver-
age per capital invested,” he erroneously took as the average in-
vestment period and, of course, found that it did not bear the
well-known relation to the quantity of capital, viz., that it failed
to increase with the increase in investment. The “average period
of construction’ plus the ‘average period of utilization” (both
averages per unit of plant value) is by no means equal to the aver-
age period of investment (which is an average per unit of produc-
tive service inclusive of those not engaged in the construction and
replacement of durable goods).

Making goods which need a longer time of construction and
making goods which last longer are perhaps only “two details
which are of the same significance as any of an infinity of other
details,” and are certainly only two ‘among an infinite number’’34
of ways of investing more capital. But ‘“‘the two together’ do cer-
tainly not constitute, as Professor Knight so strongly believes,
Bohm-Bawerk’s period of production. It is inexplicable how Pro-
fessor Knight could arrive at such an extremely narrow definition
of the production period. Bohm-Bawerk, who is now accused of
“simply selecting these two details” and giving them “the false

33 Economic Journal, March, 1935, p. 81 (italics are his).
34 Knight, Economica, August, 1934, p. 268.
3 ¢Ibid., pp. 268-70; Economic Journal, March, 193s, pp. 78, 81, 83, etc.


===== PAGE 17 =====

THE “PERIOD OF PRODUCTION” 503

designation of length of the production process,”’s® wrote the fol-
lowing sentence about investment in increased durability:

It represents but one of the many special forms of using labor for obtain-
ing consumption goods by those roundabout processes which make it possi-
ble to get out more goods per unit of original factors of production but at
points of time which on the average are more remote from the time of input
of those factors.37

Hence there is no conflict of opinion between the critic and the
criticized that “average durability of the goods” and, similarly,
the “average construction period for such goods’ 3% need not be in-
creased if total investment increases.

10. Average durability.—In spite of the fact that it was quite
foreign to the theorists who reason in terms of investment periods
to identify this concept with average construction periods plus
average durability of goods, it should be noted that the relation-
ship between “average durability’ and “capital supply’ is much
closer than Professor Knight seems to allow for. An increase in
“capital supply,” in a money-enterprise economy, leads to in-
creased investment through a lower rate of interest. (The supply
of other services is assumed to be unchanged.) A reduction of the
interest rate decreases construction cost as far as ‘carrying
charges” are concerned and increases the present value of durable
goods because of the lower discount of future returns. The latter
phenomenon is of major significance. The increase in present
value is the stronger the more durable the instruments are.

Let us compare this effect on the value of instruments of differ-
ent durability. An instrument with a service life of one year and a
gross return of $1,000 would have a present value of $940, if the
interest rate is 6 per cent, and a present value of $970 if the inter-
est rate is only 3 per cent. An instrument with a service life of
two years and a gross return of $1,0co in each year would have
under simple interest of 6 per cent a present value of (940880 =)
$1,820, and under simple interest3? of 3 per cent a present value of

36 Knight, Economica, August, 1934, p. 268.

3710p. cit. (3d ed., 1909), p. 170; ibid. (4th ed., 1921), p. 127.

38 Knight, Economic Journal, p. 78.

39 A simple example was preferred here. Compound interest should be calculated
for almost all practical (and theoretically important) cases.


===== PAGE 18 =====

504 FRITZ MACHLUP

(970+940=) $1,910. A fall of the interest rate from 6 to 3 per
cent raises the value of the short-lived instrument from $940 to
$970, 1.e., by 3.2 per cent, while it raises the value of the long-
lived instrument from $1,820 to $1,910, i.e., by 4.9 per cent. Cost
of construction (aside from interest) being unchanged—the case
is still stronger if the more durable instrument requires a longer
time for construction and, hence, becomes now relatively cheaper
in construction—it will be much more profitable to construct the
more durable instrument.

The differences in the price increases of goods of different dura-
bility become the sharper the greater the expected service life is.
Let the rate of interest drop only from 5 to 4 per cent and there
results an increase in present value of a good with ten years’ serv-
ice life by about 5 per cent whereas the increase is more than g per
cent for a good with twenty years’ service life. (Compound in-
terest was calculated here.)

There is, hence, a very strong presumption for the belief that
increased supply of liquid capital leads to increased investment in
durability, even in the sense of “average durability of goods,” the
concept emphasized by Professor Knight.

11. “More durable’ goods versus more ‘durable goods.”’—In his
criticism of the production period Professor Knight concentrates
his forces against the alleged increase in the durability of goods.
Again and again he asserts that increased investment need not in-
volve increased durability and “therefore” need not involve in-
creased “time between production and consumption,’ i.e., a
longer investment period. The crux of the whole argument is the
idea that making goods “more durable” is, but making more
“durable goods” need not be, a lengthening of the investment
period, according to Professor Knight's interpretation of the
“Austrian” theory of capital.

Investment means, for the “Austrians” at least, using produc-
tive services for future consumable services. Increased invest-
ment means using the productive services on the average for con-
sumable outputs in a more remote future. This lengthening of
the period of maturing can be done in many ways, one of which is

40 Economica, August, 1934, p. 209.


===== PAGE 19 =====

THE “PERIOD OF PRODUCTION” 595

the making of goods and instruments of higher durability, another
1s the making of a greater amount of durable goods and instru-
ments. If the primitive inhabitants of the famous isolated island,
where fishing is the only line of production, and nets are the only
type of durable instruments, abstain temporarily from some por-
tions of consumption and add one more net to their stock, they
certainly lengthen their investment period. (Labor is withdrawn
from fishing and employed for making the net. Total labor supply
remains constant, of course.) “Average durability” per unit of
stock, per net, would remain equal. But even if they make the
new net of lower durability, and decrease thereby the “average
durability” of nets, the investment period would still be length-
ened. Or if our good old Robinson Crusoe adds to his stock of
tools, which last “on the average” one year, a new knife which
lasts only two weeks, he still lengthens his production period. By
refraining from direct consumptive use of some labor hours and
using them for making the knife (and, later on, for maintaining
or replacing the knife), he changes the average time distance be-
tween input and dependent output.

That, with a given amount of productive services employed, the
making of a greater amount of durable goods per se would not in-
volve any greater length of the investment period is a unique in-
terpretation of the concept. The emphasis placed on this point by
Professor Knight is striking. ‘“More goods of the same kind,” he
says, ‘“would mean no permanent change in either investment
function or output function, as defined by Professor Hayek.”
This should be contrasted with Professor Hayek’s statement
about the possibility of “changing the investment function for in-
dustry in general without changing it for any one industry.”’4* The
proportions within any one industry may be rigid and remain un-
changed (the durability of their equipment, of course, too), but
the proportions between industries may change. The investment
function of society as a whole changes not only if goods are made
“mere or less durable” but also if more or less “durable goods” are
made.

4 Economic Journal, March, 1933, p. 78.

4 “The Relationship between Investment and Output,” ¢bid., June, 1934, p. 224.


===== PAGE 20 =====

596 FRITZ MACHLUP

III. DIFFICULTIES AND COMPLICATIONS

12. The view backward and the view forward.—In the few places
where we, thus far, had to define the production period we tried to
avoid one dilemma which probably has not been seen clearly by
Bohm and some of his immediate followers. The question how
far into the future the output that depends on present activities
is expected to reach, on the average, is different from the question
how far back into the past reach, on the average, the activities on
which the present output depends.#* Under the assumption of a
stationary economy in an unchanging world, the “historical” pro-
duction period and the “anticipated” production period may be
equal. But how are the two “production periods’ related to each
other in a changing world with changing men of changing mood?
If people decide to change the time distribution of their current
service input, how does this affect the “age distribution” of the
output of today, tomorrow, and the immediate future? Is there
not a considerable lag of the “real” time structure of production
behind the “anticipated” one? And if the “anticipated” period
undergoes heavy fluctuations, is it likely or possible that the
“real” production structure varies concomitantly? And if this is
found to be impossible, is it then not illicit to theorize about fluc-
tuations in the production structure in the short run?

When Mr. Hill raised all these doubts he did not suggest the
scrapping of the concepts altogether but a reconsideration, or the
discontinuance, of their use in business cycle theory. The recon-
sideration may lead to a number of distinctions, refinements,
qualifications of the concept, and what has been stated about it.
So much seems to be clear: (1) that the significant hypotheses
about the investment period and its changes refer exclusively to
the so-called anticipated period, i.e., to the time distribution of
the current input of services; (2) that the “historical”’44 produc-

43 Hill, “The Period of Production and Industrial Fluctuations,” 7bid., December,
1933, Pp. 601. In drawing the distinction Hill followed F. Burkhart, “Die Schemata
des stationiiren Kreislaufs bei Bohm-Bawerk und Marx,” Weltwirtschaftliches
Archiv, Vol. XXXIV (1931).

44 Mr. Hill speaks (0p. cit., p. 603) of the “completed” period rather than of a
“historical.” It is better to reserve the term “completed” for the discussion of
other phenomena in the process of lengthening or shortening the period.


===== PAGE 21 =====

THE “PERIOD OF PRODUCTION” 597

tion structure undoubtedly offers resistance to quick changes or
even to fluctuations, a resistance which is the very substance of
the explanation of cyclical and other disturbances; (3) that a pro-
duction period between past input and present output seems to
be of no heuristic value for economic theory other than as a his-
torical account of some ‘given data.”

The “view forward” determines the employment of services
available. That “it is by its nature hypothetical, representing an
average of anticipations which may not be realized’ is not in
contradiction with the fact that it is very real in its effects on what
is being done at any one moment of time.

13. Current investment versus total investment.— The clean dis-
tinction between the “anticipated” investment period and the
“historical” period is more or less the distinction between an eco-
nomic problem and the data given for its solution. The economic
problem is the problem of choice, in this case the problem of dis-
tributing the available productive services among purposes of
different time index. The data include the total equipment avail-
able which is the result of history. The historical time distribu-
tions are materialized in the stock of “capital goods,” the quanti-
ty, composition, and quality (including durability) of which are,
of course, of influence upon the present choices as to the distribu-
tion of services over time. In other words, “past” investment,
through the man-made equipment available “at present,” influ-
ences (together with other factors) the “present” or current in-
vestment for “future” output. The investment function (the only
significant one as an economic problem) refers wholly to the current
investment. Hence it does not contain what one could call “total
investment,” i.e., the current investment plus the existing capital

as Mr. Hill writes (0p. cit., p. 605): “Owing to the existence of fixed capital the
structure of production will not contract as rapidly as the anticipated production
period. The average time in which existing specific capital-goods . . . . pass into
consumers’ goods is not likely to change appreciably over a short period.” One
should not forget that the fact that less or no services are currently employed to
construct or reconstruct such fixed capital is quite an “appreciable change,” even
in the shortest run. Moreover, the “contraction” of the structure of production
may be very rapid indeed, if value changes rather than physical changes are con-
sidered.

46 [bid., p. 610.


===== PAGE 22 =====

508 FRITZ MACHLUP

goods (durable and non-durable), which are the result of past in-
vestments.” To mix existing capital goods with current invest-
ment and to make out of both one investment function is to mix
historical data with economic problems. Past investments, so far
as they are materialized in existing equipment (capital goods),
certainly influence the shape of the current investment function
(the investment period), but they are not contained in it with re-
spect to the productive services which they will yield in the future.
Only the productive services rendered at the moment of invest-
ment“? constitute (with their consumption distances) the invest-
ment function.

14. Investment of services of “original” factors versus services in
general.—At this point in the present discussion it is high time to
dispense with the idea that only “original” factors of production
should be considered. It was undoubtedly a mistake to look only
at the services of “labor” and “land,” as Bohm-Bawerk did
throughout his discussion. Whether it was an inappropriate use
of the stationary-state fiction, or whether it was the fear of reason-
ing in a circle, which led to the exclusion of the services of existing
equipment from the investment account, it was a mistake to as-
sume that the services which can be had from capital goods are,
apart from their history, different in principle from the services of
land and labor. (That there is no homogeneity within these cate-
gories should be clear.)® At any moment of time the services
available from any sources of any kind—*free gifts” of nature as
well as human labor-capacity as well as the results of past invest-
ment—are “given” for the economic choice. Labor services, land
services, machine services, are institutionally, biologically, and

47 It should be noted that this does not fit into Jevons’ terminology. Our “total
investment” would correspond to his “amount of capital invested” but not to his
“amount of investment of capital.” See his Theory of Political Economy, chap. vii
(esp. p. 229 in the 4th ed.).

48 To make it indubitably clear: The durability of existing equipment influences

the current decisions about current investment and through it the investment
period, but it does not enter as an item into the investment function.

49 See Knight, Economica, August, 1934, p. 279. Professor Knight forcefully
criticizes the fallacy of distinguishing “primary” from “produced” factors. I also
owe much to an unpublished manuscript by Professor Hayek.


===== PAGE 23 =====

THE “PERIOD OF PRODUCTION” 599

historically different, but there is no difference with regard to their
economic position in the time structure of investment.

The resources that yield the productive services may be perma-
nent or durable or perishable. The services are perishable alto-
gether because they are related to the single moment of time. The
labor hours, land-use hours, machine hours, become, in being em-
ployed, the items which constitute the current investment.

15. The “unit” of investment.—Labor is not homogeneous in
kind, in quality, in efficiency; nor is land or capital goods. There
is not much more sense in trying to compare in ‘physical units”
the floor-space hour in a factory with the water-power hour from a
river (both in some way related to “nature” or “land”) than in
comparing them with the labor hour of a bricklayer, or with that
of a stockbroker, or with the hourly service from a steel hammer or
that from a paper machine. And yet, should all of those services
and their time dimensions be “averaged” in an investment period
or, at least, brought into one investment function? (An additional
complication consists in the circumstance that the special form of
employment of transferable and versatile factors sometimes de-
pends on the length of the investment period. )s°

Is there a way out of the dilemma? Isit fair to “assume” homo-
geneity of services, to theorize on that basis, and be satisfied with
conclusions per analogiam? Or is it better to take “value” as the
only possible tertium comparationis? There may be an impenetra-
ble series of implications in adopting value as the “unit” of input.
If it is done, the time dimensions of the productive services being
invested at a moment of time would have to be weighted with the
values of these services.

16. Investment period and the amount of capital goods.— The cor-
relation between the investment period and the amount of capital

so Much emphasis was placed on this point by Franz X. Weiss, “Produktionsum-
wege und Kapitalzins,” Zeitschrift fiir Volkswirtschaft und Sozialpolitik, N.F., I
(1921), 568-72.

st This raises a number of the most serious theoretical problems. The value of
productive services is dependent on the value of their future consumable services.
Both are dependent on the length of the investment period, since an increase in that
period yields an increase in future output. It is not only a matter of space when I
abstain from entering into so trying a discussion. I may recall to the reader that


===== PAGE 24 =====

600 FRITZ MACHLUP

goods has sometimes been held to be true “by definition,” some-
times rejected as meaningless or wrong. The customary proofs for
the existence of the correlation run again in terms of a stationary
state. The assumption of the stationary state implies that the in-
vestment function does not change but remains the same through
time. The investment function of “today” would look exactly
like that of any time in the past, and exactly like that of any time
in the future. In this case the amount of capital goods in existence
at any moment of time would be a certain function’* of the in-
vestment function. It goes without saying, then, that the invest-
ment function and the amount of capital goods in existence are
correlative.

Imagine now a change in the system; people start saving and
investing; this involves a change in the investment function
through a switch of services from lower to higher consumption
distances. Assume, furthermore, that the new investment func-
tion is being maintained from then on and that the mobility and
transformability of capital goods was sufficient to make the change
take place without any losses. The changed investment function
would tell that the physical quantity of capital goods of any type

Pigou withdrew the “Pound Sterling worth of resources” from his last editions of
the Economics of Welfare and substituted physical units for it. The issue was a
different one but similar considerations are likely to be relevant in our problem. (Cf.
the recent remarks by Howard S. Ellis, Zeitschrift fiir Nationalokonomie, June,
1935, PP. 158-59.)

52 The investment function, as we defined it, represents the distribution of all
services invested at a moment (small interval) of time among uses of different con-
sumption distance. If such an investment function is (for a stationary state) drawn
only for the services of original factors, then inverted and transformed to a cumula-
tive function (now looking like Hayek’s Fig. 1b, in his article in Economic Journal,
June, 1934, p. 210), then the area under the cumulative investment function repre-
sents the stock of produced production goods. (About the simple and the cumula-
tive investment function see also supra, p. 588, n. 30.) The statement in the text
about the constancy through time of the investment function must be qualified by
the further assumption that output is continuous. While cumulative investment
functions are insensitive to discontinuities of output, and are therefore constant for
a stationary state, the investment function as we defined it would vary according
to the crop-distance. This is a technical defect. Better technicians may find an im-
proved mode of expressing the investment function.


===== PAGE 25 =====

THE “PERIOD OF PRODUCTION” 601

will be greater in the future than it is at the moment of change.
One could hold, however, that, although the physical quantity of
capital goods is still unchanged at the very moment of the change
of the investment function, there may be an immediate increase in
stock value because of the lower capitalization rate of future in-
come. In other words, the same “causes” which change the in-
vestment function change at once the rate of capitalization and,
therefore, the value of existing equipment.

Another approach seems to be better adapted to the concepts as
we have used them in this discussion. The relationship between
the amount of capital goods and the investment period shows it-
self in both directions. First, the greater the amount of service-
able capital goods in existence, the greater is the part of the cur-
rent productive services from resources of any kind which can be
invested for a more distant future. Thus, the amount of capital
goods in existence represents, so to speak, a potentiality of invest-
ing more for longer periods. It is a potentiality which has not
necessarily to be made use of. (And, indeed, capital goods may
lose in value if and when, through a change in time preferences,
given “specific” equipment cannot be utilized fully.) But, gen-
erally speaking, the investment period tends to be longer the
greater the amount of capital goods in existence. Second, if we
use an increased part of the given productive services for things of
greater time distance, the amount of capital goods in existence
will increase. This means that, other things remaining equal, a
lengthening of the investment period gradually increases the
amount of capital goods in existence.

17. Investment period and the supply of free capital disposal.—
In a single-plan economy, from the simple economy of the Crusoe
type up to a communistic economy, the change in time preference

s3 If all productive services are consumed simultaneously, no “intermediate
good” can exist. If production takes time, goods in process, and perhaps tools, are
likely to be on the “inventory” at any moment of time. If production is arranged
in stages which work continuously, the more goods must be on the inventory the
longer the production period which is chosen. With a given input (per time) of pro-
ductive services the inventory, at any moment of time, must be greater the greater
the average time interval is until output is consumed.


===== PAGE 26 =====

602 FRITZ MACHLUP

(in the plan-makers’ taste) and the change in the investment func-
tion are analytically the same. The time preference expresses it-
self in the distribution over time of the use which is made of the
productive services currently available.

In an exchange-enterprise economy the time distribution of pro-
ductive services is not uno actu changed with the time preferences
on the part of certain individuals. Individuals who save transfer
their disposal over services to other individuals who invest. In-
creased supply of “free capital disposal”’s* tends to reduce the ex-
change ratio of present to future income, which in turn leads to
increased investment of productive services for longer periods.
Individuals who disinvest, so far as they cannot draw upon the
disposal over services released from individuals who save, with-
draw the disposal over services from entrepreneurs. Decreased
supply of “free capital disposal” tends to increase the exchange
ratio of present to future income, which in turn leads to decreased
investment of productive services for longer periods.

In a money-enterprise economy ‘free capital disposal” is sup-
plied and demanded in the form of “money capital.” Insight into
the functioning of this system is much impeded by the fact that
goods in process are passed over from firm to firm by way of ex-
change against money, whereby money appears to be free capital
disposal for the individual without being so for the community as
a whole. This is why the unreal assumption of production verti-
cally integrated in a single firm is so convenient; and why under
more realistic assumptions problems of great complexity emerge
from monetary changes. So much seems to be clear that an in-
crease in the supply of money capital tends (other things, includ-
ing the amount of services, being equal) to lengthen the invest-
ment period; and, vice versa, a decrease in the supply of money
capital tends to shorten the investment period. And this is so,
whether disposal over services really is released by savers or with-
drawn by dissavers, respectively, or whether only the “elasticity”

s+ 1 follow Cassel in using the term “capital disposal.” It is better to avoid the
term “capital” simpliciter. The terms “capital goods,” “capital disposal,” “capital
value,” “money capital,” convey more definite meaning than the ambiguous word
“capital.”


===== PAGE 27 =====

THE “PERIOD OF PRODUCTION” 603

of the money system does the trick. Thus changes in the invest-
ment structure are inseparably linked with the credit cycle.

18. “At” a point of time versus “within” a period of time versus
“for” a period of time.—Discussions of “capital” and ‘“‘invest-
ment’ suffer not only from ambiguous meanings of these terms
but also from confusion with respect to their time quality. The
terms suggested by Jevons were not striking enough to be ac-
cepted by later writers. The “amount of investment of capital”
was to mean something different from the ‘‘amount of capital in-
vested,”’ss but neither of them was the “amount of capital being
invested.” The first was to be a product of a multiplication (more
correctly, integration) of the amount that has been invested by
the time for which it has been invested. The second, the “capital
invested,” was the amount that has been (and continues to be) in-
vested up to a certain point of time, but without consideration of
‘““the length of time during which it remains invested.”’s® In other
words, it was the inventory at a point of time. The third is the
amount that is being invested within a (very small) interval of
time.

“Capital invested” in the sense of capital that has been in-
vested can, in a changing world, apart from setting up a profit-
and-loss account, be significantly expressed, not in terms of its
(historical) cost—i.e., in terms of the value it had at the time
when the investments were made—but only in terms of the pres-
ent value of the goods or resources which exist as result of those
past investments. The present value is, of course, the capitalized
income which is expected to accrue from the services from these
goods or resources. Since the present value of these goods or re-
sources has not much to do with the historical amounts invested,
the term ‘“‘capital invested” had better not be used for them.
“Capital equipment” or ‘“‘capital goods’ are adequate terms.

“Capital invested” in the sense of capital that is being invested
within a small interval of time can be expressed in terms of cost
or value, both being the same under the assumptions of equilib-
rium theory. Can this ‘‘capital being currently invested” be ex-
pressed without time dimension? It certainly needs a reference to

55 Jevons, 0p. cit. (4th ed.), p. 229. 56 Ibid.


===== PAGE 28 =====

604 FRITZ MACHLUP

the unit of time chosen, since it is a time rate (such as per hour,
per day, per week). Isit necessary to add to the time within which
it is being invested, also the time for which the investment is to
be? The answer to this becomes clear if we recall what is being
invested. If we think of capital disposal or of money capital, it
can be expressed as a mere quantity without reference to the dura-
tion of investment. But if we think of the “real” investment, of
productive services, we immediately see that the time dimensions
become necessary because almost all productive services employed
—except the personal services of servants, actors, etc.—are “in-
vested” and the question is only one of their distribution over
time. Under full employment no increase in investment would be
conceivable if it were not for the investment of services for periods
longer than those for which they were hitherto invested.

This circumstance is, I think, rather significant. The money
capital which is being invested need not, and cannot, be assigned
to any certain period of investment. Even if the individuals who
supply it were quite determined as to the length of time for which
they would like to tie it up, or even if the individuals were deter-
mined to invest it for eternity, the services which are invested by
means of the money capital are invested for a period which is in-
dependent of the will of the particular saver. And the single new
investment makes, ceteris paribus (hence, with the amount of
services unchanged), the average period of investment just a little
longer than it was before. If there is full employment, the increase
in the amount of money capital being invested per unit of time in-
creases the period for which the services are being invested on the
average.

19. Capital per capita.—Productive services are distributed
among uses of different consumption distance. To distribute a
given amount of services is one problem. To distribute the serv-
ices from a given amount of resources and men is another prob-
lem, because the capacity of resources and men may not be util-
ized fully. To distribute the services from a changing amount of
resources and men is a third problem. It goes without saying that
these distinctions are highly relevant for the problem of the aver-
age period of investment or, more correctly, for the problem of
comparing investment structures.


===== PAGE 29 =====

THE “PERIOD OF PRODUCTION” 605

A greater amount of services distributed over time (over con-
sumption distances) in the same proportions as a smaller amount
of services means, of course, an unchanged average period. Should
there be an increase in the number of men employed, the amount
of other resources would have to increase too, if a change in time
distribution is to be avoided. This is often expressed by saying
that changes in the average period of investment can be seen by
changes in the amount of “capital per head.” If population grows
more quickly than capital equipment, with the result that less
“capital per head” is available, shorter periods of investment be-
come necessary. On the other hand, increased investment per-
mits lengthened investment periods only so far as it is not accom-
panied by a proportional increase in men.

This interrelation (which I chose to formulate as a statement
of tendencies) is sometimes suggested in the form of an analytical
statement. If a cumulative investment function is drawn in
the customary way (representing cumulatively the amounts of
services applied at successive moments of time in the course of
one process of producing the output finished at one moment of
time),57 the area under this investment curve is defined as capital
(more correctly, capital goods in terms of incorporated services).
The average period of investment is measured by dividing the area
under the curve by the final output (i.e., by the income measured
in terms of input or incorporated services.) This division repre-
sents, by definition, the ratio between capital and income.?® If
only labor income, rather than total income, is considered, and if
it is (according to the ordinates of these diagrams) expressed in
terms of service input, the analogous division gives the ratio be-
tween capital and labor services, or, if labor hours per laborer are
fixed, capital per head.s

57 See, among recent publications, the diagram in Marschak’s article, Economic

Journal, March, 1934, p. 147, or Fig. 1a in Hayek’s article, ¢bid., June, 1934, p. 210.
For a short explanation see above, p. 588, n. 30.

58 Professor Knight (Economic Journal, p. 88) considers this ratio “one of the
least meaningful.”

$9 Which of the two ratios is more significant for capital theory is still an open
question. The discussion in the following paragraph places more emphasis on the
second ratio. I owe much to Dr. Marschak’s previously cited article and to his
unpublished contribution to our private “round-letter’”’ discussion.


===== PAGE 30 =====

606 FRITZ MACHLUP

An increased supply of money capital with unemployed labor
services available would lead to an absorption into the production
process of more labor services rather than—or in addition to—a
redistribution over longer investment periods of the hitherto em-
ployed services. (Without increase in the supply of money capital,
such absorption would be practicable only by a shortening of the
investment period with lower returns to labor services). There
would, however, be a different sequence of events if the additional
money capital is supplied by an expansion of bank credit from
what it would be if it were supplied by voluntary saving, i.e.,
under voluntary abstinence from consumption.

20. The proportion between the factors employed.—The reflec-
tions upon “capital per head” above suggest the coherence be-
tween the concepts of the “investment period” and the “propor-
tion between the factors employed.” So far as changes in the in-
vestment period are connected with changes in methods of pro-
duction within single industries or even single firms, reasoning in
terms of proportion between factors becomes necessary. A single
firm which, under the influence of lower interest rates, buys new
machines to substitute them for less efficient ones or for labor
services, certainly does not calculate investment periods. The
firm makes cost comparisons, and the elasticity of substitution®®
will, among other conditions, be decisive for the firm’s choice. The
result is, for the particular section of the economic system, to be
stated only in terms of the proportions in which the factors are em-
ployed by the single firm or industry.

An average investment period or a certain investment function
becomes significant only for the community as a whole. In a total
system analysis it will sometimes be more convenient to theorize
in terms of the proportion between the factors employed, some-
times in terms of the investment period. The use of the latter con-
ception recommends itself in all problems where explanations of
changes through time are essential. The “analytical correlation,”

So See J. R. Hicks, The Theory of Wages (London, 1932), p. 117; Joan Robinson,
The Economics of Imperfect Competition (London, 1933), pp. 256-60; C. H. P.
Gifford, Econometrica, April, 1935, pp. 208-10; Fritz Machlup, “The Commonsense
of the Elasticity of Substitution,” Review of Fconomic Studies, Vol. II (June, 1935).


===== PAGE 31 =====

THE “PERIOD OF PRODUCTION” 607

mentioned in the preceding paragraph, is, of course, without sig-
nificance for causal relations. What are really significant are ques-
tions as to the probable effect upon the investment period of an
increased supply of money capital, and as to the effect upon the
quantity of capital goods (producers’ goods) of a lengthening of
the investment period. Lags and friction in the realization of such
sequences are not unimportant; the appreciation (increase in
present value) of the bulk of existing capital goods in the case of a
lengthening is not less relevant than the differentials for longer-
lived and shorter-lived goods and (in the case of a sudden increase
in voluntary saving) the depreciation of some specific instru-
ments adapted for the production of consumers’ goods.

The chances for a statistical measurement of the investment
period are not too good,’ for reasons which have been sufficiently
indicated in the discussion of the difficulties and complications.
Of these difficulties the greatest is the problem of comparability—
which is present throughout the field of economic statistics and is
comprehensible to anybody who has a clear understanding of sub-
jective value theory. The chief working hypothesis of every sta-
tistical attempt at measuring the investment period seems to be its
approximate correlation with the proportions in which the factors
are employed.

IV. SPECIAL PROBLEMS

21. Width of income stream versus length of investment period.—
A great many economists substitute for the concept of the “length
of the investment period” the concept of the “width of the income
stream.” The latter concept does not encounter so many compli-
cations as the former and has the advantage of not only suggesting
a picture readily grasped but also of easily fitting into any analysis
of economic dynamics, i.e., of fluctuating, progressive, or declining

6x See Marschak, “Economic Parameters in a Stationary Society with Monetary
Circulation,” Econometrica, January, 1934; and Gifford, Econometrica, April, 1935.

62 Lack of statistical “verification,” or even theoretical impossibility of measure-
ment, is no evidence whatsoever against the usefulness of a concept. As Professor
Knight has said, absolutely to the point, clear and indispensable notions often
“cannot be given exact definition; but this limitation applies to all quantitative
analysis in economics” (Economic Journal, March, 1935, p. go n.).


===== PAGE 32 =====

608 FRITZ MACHLUP

economies. Saving is pictured as a narrowing of the income
stream, which later becomes wider again, wider than it was before
the narrowing; dissaving is pictured as a widening of the stream,
which later becomes narrower again, narrower that it was before
the widening. There are possibilities of continuous and of dis-
continuous changes in the width—in short, the concept is useful
and adaptable.

Yet, it is a mistake to believe that one could dispose of the in-
vestment-period concept without disposing of one of its most sig-
nificant purposes: to solve the time problem. Of course the “in-
come stream’ is a function of time, but one which remains un-
known as long as the “investment period” is not introduced. It is
certainly true that the income stream becomes wider again after
having been narrowed by voluntary saving—but only after a
while. After how long a while? How long a time does it take be-
fore the widening of the income stream occurs, and how long be-
fore the “abstinence” is fully offset? The answer may be sought
by using a concept like the “investment period,” or it may not be
sought at all. What relates the changes in width of the income
stream with one another is the investment period which has been
thought to have been thrown overboard. That these intervals
have meaning and significance is admitted by almost every econ-
omist—even by Professor Knight.%s

22. Investment period and the “stages of produciion.”—In dis-
cussions of the length of the investment period one ordinarily
meets also propositions about the ‘stages of production.” It
would, however, be too optimistic to believe that these terms have
a clear and unambiguous meaning. Although there is much over-

83 See Economica, August, 1934, p. 273: “In reality most investments not only
begin at a fairly early date to yield their income in consumable services . . . . but in
addition they begin fairly soon to yield more than interest on cost in this form, and
entirely liquidate themselves in a moderate period of time” (italics are mine). Or, on
p. 278: “Any net saving and investment naturally means reduction in consump-
tion somewhere and subsequent increase somewhere, raising the total rate of con-
sumption in the system above that obtaining before saving started, the increase
measuring the rate of return. In terms of income sacrificed for an interval and ulti-
mately more than made up, the notions of roundaboutness and waiting have some

meaning . . . .” (italics are Professor Knight's). Any investment is displaced earlier
consumption income; this is the very point of the theory of the investment period.


===== PAGE 33 =====

THE “PERIOD OF PRODUCTION” 609

lapping, one can distinguish “stages” of production in terms of
(a) units of technical acts or divisions of the process (different
acts), (b) locational units (different plants, buildings), (¢) voca-
tional units (different crafts, professions), (d) financial units (dif-
ferent firms), (e¢) time units (different consumption distances).
Only the last of these concepts is inherently connected with the
concept of the investment period, while the others are sometimes
merely phases of the underlying technical conditions, sometimes
among the more or less probable accompanying phenomena, but
sometimes entirely unconnected.

In the attempt at illustrating an abstract exposition by con-
crete examples many authors pictured the lengthening of the in-
vestment period by pointing to the introduction of certain tech-
nical or financial stages of production. Readers—indeed, some-
times the authors themselves—then took the technical or financial
properties of the process for the essential factors. In the discussion
of production processes without durable instruments, the diver-
gence of the technical stages from those in terms of “time distance
from consumption’ is not too disturbing; but when durable in-
struments are introduced into the scheme, the divergence between
the two concepts becomes considerable. The steel plant and the
machine shop, two definitive stages, of course, in the technical
process, represent by no means two stages in the sense of con-
sumption distance but an infinite number of such stages. One has
to keep in mind that services employed for the making of a ma-
chine are, as to their consumption distances, distributed over a
great range of time. Unless one goes through the procedure of
averaging these time distributions for any single unit of equip-
ment, or unit of plant—which is neither necessary nor possible—
one must not consider a certain technical stage of production as a
certain time stage of production.

The divergence between financial stages and time stages is not
much less remarkable. This was not made clear enough by Pro-
fessor Hayek in his Prices and Production, and led to much con-
fusion. His diagrams are devised to show the money stream be-
tween the financial units engaged in the production process. That
the lengthening of the investment period (i.e., the increase in the


===== PAGE 34 =====

610 FRITZ MACHLUP

time stages of production) would involve an increase in money
transactions to be performed because of an increase in the number
of financial stages of production was an assumption of Professor
Hayek's which he certainly did not state clearly enough and which
was justly criticized by Professor Ellis.®4 In anticipation of the
confusion, Bohm-Bawerk had expressly given warning that the
length “of the process, its extension or its curtailment . . . . is not
measured by the number of independent intermediate members
which the production process embraces.

23. The investment period during new capital construction.— ‘In
a sense, all work done in a stationary society is replacement
work.”% This mode of approach has perhaps much more justifica-
tion than the crude division of the production system into pro-
ducers’ goods and consumers’ goods production, whereby the
“early stages’ of production are assumed to be engaged in the
one, the “later stages’ in the other part of that process. Accord-
ing to this last view, the number or the work of the earlier stages
1s imagined to be increased if society becomes progressive. The
question then arises whether lengthening of the investment period
is not confined to the period of capital accumulation, while short-
ening occurs when the process of accumulation comes to an end
and a stationary state sets in.

This is Professor Knight's view. If more capital goods are con-
structed, he says, “there will be, temporarily (while the expansion
1s taking place, but not after it is completed), a slight increase in
the proportion of goods in the earlier stages of processing opera-
tions, in comparison with later stages.”*” It will be shown, never-
theless, that a society which stops constructing new capital and
confines itself to maintaining capital does stop lengthening its
investment period; it need not shorten it if the transition is not
too sudden.

¢4 Howard S. Ellis, German Monetary Theory, 1905-1933, (“Harvard Economic
Studies,” Vol. XLIV), p. 354. “A given increment of capital has nothing to do
with the number of independent productive stages through which the intermediate
products pass.”

8s Positive Theory, p. 91.
66 Marschak, Economic Journal, March, 1934, p. 151.
67 Knight, Economic Journal, March, 1935, p. 78 (italics are his).


===== PAGE 35 =====

THE “PERIOD OF PRODUCTION” 611

The essential point in the argument is the gradual transition
from a state of new capital formation to a state of pure capital
maintenance. It is easy to sce that there is a rate at which new
saving and investing may slow down, and finally dwindle, with-
out necessitating a shortening of the investment period (and diffi-
culties in the construction-goods industry). Any increment of
capital involves an increment in replacement funds necessary for
full maintenance. If society is “progressive” but at a decreasing
rate, the funds supplied for maintenance still continue to increase.
The decrease in the supply of new savings may, therefore, be com-
pensated by the increase in the supply of replacement fund, so
that the total supply of free capital disposal remains equal.®® In
other words, the increase in construction of equipment for re-
placement purposes may offset the decrease in construction of
new capital equipment. If the rate at which saving and investing
falls off does not surpass these limits % no shortening of the invest-
ment period need ensue.

On the other hand, it should be noted that if the increase in
capital takes place within too short an interval, it will not be suffi-
cient to maintain the lengthened investment period. If in an
economy which has been stationary hitherto one single investment
is made and not followed by any further complementary invest-
ments, the single investment is, generally, bound to be wasted.”
The lengthening of the production process is itself a process which
needs time to be completed and, hence, to become definitive. But
if after a sufficiently long phase of capital formation the transition
from this phase to a phase of pure maintenance of capital is grad-
ual enough to ‘‘permit the processes which have already been

Suppose a current replacement fund of 300 units and new capital formation of
go. In the next interval the replacement fund will have to be (if the replacement
quota is 1o per cent) 304 so that new capital formation can fall to 36 without causing
a reduction in the supply of capital disposal (see Fritz Machlup, Borsenkredit,
Industriekredil und Kapitalbildung [Vienna, 1931], p. 100).

69 These limits are determined, as shown by Fowler (Te Depreciation of Capital
[London, 1934], pp. 30-54), by the average durability and its relation to the average
construction period of capital goods. Here is the place where these concepts used
by Professor Knight (see above, sec. 9) adequately come in.

7 For detailed explanations, see my book, op. cit, pp. 103-11; and Hayek,
Econometrica, April, 1934, pp. 153-58.


===== PAGE 36 =====

612 FRITZ MACHLUP

YY ~

started to be completed,” the lengthening of the investment
period will have been not temporary but definitive.

24. Changes in demand.— Changes in the investment period are
not always caused by changes in time preferences. Purely quali-
tative changes in demand of intentionally equal time index may
cause changes in the investment function. For a single-plan econ-
omy, however, this is not true, since the omniscient plan-makers
will see whether or not the changed tastes can be satisfied without
any different, or more or less, equipment and, hence, without any
changes in the time distribution of services. If changed tastes
could be satisfied only by a changed time structure of production,
the plan-makers would immediately have to balance the dis-
placed alternatives, which in this case would be alternatives of
different time index. In a money-enterprise economy the situation
is different, because of the lag of cost reactions behind consumers’
choices. The prices of consumers’ goods between which the indi-
viduals choose may be, at the time of the choice, different from
what they would be after all rearrangements of productive serv-
ices necessary for the provision of the goods demanded had been
performed or calculated. The change in investment periods comes
about by way of interactions of entrepreneurs, and it is only
through consecutive changes in prices that consumers find their
changed tastes to be incompatible with unchanged time prefer-
ences.

Using less abstract language, we may say that the consumers’
choices between goods made from different materials or made in
different types might cause major adaptations in the productive
equipment. If the specificity of equipment does not permit tech-
nical adaptation, the change can be made only out of replacement
funds. The quasi-rent in the still underequipped line of produc-
tion becomes apparent in higher rates of interest offered for free
capital disposal. In this way replacement funds not only from the
overequipped line of production (overequipped relative to the
changed demand) but also from quite different industries are at-
tracted by the more profitable production. Resulting price in-
creases in other products force the consumer to reconsider the

m Hayek, ibid., p. 157.


===== PAGE 37 =====

THE “PERIOD OF PRODUCTION” 613

distribution of his budget. (This redistribution may include the
item “saving” because it is not only prices of commodities but
also the rate of interest which undergo change.)

The adaptation in the productive apparatus to a change in de-
mand may lead either to merely temporary or to definitive
changes in the investment structure. This depends on two cir-
cumstances: first, a change in demand so sudden and sharp that
the producers of the product that fell into disfavor cannot earn
any replacement quotas (definitive loss of the sunk investment)
will tend to definitive shortening of the investment period; sec-
ond, the quasi-rents of the producers of the favored product, if
saved and invested, might counteract the first tendency.

There is another point calling for attention. Apart from the
losses and gains of transition, differences in the “capital intensity”
of the goods between which the demand has shifted may be im-
portant. The favored product may not only need different equip-
ment but more or less equipment than the product in disfavor has
needed for its production. The shift in demand may, for instance,
finally be “capital saving.” (Crude examples: Tents preferred to
houses; services of stage actors preferred to services of automo-
biles.) Whether or not such changes lead to shorter investment
periods depends on many questions, e.g., whether consumers use
all the money diverted from capital-intensive products for pur-
chases of less capital-needing products (or what else they use it
for); whether the sunk capital can be extracted from the indus-
tries in disfavor; whether the capital extracted from these indus-
trics is used for reinvestment elsewhere (probably under reduced
interest rates), etc.

In an economy in which new savings are continually forthcom-
ing, a part of the new savings will be used for financing the adapta-
tions of equipment to changed demand. Since the presumption
that net losses of capital are connected with these adaptations is
very great, it is fair to assume that an economy can be stationary
in respect to its capital base only if it provides new savings to re-
build its equipment needed for changed tastes. An economy
which is stationary in terms of new savings (i.e., savings are zero)
is declining in respect to its capital base, if changes in demand oc-


===== PAGE 38 =====

614 FRITZ MACHLUP

cur. Or, to put it in another way, “quick change in the objects of
consumption without the emergence of new savings is itself a
form of consuming capital.”’”” Changes in demand represent,
therefore, one of the qualifications of the proposition that any in-
crease in the supply of capital disposal leads to a lengthening of
the investment period. Changes in demand represent, further-
more, one of the important, but neglected, points in the theory of
the natural rate of interest. Changes in demand tend to raise the
equilibrium rate of interest” and may involve a sufficient impulse
for an (inflationary) expansion of bank credit.

25. Changes in technical knowledge (invemtions).—The inven-
tion case was the basis of the very first critical discussions follow-
ing publication of Bshm-Bawerk’s Positive Theory. The greatest
part of these discussions was occupied by repetitions of simple
misunderstandings. The misnomer ‘period of production’ led
critics to look at the duration of the direct technical production
process proper, and then to believe that almost all inventions
shorten this period. Sewing by hand is a much slower method of
making dresses than sewing by machine. Hence, thought the
critics, the production period was shortened by the invention of
the sewing machine. Bohm-Bawerk had a hard time to explain
that sewing of the dress was only the last part of the process,
while in the process as a whole services were invested in iron pro-
duction and machine shops and in the sewing machine and only
finally in the dressmaking.

In his ardor to refute the assertion that the production period
is shortened by such inventions, Bohm went so far as to assure us

2 Machlup, “The Consumption of Capital in Austria,” Review of Economic
Statistics, XVII (January, 1935), 14.

73 The effect upon the equilibrium rate of interest of changes in demand can be
the result of several opportunities of investment, e.g.: (a) financing technical
adaptations; (0) financing quick construction of the different equipment needed for
the newly favored product; (¢) financing construction of the increased equipment if
the newly favored product is capital-intensive; (d) financing construction of differ-
ent and increased equipment which was, due to lumpiness, hitherto not to be used
profitably but becomes so if the demand for the newly favored product reaches a
certain minimum.

74 See especially Bohm-Bawerk’s “Exkurse,” e.g., “IExkurs II,” in Positive
Theorie des Kapitales, IT (4th ed.), 44.


===== PAGE 39 =====

THE “PERIOD OF PRODUCTION” 615

of the contrary, which is not true either. The invention neither
shortens nor lengthens the investment period. All it does is to
make investment opportunities more or less profitable. If and
when new capital disposal is supplied (e.g., by new savings), then
the investment period can be lengthened. But if these ne
ings would have come forth in any case, then the inve
period would have been lengthened in any case. The new
tion, if it was of the mentioned type, merely made the i
rate higher than it would have been otherwise.

If there is no increase in the supply of money capital above the
current supply of replacement funds, then there is generally no
possibility of a lengthening of the investment period. But if the
new invention is connected with some especially durable instru-
ments, will not substitution of new instruments of itself involve a
lengthening of the investment period? If exploitation of the in-
vention appears to be profitable in spite of its taking so much
capital away from other industries (under a sharp rise of the in-
terest rate), so many other investments which would otherwise
have been made are omitted that the investment period is left un-
changed. The high rate of interest might, of course, induce people
to save or to dissave, and in this way lead to a change in the in-
vestment period.

New inventions may suggest new sorts of equipment which
compete with, and render obsolete, existing equipment. This may
cause major losses of sunk investment ; the investment period may
be shortened (in case replacement funds have to be attracted from
other industries) or may be left unchan
are just sufficient to finance the new
a smaller degree than would otherwi
would be quite unwarranted to con
undiminished or even increased pr
less capital or in a shorter investme
resenting such a development. (O
been known many years earlier, many past investments would
have been made differently and their obsolescence saved.) Ob-

solescence occurs if the new equipment is relatively more efficient
orif it is cheaper. The latter is a case of a‘

Ww sav-
stment
inven-
nterest

ged (in case new savings
investment) or lengthened in
se have been possible. But it
sider this a social loss. If an
oduct can be provided with
nt period, there is no reason for
f course, if the inventions had

capital-saving” inven-


===== PAGE 40 =====

616 FRITZ MACHLUP

tion. The old equipment is depreciated to the lower value of
the new, because the price of the products of this equipment
would no longer cover replacement of the former value. What
the consumer had to contribute to this replacement now becomes
his increased purchasing power, which may be spent for increased
consumption or which may be saved. The capital loss through
obsolescence is capital consumption in the sense that it is accom-
panied by increased consumption, but it is not capital consump-
tion in the sense that it renders impossible permanent mainte-
nance of consumption on the higher level.

In substance, the effect of inventions, capital-needing or
capital-saving, upon the investment period is per se neither
lengthening nor shortening; their indirect effects depend on the
incidental losses of transition and on the incidental changes in
saving and dissaving.

26. Investment period and disinvestment period.—Changes in
technique, it has just been shown, need not involve changes in the
investment period. But neither need changes in the investment
period involve changes in the technical methods employed. In
an economy where no durable goods exist and where no technical
substitutability exists, so that the proportions between the factors
in the production of any one product are fixed, shortening of the
investment period would be possible through changes in the rela-
tive proportions of different industries. There are two ways of
shortening the investment period without changing the technique
of production in an economy which employs durable goods. One
way is changing the proportions between industries; the other is
neglecting temporarily the replacement of durable goods. To say
this is not to assert that these are the only or the most likely pro-
cedures in disinvesting societies; but it is necessary to warn
against the belief that an unchanged technique of production
offers any evidence for an unchanged investment structure.

Consideration of the “disinvestment period” is necessary for
the analysis of a number of problems connected with capital dis-
investment either as a phase of the business cycle or as a general
trend,’”s as well as of problems connected with the adaptation of

7s See Machlup, “The Consumption of Capital in Austria,” Review of Economic
Statistics, XVII (January, 1935).


===== PAGE 41 =====

THE “PERIOD OF PRODUCTION” 617

specific capital equipments to changing conditions. It is neverthe-
less not true that immobility or specificity of equipment are the
only factors that matter. Even if every good were perfectly trans-
formable into any other good, if we, therefore, could eat up ma-
chines and buildings directly rather than through a lengthy proc-
ess of undermaintenance, even then the disinvestment period
would remain significant for the questions how long ‘“‘overcon-
sumption’’ can be continued, and after what interval the consecu-
tive shrinkage in width of the consumption stream will come
about.

The statement that “‘in equilibrium the period of investment
for the total volume of capital . . . . is equal to the period of dis-
investment’? is liable to misinterpretation. Within what time
the total capital of society can be disinvested is not the point in
question. The notion of the investment period, being an “‘aver-
age” of an infinite number of time indices, would not help to an-
swer such a question. To put the question for the length of the
disinvestment period in terms of an individual investment would
be still less sensible. The individual investor can exchange his
property rights against present consumption income (of course
not at a fixed ratio) at any time, whether they are taken over out
of new savings or, in want of new savings, out of some replace-
ment funds. And the marginal investment at one moment of time
will hardly concern the same lines of production as the marginal
disinvestment at another moment.

The use of replacement funds for ‘‘taking over” property rights
of disinvesting individuals, hence the use of replacement funds for
the disinvestors’ consumption,’ is the essential point. The speed
or rate of supply of replacement funds, on the one hand, the con-
sequences of withdrawing parts of them from reinvestment, on the
other hand, are dependent on the time structure of investment.
The average time interval from investment to consumption is also
the average time interval from investment to the forthcoming of
the capital disposal free for reinvestment, because it is the receipts

i Vowler, The Depreciation of Capital, p. 35.

i7 Not necessarily the disinvestors’ personal consumption; if they disinvest, say,
in order to pay taxes, state employees might be the persons who do the consuming.
Disinvestment for the sake of financial liquidity is to be discussed later.


===== PAGE 42 =====

618 FRITZ MACHLUP

from the sales of the consumption goods which provide the re-
placement quotas for all the earlier stages of production.

Imagine a barter-enterprise economy with vertically integrated
production. The entrepreneur exchanges the final products of his
production, or other consumption goods which he received in ex-
change for his, against productive services, if he wants to con-
tinue the enterprise, i.c., if he wants to reinvest. But he eats them
himself if he does not want to reinvest. In a money-cnterprise
economy the money receipts for the final consumption goods are
the funds available for reinvestment or disinvestment. The in-
dustry where a disinvestment takes place—i.e., where reinvest-
ment is omitted——need by no means be the same as, or be closely
connected with, the industry which received for its products the
money that “financed” the disinvestment, first, because replace-
ment, though continuous in the economy as a whole, is not con-
tinuous in single lines of production; second, because the money
capital circulating in an economy with vertically differentiated
production (i.e., not integrated in one firm) is a multiple of the
“real” replacement funds. This monetary phenomenon is partly
responsible for the difficulties, in practice as well as in theory,
which result from the fact that the money capital de facto avail-
able for disinvestment exceeds in so large a degree the money
capital “genuinely” released from production through the final
sale of consumption goods. The meaning of these phrases is more
readily apparent if one imagines the money funds which would be
available for reinvestment or disinvestment in an cconomy with
vertically integrated production.

V. THE CREDIT CYCLE
27. Credit expansion and investment period. —Analysis of the
consequences of an increase in the amount of money capital
given to entrepreneurs has to cover a number of possibilities.
Apart from an increase in the holdings of cash reserves (hoard-
ing), and from an increase in the money work to be done on ac-
count of a possible increase in “interfirm transactions,’ the incre-

8 On this point see my Borsenkredit, Industriekredil und Kapitalbildung, pp.
15-10; and Strigl, Kapital und Produktion, pp. 30-37.


===== PAGE 43 =====

THE “PERIOD OF PRODUCTION” 619

ment in money capital may be used (1) for employing a given
amount of services for other work than that for which they were
hitherto employed; (2) for paying higher rates per unit of service;
(3) for employing a greater amount of services from the factors
hitherto employed (longer labor week); (4) for employing a greater
amount of services from factors hitherto unemployed.

An analysis that begins with full employment and given supply
of services has to take into consideration only the possibilities
(1) and (2). But these two will have to go together under almost
all assumptions. If the increase in the amount of money capital
given to producers is due to voluntary saving, hence to a decrease
in the amount of money spent directly on consumers’ goods, the
tendency toward a rise in the prices of productive services will be
smallest. But there will still be a tendency for such a rise, be-
cause marginal-value productivity of labor services should be as-
sumed to rise in those employments which are made profitable by
the interest-rate reduction that accompanies the increased supply
of money capital. The rise in wage rates will, of course, be sharper
if not voluntary saving (abstinence from consumption), but credit
expansion is the source of the increased supply of money capital.
But in both cases the first of the mentioned possibilities is rather
certain to come true, and, according to what has been said above
(see especially secs. 17 and 18), the work that will be favored at the
expense of some other work will be of a type that yiclds consum-
able services at, on the average, more distant points of time.

The sequences of events characterized as lengthening of the in-
vestment period have mostly been described for the case of a fixed
supply of productive services. Increased employment of services
—i.e., the possibilities (3) and (4) above—have been ruled out.
This was correct as a first approximation, but a further step is to
analyze the sequences without the assumption of fixed supply of
productive services.

The inclusion of the possibility (3)—i.e., of employing more
services from an unchanged number of men—is theoretically fas-
cinating (because of the interesting complications in assuming
various shapes of labor supply curves), but it has less practical sig-
nificance than the possibility (4)—i.e., employing formerly unem-


===== PAGE 44 =====

620 FRITZ MACHLUP

ployed factors of production. The absorption into the productive
process of unemployed factors has been regarded by many writers
as the realistic case where the irrelevance of an investment-period
theory for the analysis of credit expansion would become apparent.
And this for two reasons: first, it has been said that if sufficient
productive capacity is available no services have to be withdrawn
from consumption-goods industries as, owing to an expansion of
credit, the production of investment-goods industries is extended;
second, the amount of service input being increased, the increased
investment need not involve any lengthening of the investment
period. There is more investment, but there are also more pro-
ductive services, hence no change in proportion and no “distor-
tion of the investment structure” will ensue.

The correctness of this view and the applicability of an invest-
ment-cycle theory are not incompatible. One has to bear in mind
that, according to the capital theory under consideration, an in-
creased amount of services can be employed, without additional
voluntary abstinence from consumption, only if the investment
period is shortened” and the earnings per unit of service are re-
duced. By employing more services, through credit expansion, at
unreduced rates of earnings without voluntary saving, the invest-
ment period, absolutely unchanged, becomes longer than it would
be without credit expansion.

The case of credit expansion with full employment and that
without full employment are, therefore, equivalent in respect to
the fact that the investment period becomes longer than it would

79 Less capital per head is available. See above, p. 590, and also p. 603. It need
not be explained that the return to labor is smaller the shorter the period of invest-
ment. In terms of cost theory one would speak of decreasing returns of increased
application of labor, or of increasing marginal cost of production. A qualification
should be made with reference to the unused capacity argument. If unused capacity
of equipment of all sorts is available in all stages of production, and if materials,
goods in process, and finished goods of all kinds are on stock in sufficient quantities,
synchronized production (and consumption) could set in at a stroke all at once. No
additional waiting for intermediate or finished goods would be needed. (Why there
should be unemployment, under such circumstances, is, however, hardly intelli-

gible.)


===== PAGE 45 =====

THE “PERIOD OF PRODUCTION” 621

be if the supply of money capital were confined to its ‘natural’?
sources: namely, replacement funds and voluntary savings.

28. Investment period and inevitableness of the breakdown.—Ab-
solute or relative overlengthening of the investment period is the
probable result of expansion of producers’ credit. The length is
excessive in the sense that without any “outside” influences (with-
out change in data) internal forces will lead to a reshortening of
the investment period, sooner or later. The “internal forces’ con-
sist in a divergence between the individual time preferences (as
expressed in the proportions in which the money incomes are
saved and invested or spent for consumers’ goods) and the time
structure of production. The divergence is concealed so long as
the credit system continues to subsidize successfully the employ-
ment of productive services for work of great consumption dis-
tance—work which would not be done if the voluntary distribu-
tion of the existing money funds alone were decisive for produc-
tion.

A discussion of the details of business-cycle theory and mone-
tary management cannot be embarked upon here. But all that is
needed here is to point to the essential connection between invest-
ment period and credit cycle. While some writers find the causes
of the breakdown of prosperity in the imperfections of monetary
institutions or in lack of skill of monetary authorities, the invest-
ment-cycle theory suggests that there is no money system and no
credit management conceivable which could permanently main-
tain a production structure (i.e., time distribution of productive
services) which does not correspond to the structure of expendi-
tures (i.e., time distribution of purchasing power). Though mone-
tary forces help to bring about “prosperity,” and lead to the ex-
cessive length of the investment period, monetary forces do not
seem to be capable of maintaining it permanently.

Why the boom is doomed is explained by the investment-cycle

8 No value judgment is implied in using the word “natural” as distinguished
from “artificial.” The “natural” sources refer to the concept of the “natural” rate
of interest, which is simply that rate which equalizes the demand for money capital
to the amount supplied out of replacement funds and voluntary savings.


===== PAGE 46 =====

622 FRITZ MACHLUP

theory. This theory does not make the pretense of being the only
explanation of all cycles and crises that have ever occurred, nor
does it pretend that it states unconditional necessities. The gist
of the theory is that there is a high chance that increased bank
credit,’ unless offset by hoarding or similar phenomena, makes
for investment periods different from what they would be other-
wise; that there is, furthermore, almost no chance that the dis-
tribution of expenditures in consecutive phases will adapt itself to
the investment structure;®? and that, therefore, the investment
structure will be forced to undergo a process of readjustment.

29. Liquidation for consumption versus liquidation for liquidity.
—The “liquidation mania” commonly observed during crises and
depression, the urgent attempts to exchange capital goods and
property rights for liquid funds, are not motivated by the wish to
use the liquid funds for buying consumption goods but by the
wish to keep the liquid funds on reserve. These rather obvious
facts are taken by some writers for the whole substance behind the
notion of depression. The struggle for liquidity alone is held to
account for the disinvestment after prosperity. Hence, the break-
down is considered either as the result of psychic (or psychopath-
ic) phenomena or as the result of the failure of the credit system
to supply the money necessary to satisfy this increased demand
for liquidity.

Nobody can seriously question the fact or the significance of
this “liquidation for liquidity.” But this clearly visible fact should
not lead us to forget that it is as a rule preceded and, in a sense,
““caused” by “liquidation for consumption.” This “liquidation
for consumption’ is the discontinuance of activities of great con-
sumption distance forced upon, or suggested to, the entrepreneurs
through lack of capital disposal or through cost increases which
are but the expression of the comparatively more urgent demand

81 Increased utilization of bank balances may take the place of increased amounts
of bank balances. (Increase in V’ rather than increase in M’, i.e., “dishoarding.”)

82 The adaptation of the expenditure distribution to an excessive period of in-
vestment would consist in providing increased savings to take the place of infusions
of new bank credit.


===== PAGE 47 =====

THE “PERIOD OF PRODUCTION” 623

for consumption goods. That the use of productive services ceases
to be profitable for a certain kind of work is the expression of the
buyers’ preference for other products, e.g., for “present” con-
sumption goods. The first attempts at disinvesting—i.e., the fail-
ure to reinvest in work of great consumption distance—may
easily, and actually did mostly, start a general flight from invest-
ment into greater holdings of cash. Although in reality the two
types of liquidation are ordinarily combined, they should be dis-
tinguished in theoretical analysis. Complete absence of the li-
quidity rush would not imply elimination of the business cycle or
of depression; only the depth of the trough, the amplitude of the
cycle, would be much smaller.

30. Deflation and shortening of the investment period.—How
much of the difficulties and frictions would be taken away from
society if deflationary liquidation did not accompany and aggra-
vate consumptive liquidation cannot be known. But that the
shortening of the investment period is connected with difficulties
and frictions of its own seems to be certain. They may result from
several circumstances: (1) Many capital goods are specific, i.e.,
not capable of being used for other purposes than those they were
originally planned for; major losses follow then from the change in
production structure. (2) Capital values in general—i.e., antici-
pated values of future income—are reduced by higher rates of
capitalization; the owners of capital goods and property rights
experience, therefore, serious losses. (3) Specific capital goods
serviceable as ‘‘complementary’’ equipment for those lines of pro-
duction which would correspond to the consumers’ demand are
probably not ready; employment in these lines is, therefore, small-
er than it could be otherwise. (4) Marginal-value productivity of
labor in shorter investment periods is lower; wage rates are, there-
fore, depressed. (5) Under inflexible wage rates unemployment
ensues from the decreased demand prices for labor.

That capital loss, wage cuts, and unemployment can be ex-
plained apart from any deflation does not mean that in reality
they are not entangled with deflation. We said before, and repeat
here, that deflation, though sometimes ‘‘caused’ by the process of


===== PAGE 48 =====

624 FRITZ MACHLUP

shortening the investment period, aggravates this process con-
siderably?

31. Immobility of factors and rigidity of costs.— The immobility
and specificity of capital goods—strengthening the resistance to
adaptation and adjustment—has something to do with the dura-
bility of instruments and thereby with the investment period.
But, as Professor Knight points out correctly, there is no definite
relation between durability and the lack of adaptability.®4 And
there is certainly not the least economic connection between in-
vestment period and the main force of resistance to adjustment:
cost rigidity. Hence, with immobility of factors (human and pro-
duced) and rigidity of costs (labor and material) as the essential
factors of the trouble, the period of investment would have no
place in the analysis of depression.

If capital goods were mobile and perfectly adaptable there
would be no unused plant capacity. And ‘if labor were mobile
and wages flexible, no fixity in the capital structure would give
rise to unemployment.” And yet, to explain unemployment
(through wage stickiness) is one thing; to explain the business
cycle is another. If wages were perfectly flexible there would per-
haps be no sharp fluctuations in employment, but there would be
fluctuations in wage rates instead. Why the perfectly flexible
wage rates go up through some length of time and why they fall
most heavily through some succeeding length of time—in other
words, why there is boom and depression, and whether and why
depression must follow the boom—would still remain open ques-
tions. For a theory of the business cycle, with no unemployment
but wage fluctuations instead, the concept of the investment peri-
od would be as valuable a tool as it is for the theory of the business
cycle with little wage flexibility and unemployment.

83 To state this is by no means to advocate reflationary measures. Although they
may offset hoarding and mitigate the difficulties connected with it, they may also

interfere with, and actually delay, the necessary adaptations in the production
structure.

84 Economic Journal, March, 1935, p. 93; Canadian Journal, February, 1933,
p. 21.
85s Knight, Economic Journal, March, 1935, p. 94.