CLASS S T R U G G L E

AND T H E D I S T R I B U T I O N OF
N A T I 0 NAL I N C O M E *

1. Until fairly recently it was generally accepted that profits would

decline pro tanto if wages were raised. Even though in the analysis
of other phenomena SAY’S Law was not adhered to, at least not
strictly, in this case the preservation of purchasing power was not
put to doubt. The analysis of an increase or reduction in wage rates
dealt with the physical consequences of this absolute shift from profit
to wages or vice versa. I n the case of rising wage rates the reconstruc-
tion of capital equipment in line with the higher spending on wage
goods and lower outlays on investment and capitalists’ consumption
was emphasized. At the same time, therc was assumed a tendency
to higher unemployment as a result of substitution of capital for
labour that has become more expensive.
Although quite a number of economists would argue in this
fashion even today, the fallacy of this approach is widely recognized.
I t may be countered by various economists in somewhat different
ways. My counterargument is based on the following assumptions:
There is a closed economic system and a proportional rise of all wage
rates. I n a certain short period the annual wage bill increases as a
result of the rise of wage rates by A W. The workers spend all their
income immediately. The investments and capitalists’ consumption
are determined prior to the short period considered and are therefore
not affected by the wage rise in this period.
If we subdivide the economy into three sectors producing (I) in-
vestment goods, (11) consumption goods for capitalists, and (111)
wage goods-including the respective intermediate products-, it
follows that employment in the first two sectors is not affected by the
rise in wages. Thus, denoting the wage bills in these sectors measured
in ‘old’ wage rates by W1 and Wz and the fraction by which wages
are raised by tc, we obtain for the increment in aggregate wages in
* Managing Editors’ Comment: As far as wc arr informed, this is Professor
KALECKI’S last article. He sent us this paper one month before his death on
April 17, 1970. Some minor stylistic improvements were made by the editors.
1
 MICHAL KALECKI

+
sectors I and II K ( WI W2).The profits in these two sectors decline
pro tanto provided the prices of their products have not risen, which
is assumed in the argument based on ‘preservation of purchasing
power’.
The position in sector 111, however, is quite different because the
workers immediately spend their additional proceeds. The in-
crements of the wage bills in sectors I and 11, equal to a( W I + W2),
must unavoidably cause profits in sector I11 to rise pro tanto, as the
profits of this sector consist of the proceeds from the sale of the goods
not consumed by the workers in sector 111. Thus, the increments in
the wage bills of these two sectors, K ( W1+ W Z )will
, cause an equal
rise of profits in sector 111. This may come about either through the
rise in output in that sector or through a rise of the prices of its
products. As a result total profits remain unaltered, the loss of sectors
+
I and I1 of a(WI W2) being counterbalanced by an equal gain
of sector 111. I t follows that no absolute shift from profits to wages
occurs and the argument based on SAY’S law would thus prove
fallacious-at least in the short run.
This last qualification is essential. For it may be argued that the
decline in the volume of investment and capitalists’ consumption as
a result of the wage increase would admittedly not come about
immediately but would be delayed-possibly to the next short
period. This would be true if the capitalists at least decided to cut
their investments and consumption immediately after having agreed
to wage increases. But even this is unlikely, for their decisions are
based on current experience. This assumption according to the above
will show that no loss in total profits occurs in the short period
following the wage rise and that thus there will be no reason for a
cut in investments and capitalists’ consumption in the following
period. If a decision for such a cut is not taken right away on the
basis of the bare fact of the wage rise, it will not be taken at all. As
a result, profits will not decrease in the next period either. The
argument of a shift from profits to wages as a result of a wage rise
based on SAY’Slaw is therefore Fallacious even if we consider all the
ramifications of this event.
The same obviously applies to a wage cut: no increase in profits
will occur either in the short period following it or subsequently.
2. So far we have assumed that prices of investmcnt goods and
2
 CLASS S T R U G G L E , D I S T R I B U T I O N O F N A T I O N A L I N C O M E

consumption goods for capitalists remain unchanged when wages

rise. This was in line with the theory of the shift from profits to
wages to the extent of the wage rise. (The preceding section in a
sense amounted to the reductio ad absurdurn of this theory.) However,
stable prices are unlikely in our case; they will rather rise under the
impact of the wage increase (perhaps not in the short period following
the wage rise but later on). But to discuss this question as well as
other repercussions of a wage rise (or a wage cut) we want to know
more about price formation in the system considered.
We shall first abstract from all semi-monopolistic and mono-
polistic factors, i. e. we shall assume the so-called ‘perfect compe-
tition’. This is, however, a most unrealistic assumption not only for
the present phase of capitalism but even for the so-called competitive
capitalist economy of past centuries; surely this competition was
always very imperfect in general. Perfect competition-when its real
nature, that of a handy model, is forgotten-becomes a dangerous
myth.
As follows from the argument in the preceding section, the volume
of capitalists’ investments and consumption is maintained in the
short period following the wage rise and consequently thereafter.
O n the assumption of perfect competition and of supply curves
sloping upwards a t some point, the rise in wage rates must cause a
proportional rise in prices at given levels of respective outputs (per-
haps not in the first short period but later on). As a result profits
in sectors I and I1 will rise in the same proportion as wages, i.e.
+
1 M times.
It is now easy to prove that the volume of production and con-
sumption of wage goods remains also unaltered. Indeed, in such a
case profits in sector 111,similarly to those in the other two sectors,
+
increase in proportion with the wage rise, i. e. 1 M times. As men-
tioned in section 1, profits in sector I11 are equal to the proceeds
from sales of wage goods to the workers of the sectors I and I1 and
therefore they must increase in the same proportion as the wages in
+
these sectors, i. e. 1 M times. Should the volume of production and
consumption of wage goods increase or decline, that could not be
the case.
Hence, with perfect competition the volumes of production in
all three sectors remain unchanged while their values increase
3
 MICHAL KALECKI

each by 1+ a. The total wage bill and total profits increase therefore
in the same proportion and the distribution of national income
remains unaltered.
We have shown the fallacy of the theory based on SAY’Slaw, which
maintains that wage movements have a direct and full impact upon
the distribution of national income, and arrive at the opposite state-
ment that they have no influence upon this distribution a t all. But
such a conclusion is based on the untenable assumption of perfect
competition. Only by dropping it and penetrating the world of
imperfect competition and oligopolies are we able to draw any
reasonable conclusion on the impact of bargaining for wages on the
distribution of income1.
3. I n fact, a major part of the economy may be plausibly rep-
resented by a model very different from perfect competition. Each
firm in an industry sets the price p of its product by ‘marking up’
its direct cost u consisting of the average costs of wages plus raw
materials in order to cover overheads and achieve profits. But this
mark-up is dependent on ‘competition’, i.e. on the relation of the
ensuing price p to the weighed average price of this product for the
industry as a whole. O r :

where f is an increasing fuiiction: the lower p is in relation to p, the

higher the mark-up will be fixed. From formula ( 1) we obtain :

I t should be noted that the function f may be different for various

sirms in an industry. It will reflect semi-monopolistic influences
resulting from imperfect competition or oligopoly. The more inten-
five these factors are, the higher is f(pip) corresponding to a given

1. We abstract here from the influence of the increase in the price levcl upon
the rate of interest by assuming tacitly that the supply of money is elastic. Other-
wise the higher demand for money would have increased the rate of interest which
would affect adversely investments and consequently profits. Such a n effect seems
unlikely to be of any greater importance, especially because the changes in the
bank rate are reflected in the long-term interest rate on a much reduced scale.

4
 CLASS S T R U G G L E , D I S T R I B U T I O N O F N A T I O N A L I N C O M E

relation pip. Prices p will be in general different for various firms

because of the differences in direct costs u and because of those in
the function$
The price system is determined. Indeed, with s firms in an industry
+
there will be s 1 price values to be determined, i. e. p,, pz, ...,ps,p,
and
-
as many equations : s equations of type (2) and one determining
p in terms ofp1,pz, . .. ,ps.
If all direct costs u increase with given functions f by 1 tc, prices +
p1, pz, . . . ,Is will do so as well. Such a solution indeed satisfies (1)
+
and (2) because u by assumption increases by 1 ct and p / p remains
unaltered. If, however, the direct cost uk increases for one firm only
(again with given functions f),it is easy to see that p k increases in
a lesser proportion becausep will not rise in the same proportion as uk.
4. Since prices p for a product are generally not equal, the above
applies strictly to imperfect competition only or to differential oli-
gopoly but not to non-differential oligopoly or monopoly. However,
apart from basic raw materials produced frequently in conditions
approaching perfect competition, most of the products do have differ-
ential prices. (Let us not forget that absolutely identical products
with the same transport costs but different periods of delivery may
have different prices.)
I t seems therefore a fairly good approximation to an actual econ-
omy if we assume the model described above, while the sector of
basic raw materials conforms in its price formation to the model of
perfect competition.
Let us now imagine that in a closed system of this type wage rates
in all industries increase in the same proportion, by 1 a. It follows+
+
easily that all prices will also incrcase by 1 M provided that thefunc-
tions f in industries to which they are relevant remain unchanged. If these
conditions are fulfilled, it follows that we shall reach the same con-
clusions as for a perfectly competitive economy described in sec-
tion 2 : a general increase in money wage rates in a closed economy
will not change the distribution of national income. The same would
apply to the case of decreasing money wage rates. However, we shall
argue that the functionsfdo depend on the activityofthe trade unions.
5. High mark-ups in existence will encourage strong trade unions
to bargain for higher wages as they know that firms can afford to
pay them. If their demands are met but the functions f are not
5
 MICHAL KALECKI

changed, prices will also increase. This would lead to a new round
of demands for higher wages and the process would go on with price
levels constantly rising. But certainly an industry will not like such
a process making its products more and more expensive and thus
less competitive with products of other industries2. T o sum up, the
powcr of the trade unions restrains the mark-ups, i. e. it causes the
valuesf(p/p) to be lower than they would be without their existence.
The power of the trade unions manifests itself in the scale of wage
rises demanded and achieved. If an increase in bargaining capacity
is demonstrated by spectacular achievements, there is a downward
shift in functions f ( p / p ) and the mark-ups decline. A redistribution
of national income from profits to wages will take place. But this
redistribution is much smaller than it could be if priccs were stable.
The rise in wages is to a great extent shifted to consumers. ‘Normal’
wage increases will usually leave the functions f unaffected while
otherwise mark-ups may tend to get higher because of the rise in the
productivity of labour.
6. Let us imagine that a spectacular wage rise somewhat depresses
the mark-ups so that a redistribution of national income from profits
to wages occurs. From section 1 it follows that profits in sector 111
will increase in the same proportion as wage rates. However, as there
is a redistribution of income from profits to wages as a result of the
reduction of mark-ups in sector 111, the wage bill in sector I11
increases more than wage rates, i.e. there will be a rise in employ-
ment and output. Output and employment will stay unaltered in
sectors I and 11. I n other words, the volume of investments and
capitalists’ consumption remains constant, while the workers’ con-
sumption will increase. Such an expansion of total output and em-
ployment will be feasible because our model of semi-monopolistic
price fixing, as developed in section 3, presupposes the existence of
excess capacities.
The money value of the wage bill will clearly increase more than
wage rates. However, total profits will increase less than wage rates:
Indeed, profits in sector I11 increase proportionately to wage rates,
employment in sectors I and I1 remaining unaltered, but profits in
2. Despite the fact that for the sake ofsimplicity wc assumrd that all wage rates
are raises simultaneously in the same proportion, we consider realistically that
bargaining is proceeding by industries.

6
 CLASS S T R U G G L E , D I S T R I B U T I O N O F N A T I O N A L I N C O M E

the latter two sectors increase less than the wage rates as a result
of the decline of mark-ups3.
If the power of the trade unions declined, the process described
would be reversed. Employment and output in sectors I and I1
would remain unchanged, but in sector I11 they would decline. O r
the volume of investment and capitalists’ consumption would remain
unchanged while the workers’ consumption would fall. Total output
and employment would thus decline. The value of the wage bill
would fall more and the value of profits would fall less than wage
rates4. Since the decline in mark-ups tends to increase aggregate
output, this would cause a rise in the prices of basic raw materials,
subject to conditions of perfect competition, in relation to wages. As
a result the increase in output and employment would be somewhat
restrained. I n the same fashion this factor would somewhat restrain
the decline of output and employment caused by the rise of the mark-
ups. I t follows from the above that a wage rise showing a n increase
in the power of the trade unions leads-contrary to the precepts of
classical economics-to an increase in employment. Conversely, a fall
in wages showing a weakening in their bargaining power leads to
a decline in employment. The weakness of trade unions in a depres-
sion manifested in permitting wage cuts contributes to the deepening
of unemployment rather than to relieving it.
7. I t follows from the above that the class struggle as reflected in
trade union bargaining may affect the distribution of national in-
come but in a much more sophisticated fashion than expressed by
the crude doctrine: when wages are raised, profits fall pro tanto. This
doctrine proves to be entirely wrong. The shifts that occur are
(a) connected with widespread imperfect Competition a i d oligopoly
in the capitalist system, and (b) contained in fairly narrow limits.
However, the day-by-day bargaining process is an important co-
determinant of the distribution of national income.

3. This, however, is subject to the following qualification. As a result of the

increase of total output there will be an increase in prices of basic raw materials,
inter uliu of those used alike by sector I or I1 and sector 111. This, although not
very likely, may offset the influence of the decline in mark-ups in the sectors I
and I1 upon the distribution of income between profits and wages. In any case,
however, total profits will rise less than the total wage bill.
4. Subject to a qualification analogous to that stated in the preceding footnote.

7
 MICHAL KALECKI

I t should be noted that it is possible to devise other forms of class

struggle than wage bargaining, which would affect the distribution
of national income in a more direct way. Actions may e. g. be taken
for keeping down the cost of living. This may be achieved by price
controls, which, however, may prove difficult to administer. But
there exists an alternative: the subsidizing of prices of wage goods
financed by a direct taxation of profits. Such an operation, by the
way, will not affect aggregate net profits. The argument here is the
same as in section 1 in the case of a wage increase. The same applies
to the effect of price controls. If such measures cannot be carried
out in parliament by political parties associated with trade unions,
the power of the trade unions may be used to mobilise supporting
strike movements. The classical day-by-day bargaining for wages is
not the only way of influencing the distribution of national income
to the advantage of the workers.
8. The redistribution of income from profits to wages as described
in the last two sections is feasible only if excess capacity exists. Other-
wise it is impossible to increase wages in relation to the prices o
wage goods because prices are determined by demand, and the func-
tions f become defunct. We return then to the position described in
section 2 where the wage rise could not effect a redistribution of
income.
Under these circumstances price control of wage goods will lead
to scarcities of goods and haphazard distribution. Subsidizing prices
of wage goods, too, can reduce prices only in the long run by stimu-
lating investment in wage good industries.
I t should be noted, however, that even contemporary capitalism,
where deep depressions are avoided as a result of Government inter-
vention, is generally still fairly remote from such a state of full
utilization of resources. This is best shown by the fact that prices of
finished goods are fixed on cost basis rather than determined by
demand.

Warsaw MICHAL
KALECKI
 CLASS S T R U G G L E , D I S T R I B U T I O N O F N A T I O N A L I N C O M E

SUMMARY

The author believes that the class struggle manifested by trade union bargaining
may lead to a redistribution of national income from profits to wages. With a
three-sector-model (investment goods, consumption goods for capitalists, wage
goods) he shows that under perfect competition the share of wages certainly will
not change when wage rates alter. However, an oligopolistic market structure,
excess capacities, and mark-up pricing are the basis for a successful wage bargain.
The more powerful the trade unions are, the more they will be able to restrain the
mark-ups and thereby to increase the share of wages in national income. More-
over, a risc of wage rates will-contrary to the results of classical theory-lead to
a rise of cmployrnent.

ZUSAMMENFASSUNG

Drr Autor ist der Ansicht, dass unter bestimmten Voraussetzungen der Lohn-
kampf der Gewerkschaften zu einer Umverteilung des Volkseinkommens fuhren
kann. Mit seinem Drei-Sektoren-Modell (Investitionsguter, Konsumguter fur
Kapitalisten, Konsumguter fur Lohnempfanger) zeigt er zwar, dass bei vollkom-
mener Konkurrenz die Lohnquote durch Veranderungen des Lohnsatzes nicht
beeinflusst werden kann. Oligopolistische Marktstruktur, Leerkapazitaten und
Preisbildung durch Zuschlag auf die Durchschnittskosten bilden jedoch die Vor-
aussetzungen fur einen erfolgreichen Lohnkampf. Je machtiger die Gewerkschaf-
ten, desto eher wird es ihnen gelingen, den Kostenzuschlag zu vermindern und
damit eine Umverteilung des Volkseinkommens zu ihren Gunsten zu erwirken.
Gleichzeitig fuhrt eine Erhohung des Lohnsatzes - im Gegensatz zur klassischen
‘Theorie - nicht zu eincr Senkung, sondern zu riner Erhohung des Beschafti-
gungsgrah.

L’auteur du prdscnt article pense que, dans des hypothtses bien dkfinics, la bataille
pour les salaires, menbe par les syndicats, peut entrainer une redistribution du
rcvenu national. En s’appuyant sur un modtle B trois secteurs - biens d’investisse-
ment, biens dc consommation pour capitaiistes, bien de consommation pour sala-
rids - l’auteur montre en effet que, dans une situation de concurrence pure et par-
faite, la part dcs salaires ne peut Stre influcncde par des variations du taux de
salaire. Structures de march6 oligopolistiques, capacitks de surproduction, aug-
mentations des prix constituent cependant les hypothtses prkalables B une bataille
pour les salaires victorieuse. Plus les syndicats sont puissants, plus ils seront capa-
bles de freiner les augmentations et, par lB-meme, d’augmenter la part des salaires
dans le revenu national. De plus, un accroissement du taux de salaire - contraire-
ment aux conclusions de la thCorie classique - entrainera une augmentation de
l’emploi.