15. Labor’s Share

IN CHAPTER 3 I referred to the widespread conviction that, through the “bitter struggles” of more than two centuries, unions managed to acquire a larger share of aggregate income for the workers. This quite general belief that history records a redistribution in favor of the working class as a whole, through victorious strike-threat warfare (whether defensive or aggressive), is a formidable, illusory stereotype. In this and the next chapter we shall be concerned with changes in the proportion of income accruing to labor.

Some economists have wondered why anybody could be interested in such a question. If it should happen that, in the absence of the strike-threat, aggregate nonlabor income would rise proportionately to labor income, while aggregate income was greatly increased and more equally distributed (and I have tried to show that a more equal distribution would be the consequence of a strike-free economy), why should a possibly smaller percentage of wages to aggregate income be a matter of concern? The actual flow of wages would be much greater; and given the greater real earnings, why should not the individuals making up “labor,” through that very process, have become owners of a greater proportion of assets, receiving an additional income in that manner? Why then worry about labor’s share?

The answer is that the strike threat system is applauded, or respected, or tolerated because it is believed to bring about a more equal or a more “just” distribution of income between labor and capital. From the standpoint of the unions and their apologists, the “prime objective” of the system is, according to Golden and Ruttenberg, to exercise “constant pressure for a larger share of the nation’s annual income,”¹ or, as Mathew Woll puts it, “organized labor’s obligation to its members is to pursue wage increases until . . . the national income is distributed equitably and stays so distributed.”²

Moreover, the growth in labor’s absolute real income during the period over which unionism was growing has left an impression that a rising relative share of income was gained for tabor. It is important that this stereotype be disturbed. I have found from personal experience that academicians of distinction—open-minded in the spheres in which they are expert—believe, almost without exception, that the higher standards of living (real wage rates and working conditions) which the working classes enjoy today, in comparison with, say, 1880, or 1900, or 1930, were won against reluctant “employers” after years of strikes and strife. They are nearly all convinced that a large gain was gradually achieved at the expense of “property.” The illusion has had a profound influence on the attitudes of sociologists and scientists with a smattering of economics, to say nothing of judges, schoolteachers, clergy, journalists, and union officials. The fact is that, in a country like the United States, real income per head has been growing so rapidly that, even if labor’s proportion had fallen materially this century, the average real wage rate could still have risen prodigiously.

The causes of the phenomenal rise in working-class affluence since the industrial revolution have been summarized in Chapter 3. It will suffice here to stress that income redistribution has been a factor in raising the material condition of the poorer workers only insofar as a reduction of inequalities in the natural scarcity values of the workers’ powers (their efforts and skills) has been brought about; and this has been a process in which union pressures have played a negative role. But the results contribute to a general illusion that a rise in labor’s proportion of aggregate income has been achieved.

There are grounds for holding that the most important consequences of (a) the better use of the qualities of the people (via upward mobility), and (b) the better use of a growing stock of assets (which we have seen are the true origin of growing working-class well-being) must have been to cause the remuneration of both income categories to rise more or less in the same proportion even if not always in exactly the same proportion (because the one can hardly be expected perfectly to balance the other). I have, I maintain, already explained why the unions’ conventional influence on wage rates must necessarily fail to bring about any lasting net redistribution in the “private” sector if they rely on the strike-threat weapon. But I must be in a position to answer critics who may say, “Well, that’s all theory! What are the facts? The henchmen of the strike-threat system hold that it can change the proportions in a socially defensible way. Can you prove them to be wrong? Can you show that income has not been redistributed from the rich to the poor, or from investors to workers, during the period in which the strike threat has been increasingly used?” I certainly can show this; but no one can prove from the findings of statistical inquiries alone that labor’s share would not have been still smaller in the absence of union pressures!

In his great study of human action,³ Von Mises has explained why the type of reasoning on which the foregoing analysis has relied does not require the type of empirical confirmation to which I am about to appeal. If statistical evidence seems not to confirm the conclusions, that is presumably proof that the data, or the treatment of the data, must in some way be defective. But the economists’ logic has never been infallible: still less have empirical or political assumptions at the stage at which inferences are made (to put it mildly) always been unchallengeable. Hence in the following chapter (Chapter 16) I shall submit my general conclusions to the test of broad consistency with statistical studies.

The matter of labor’s share in income is among the topics to which several distinguished statistical economists have directed ingenious studies. Through analysis of income, production and pricing data, they have tried to discover what change there has been in this share over various time periods. I shall in due course quote their verbal summarizations of what I regard as their most relevant conclusions. Unfortunately, the difficulties met with in such investigations are formidable, and in spite of all the ingenuity the investigators have exercised, the significance of their findings is perforce sometimes rather limited.

My reference to these difficulties implies no criticisms of those statistical economists whose inferences, including policy inferences, are conditioned by efforts at reconciliation with the conceptual clarity of “orthodox” analysis. This is exactly what Schultze and Weiner are attempting in their scholarly introductory essay to The Behavior of Income Shares, an important symposium of statistical and econometric studies of the determinants of the relative shares of labor and capital.⁴ And it is what Tibor Scitovsky also is concerned with, in respect of the whole field of “empirical” contributions on this subject, in the learned article with which that symposium opens.⁵

During the last quarter of a century there has been a remarkable interest shown by economists in the visible income redistribution consequences of strike-threat policy—indeed, we have witnessed what Gregg Lewis has called “an outpouring of empirical research on unions and wage differentials.”⁶ This “outpouring” reflects, indirectly, widespread misgivings about the labor movement. Nevertheless the investigations have been carried out with a scrupulous objectivity and the researchers have mostly allowed their figures to speak for themselves. But partisans of the unions seems to be uneasy about it. Dunlop almost suggests that the statistical economists are showing a morbid or unhealthy interest in the unions as determinants of wage rates and labor cost! He says, “The persistent concern with the impact of unionism as an institution perhaps reflects a preoccupation with defending or condemning the institution as a whole.” But, insists Dunlop, “the institution is here and is likely to stay,”⁷ as though to suggest that whether the determination of labor’s remuneration under duress is beneficient, innocuous or pernicious is hardly a matter which should worry the serious scholar. It will happen anyway!

At the same time, Dunlop thinks it worthwhile to refer to some of the limitations of empirical investigations, as though they must necessarily invalidate such studies. But only if, which I do not think is true, the investigators are blind to the pitfalls do his warnings have applicability. In some fields, admittedly, there is no way of bringing the categories and concepts which economists have found useful into close correspondence with the cruder categories and concepts with which the statistician must be content. Yet it is possible to attempt to determine how the relative shares in aggregate income of “labor,” variously defined, on the one side and all other factors lumped together on the other appear to have changed over periods in which the strike-threat influence has been growing. And it is possible to compare movements of wage rates in the unionized and nonunion sectors. For such studies to be done satisfactorily, however, there are conceptual problems which must be faced.

What I have just termed “all other factors” are usually described as “capital” or “property.” A better term for this would be “investors,” for all assets have been invested in. We sometimes talk, not misleadingly, of “human capital” and of the “investment” in it of time, effort and resources. But the yield to capital in such a form is conventionally treated as the remuneration of labor. Fortunately, this usage does not greatly hinder our present task.⁸

At one time, in discussions of income distribution, economists classified agents of production as “land, labor, capital and enterprise.” Today, “land” is classified as “capital” and the remuneration of “enterprise” is classed as part of the yield to “capital.” Such a classification is justified simply because a certain division of labor in respect of the risk-taking or entrepreneurial function realistically links enterprise with property ownership. But properly visualized, entrepreneurial remuneration is neither remuneration of capital nor remuneration of labor. It may accrue to either (positively as profits or negatively as losses) according to which assumes the risk of the entrepreneurial decision turning out to have been wise or lucky (or unwise or unlucky). Those earlier economists showed real insight who classified entrepreneurs hip separately, as “enterprise,” an agent of production distinct from the other three agents—land, labor, and capital. For the yield to “enterprise,” namely, “profit,” is payment by results for the most important function that is performed on behalf of the community—prediction and responsible action to determine the composition of the stock of assets and/or valuable skills. Through this function, entrepreneurs determine the form of economic activity. In the vast majority of cases, however, it is the owners of capital who assume most of the risk which the entrepreneurial factor involves; in empirical studies there is no satisfactory way of isolating either the interest element or the labor element in the owners’ income; and hence for practical reasons it is usually appropriate to classify profits (positive or negative) together with interest as nonlabor income. Under this classification, the part of income with which “labor’s share” is usually compared is the whole of the remainder, in other words, that earned by “property” or “capital” and “enterprise.” It is all nonlabor income (as I have said) “lumped together,” and consists of interest (which includes rent) plus profits and minus losses. Nevertheless, we must remind ourselves that there are no legal or institutional barriers to contractual arrangements under which labor becomes the residual claimant and takes the profit.⁹ A rather different complication is that the worker, in seeking the most remunerative fields for acquiring skills or for selling his services, is acting as an entrepreneur; and the yield to his shrewdness or enterprise, although received as “wages,” is in principle a form of “profit.” This is a matter of small importance in the present context, although it will be of great importance in another (see pp. 223-224).

In some studies, “labor’s share” covers the earnings of artisans and laborers only. In others, it covers all “employee compensation,” that is, all forms of contractual remuneration for services rendered by people, and it is this connotation that has been most often used. “Salaries,” which are conceptually distinguishable from “wages” only in the most nebulous way, and statistically distinguishable only by arbitrary definition,¹⁰ have usually been reckoned as part of the remuneration of labor in the broadest sense, and regarded as pan of “labor’s share.” But this share covers, as we have seen, a yield to investment in human capital.

And we have another conceptual difficulty to face. To the extent to which investors who have not anticipated or adequately evaluated strike-threat consequences are exploited, it is, as we have seen (pp. 135-137), through the seizure of some part of investors’ property and the income stream which that property yields. There is clearly no way in which empirical studies can isolate this element and count it as part of the yield to property.¹¹

Before referring to the findings which meticulous statistical investigations have reached about changes in the income distribution pattern, as reflected in the relative shares of “labor” and “investors,” it is important to consider, in the background of the analyses presented in the earlier chapters, what we should have expected to find has been happening to these two great statistically defined magnitudes since the beginnings of this century.

Firstly, we have certain “observable circumstances” (including the un-predicted use of the strike threat) which would have led us to expect that, in the United States, since the 1880s, and especially since 1935, labor’s share would have been increasing (circumstances discussed on pp. 226, et. seq.).

Secondly, we have certain other “observable circumstances” of an offsetting nature, which we should have expected more or less to have neutralized those tendencies.

I hope to show that these discernible determinants of relative shares are probably responsible for the rough constancy for labor’s share that empirical studies have established, despite continued efforts to raise that share, and in spite of illusory evidences that those efforts may have had some success.

All economizing displacements, considered in isolation, tend to reduce the percentage share of those whose assets or labor provide the services economized. For example, every time unskilled or unprivileged workers manage to sneak through the fences and work their way into a skilled or privileged occupation, their entry tends, ceteris paribus, to reduce the relative share of labor in any activity affected. But we cannot validly draw inferences after merely considering the consequences “in isolation.” For firstly, the cheapening of any kind of labor tends to attract in additional assets to cooperate with it, and this may be expected, over the whole economy, to work toward restoring the former proportions. That is, the more easily capital can be attracted by relatively low tabor costs in any activity, the weaker will be the tendency for the better use of tabor to reduce labor’s proportion. And secondly, the more easily labor can move into fields in which the tools of production are increasing in efficiency, the weaker will be the tendency for technological progress to raise labor’s proportion.

The point can be explained in more general terms. In every productive activity the workers are, so to speak, demanding the services of complementary assets and of risk-taking, while investors (as owners and entrepreneurs) are demanding the services of the workers Both invest inputs (services embodied into outputs), the value of the resulting outputs being shared according to a contract (influenced by free or restrained market forces). In each case, any economization of labor expresses an increased demand for the services of the complementary assets used, just as any economization of such assets expresses an increased demand for the services of the labor used. Each economy tends therefore to raise the opposite party’s relative share.

Exactly the same principle is relevant with diseconomies (such as an earthquake, or the raising of a wage rate by a strike threat, or a collusively enforced output restraint). The reduced contribution of a factor will mean a rise in its percentage share in the value of the output. The general principle can be stated as follows.

Given unchanged knowledge, the application of further increments of any factor of production to a fixed “amount” of any other factor of production, or to a fixed “amount” of any unchanging combination of other production factors will, in the absence of any economies of scale, yield less than proportionate average returns. Thus, a reduction in the number of man-hours worked,¹²the volume of complementary assets being assumed unchanged in magnitude and composition, must mean a rise in labor’s share. A growth in the stock of assets in an industry, with the number of man-hours unresponsive, must have a similar effect. To put it differently, if the quantity of services rendered by any factor of production rises or falls relatively to an assumed fixed quantity of services which owners of complementary factors of production find it profitable to retain or bring into a particular productive activity, the proportion of the value of the product which accrues to those who provide relatively larger or smaller inputs will fall or rise respectively.¹³

Such are the basic factors determining the division of aggregate income between investors (property) and labor. We are concerned, that is, with changes in the relative supplies of productive services rendered by assets and those rendered by labor; and various changes in these magnitdes (and in other factors) can be observed to have been occurring over history—changes which, superficially considered at any rate, could have been expected to be affecting the proportions. The following changes are relevant:

(1) in the size, race and sex distribution of the population;

(2) in society’s valuation of leisure, including changes in (a) the ages at which different classes of juveniles are allowed to compete with their elders and enter various remunerated employments, and (b) the ages at which people are encouraged or forced to retire from remunerated activity (generally or in specified sorts of occupation);

(3) the extent to which certain kinds of work are regarded as properly reserved for a particular class, sex or race;

(4) in the extent to which the fixing of wage rates under duress occurs (that is, the extent of deliberate contrivance of labor scarcity in that manner);

(5) in saving preference schedules (that is, in people’s desire to provide for the future);

(6) in the rate at which (given saving preference) the process of economizing displacement assists the net accumulation of assets (through raising prospective yields to investment) and thereby increases investors’ bidding (as intermediaries) for labor, which in turn multiplies the yield to labor;

(7) in the extent to which the economizing-displacement process tends to be neutral, or to cheapen either labor or assets relatively to one another;

(8) in consumer preferences as between outputs of labor-intensive and capital-intensive production.

It should be noticed that (1), (2), (3) and (4) (empirically representable by the number of man-hours actually worked and the distribution of those hours over tasks of different degrees of productivity) represent labor’s bidding for the services of the tools (that is, of the assets) which multiply labor’s yield, while (5) and (6) represent investors’ bidding for labor, a bidding which actually expresses the multiplication just referred to. Changes (7) and (8) bring in a different kind of influence.

Now superficially this notion of labor’s bidding for the services of complementary factors may appear to involve a paradox and scope for confusion in other ways. The composition of the great complex of demands for the productive services of both men and assets is determined, as we have seen, by people in their consumer role. That is, consumers—the ultimate employers—demand the joint product. Hence when investors (as intermediaries) demand services for incorporation into assets (work in progress), theirs is a derived demand. That demand is expressed through their initiative when they are residual claimants, as they virtually always are. But because labor hardly ever takes the residual share, this does not mean that the workers are not entrepreneurially involved. The wage system relieves the workers from the greater part of the risk burden; but they are still buying the services of assets, and profiting or losing from the wisdom or unwisdom of their policy in so doing. The wage terms on which they work in any activity are a major determinant of the rate of flow of services into replacement and growth of the assets they use. Ceteris paribus the cheaper their services the greater will be the investment their “bidding” calls forth.

It is through interpretation of this empirically observable complementary relationship between assets and labor that we can perceive, I suggest, the main reason for the constancy of proportions. Assets (of all degrees of physical or economic perishability) are the tools of labor, and increases in their quantity and quality (their cheapening) mean increases in labor’s earnings. An overwhelming proportion of the value of assets is in constant process of consumption and replacement at various rates, and the costs of replacement as well as of growth are borne jointly (out of the realized value of the output) by the workers and the owners of the assets. Hence, a general cheapening of labor will mean the cheapening of assets;¹⁴ the magnitudes (a) aggregate real value of the services of assets and (b) aggregate real value of the wages flow, are not wholly independent of one another; and some tendency to stability in the relative value of their shares is therefore to be expected. It is true that a general cheapening of labor would otherwise tend to reduce labor’s relative share (while raising its absolute share); but because cheap labor means cheap tools, there will be a countervailing tendency to maintain the relative shares which other factors have determined.

What may at first seem to be a separate reason for the hardly changing proportions (in spite of strenuous efforts to transfer income from the one sector to the other) is put forward in a rigorously argued contribution by Lebergott. On the realistic assumption that wage rates in the industries which produce capital goods will in practice change more or less in the same proportion as wage rates in the industries which use capital goods, he infers that this explains why in practice the price of capital service tends to “bear a long-term proportionality to that of labor.”¹⁵ This long-term proportionality, he shows, “derives from the fact that the supply forces working to fix the price of capital are dominantly wage costs in the capital-producing industries and those that supply them. In the competitive market these wage costs parallel wage cost changes in capital-using industries because wage changes for identical occupations must bear a parity with one another in all employing industries.”¹⁶ G. Garvy restates Lebergott’s conclusions as follows: “In ultimate analysis, the cost of capital goods can be reduced in essence to wage costs incurred in previous periods. Therefore, in the long run, the price of capital goods must bear constant long-term relation to that of labor.”¹⁷ As I see the issue here, it is that rising labor costs of supplying and replacing relatively long-life assets (which are assumed to rise more or less proportionally to rising labor costs imposed on industry generally) affect labor adversely in the industries which must use such assets and meet the interest, depreciation charges and upkeep costs. Lebergott is, I think, envisaging labor in what I have termed its “entrepreneurial role.” He sees it as I do, as continuously demanding the services of capital equipment—demanding with the value of the services the workers contribute as inputs. Duress-imposed real labor costs tend therefore to recoil to labor’s disadvantage and bring about no gain to the workers as a whole.

The lesson is perhaps clearest if we think of labor costs in the iron and steel, the construction and the machinery-manufacturing industries. Rising output prices in these activities adversely affect yields to labor of all kinds. But the rising costs of supply of assets which labor has to bear in the circumstances imagined are simply the consequence of a smaller real value of assets being retained, replaced or added to in the industries which manufacture fixed assets. This is the reaction to the strike-threat system explained in Chapter 10. It may be that Lebergott’s way of stating the principle assists our understanding of the simple reality that “assets are labor’s tools,” while duress-imposed costs of manufacturing the tools are against labor’s advantage.

The services of assets and those provided by labor can, then, be envisaged as demanding one another, the relative values being, in each case, market-determined. The fact that the market is seldom “free” but constrained by various contrived scarcities and plenitudes creates no tendency for the relative values of the two broad kinds of services as a whole to change in the long run. especially when the argument of Chapter 7 and the dynamic factors referred to above are given due weight.¹⁸ For instance, if we imagine a reduction of the “labor supply” in existing employments through widespread collusive action, we can hardly usefully assume the survival of an unchanged stock of assets. A reduced real value of complementary assets will be profitably replaceable or accumulable in each activity affected by contrived labor scarcity. Not only will the costs of replacing or adding to the stock of assets as presently composed be higher than previously, but the real value of assets which compete with labor (that is, noncomplementary) will tend to expand and ceteris paribus reduce labor’s share. In addition, the assets structure (the composition of replacement and growth) will be molded to suit less productive employments; for some labor will be diverted to less productive (that is, less income-generating) employments (including, perhaps, unemployment). Demands for labor and demands for services of assets will tend to contract in correlation.

A reduction of man-hours supplied in relation to the capital stock can be relied upon, then, to bring about a rise in labor’s share only while the composition of the stock of assets can be assumed still to be in process of becoming fully adjusted. If we assume that no such adjustment occurs, and for the short run an assumption of that kind may at first seem to be reasonable, union policies which have the effect of reducing labor inputs must indeed tend to raise labor’s percentage share (to labor’s absolute disadvantage). But over a period as long as a decade, the effect of contrived labor scarcity can certainly be realistically expected to bring about compensating changes in production functions.¹⁹

We are in practice concerned, however, not only with the substitution of factors in the production of a defined bundle of outputs, but also with the substitution of one kind of output for another, that is, for a changing composition of outputs in general.²⁰ Every time the price of one input rises, the other complementary inputs become less profitable; and as time passes, the providers of the inputs will tend to divert them to different kinds of output. Any increased share acquired at first by the contriver of a scarcity is likely to be gradually whittled away through reactions from the great society outside (that is, external to the firm or industry in which the contrived scarcity is imposed). Ceteris paribus, the real value of complementary factors used in the activity will gradually diminish.

The fact that capital can sometimes substitute for labor is of course a vital consideration. Such a substitution is one of the most conspicuous ways in which assets take on less exploitable forms. It tends to reduce labor’s share. Admittedly investment in labor-economizing plant may sometimes itself be highly exploitable, although vigilant entrepreneurs will avoid the trap. For this reason, however, other reactions upon the composition of the stock of capital resources may be more important. But in my judgment, labor-economizing developments must have had a formidable influence in reducing labor’s share.

To recapitulate. To the extent to which assets are initially mainly complementary in their relation to labor, the burdening of activities which produce such assets with strike-enhanced labor costs will be a self-defeating way of trying to augment labor’s relative share. Not only must labor incur higher costs for the assets they have to use (which will ultimately offset, at least partially, any immediate transfer at property’s expense), but even more important, it will induce a change in assets structure through which investors will be able to avoid continued exploitation. In other words, while strike-threat pressures statically considered must tend to raise labor’s proportion (as distinct from its absolute earnings), dynamically considered the process can be expected to have a neutral effect upon relative shares. Labor’s tools will assume a less wage-multiplying form while the workers will be driven to cooperate with different assets, in employments which will, on the average, be less remunerative.

Assuming now that the community’s savings-preference schedule remains unchanged, any former rate of growth in the real value of the aggregate capital stock must fall (perhaps become negative) following a wave of duress-imposed labor costs. For prospective yields must decline and the value of profitable investments in inputs must shrink. This also will tend to offset any tendency for labor’s percentage to rise. On the other hand, rising thrift, especially if accompanied by autonomous capital-economizing developments, in accelerating the rate of additions to the stock of complementary assets, will tend to raise labor’s share.

My own interpretation of the data of United States experience is that if (and I must place strong emphasis on the if) the rate of growth in the stock of assets which has been witnessed this century had taken a form determined under free market incentives (the proportion of complementary to competing assets being an important consideration), yet reached the magnitude in relation to man-hours to which reference was made above, there would have been a very large increase indeed in labor’s share for this reason alone.

In fact, I suggest, we have found the reverse. Technological and managerial ingenuity has, for instance, been canalized toward labor-substituting, automating forms. Growing recourse to the strike threat has biased the form of economizing displacement. Undoubtedly, this is one of the most important factors explaining why labor’s relative share has been kept down. Certainly we should have found some tendency toward growing automation and the assembly-line and mass-production methods of the western world, even had history been different and a relatively strike-free economy been experienced. But far more effort would then have gone into developing machines which increase demands for labor than into those which have the opposite effect.

I have throughout been occasionally reiterating the deliberately challenging yet accurately descriptive term, “wage-multiplying assets.” The more cheaply complementary assets may be replaced or accumulated in any activity, the greater will be the relative yield to effort and skill in that activity and the greater its share. But even such assets as compete with labor are still wage-multiplying in the sense that economies achieved in their production tend to multiply the absolute wages flow, although they must (subject to the important qualifications we have just noticed) tend to reduce the relative value of wages as a component of total income. That is, while the cheapening of labor-economizing assets is not wage-multiplying for the workers in any occupation directly affected, it does have this effect for workers in all the other occupations which stand in a noncompeting relationship. It tends equally, of course, to multiply yields to capital in noncompeting fields.

Some economists hold that a rise in the demand for leisure (not uninfluenced by strike-threat influences) over the past century must have been tending to raise labor’s percentage of the value of the product—a reduced aggregate physical product. These economists feel that (1) the subsidization of early retirement in various ways; (2) prolonged schooling for such young people as do not benefit therefrom, and who are often deprived in some measure thereby from training in wage-multiplying skills;²¹ and (3) a reduction in the hours of labor in privileged occupations, must have been contributing to an augmented relative share for labor (although at the expense of labor’s absolute share). Certainly a general enforcement or subvention of a preference for leisure (against the alternatives of greater material well-being and security) combined with all the other relevant factors, including the age distribution of the population, can be observed to have been causing the number of man-hours worked (in terms of “efficiency units”) to increase less rapidly than the stock of assets.²²

In considering this issue, we must remember that the aggregate number of man-hours supplied is influenced not only by voluntary or duress-imposed demand for leisure but by every contrived labor scarcity. Greater leisure may well be one of the products purchased by the private beneficiaries of the scarcity contrivance. But more important, once man-hours are measured in “equal productivity units,” it becomes clear that their number is reduced whenever the price of labor is raised by force. Even if there had been no change in conventional working hours in any occupation, the raising of wage rates above the natural scarcity level must have meant a withdrawal of labor supply in terms of “equal productivity units”; for ceteris paribus a larger proportion of the workers must have been confined to work of lower productivity. Hence there is no special problem due to the reduction of labor supply via abnormally early retirement, prolonged useless schooling or shortened hours of labor.

Let us now consider the United States where it has been estimated that the volume of physical capital has increased this century more than three times as much as the aggregate number of man-hours has increased.²³ One would have expected (relying solely on static assumptions) such a relative growth in the productive power of assets to have raised labor’s share in a marked degree. But this expectation is always subject to the crucial qualification enunciated on p. 22 J. namely “the volume of complementary assets being assumed unchanged in magnitude and composition,” It is just this assumption which we cannot make. The real value and the form of assets are in process of constant adaptation to the price and type of labor available. That seems to be why no increase in labor’s share is discernible.

In continually insisting that what happens in the short run is no necessary indication of what will happen in the long run, I have so far only briefly referred to the business cycle, over the period of which changes in relative shares certainly do occur. During the downturn, costs tend to be more rigid than prices, and fixed assets become underutilized, leading to a rise in labor’s share; while during the upturn, labor’s share declines. Thus, when a reduction in man-hours worked occurs during a developing recession (mainly through unemployment due to the maintenance of wage rates), there is a discernible cyclical redistribution in favor of labor’s share. Labor’s percentage is increased to labor’s disadvantage.²⁴

During wars in which public opinion strongly supports the war effort, it is possible to increase the relative flow of labor inputs because the workers agree, so to speak, in some measure to sacrifice leisure for the common good, or in return for high “overtime” payments. Further, in these circumstances, the unions often permit (voluntarily or otherwise) a measure of “dilution” (unprivileged workers doing privileged work). But recognition of the anticyclical movement of labor’s share must not lead the reader to the conclusion that during a period of growing prosperity, it is essential for labor’s percentage to decline. Demands for labor are derived from prospective yields from investment in labor’s inputs. Hence the optimism and feeling of entrepreneurial security which would exist under boom conditions if strike action were ruled out would almost certainly cause most demands for labor to increase, and this would tend to preserve labor’s proportion. Entrepreneurs would strive to increase their activity ahead of their competitors. Nevertheless, in considering a long period with an inflationary trend but covering several cycles of recession and boom about the trend, we should expect a tendency for labor’s share to decline unless we must give much weight to what has been called the “ratchet” effect (see pp. 229-230);. But two contrary factors are likely to offset this tendency: (a) price controls intended immediately to reduce predicted residual claims and (b) the fact that wages and salaries in the armed forces are paid out of income transferred via taxation, while the wages and salaries of those employed in producing for the war effort are also remunerated from this source. Empirical studies do, indeed, seem to confirm such conclusions (see Chapter 16).

I have been discussing the tendencies which (allowance made for cyclical disturbance) appear to be stabilizing labor’s relative share, despite factors which one might at first think would be inclined to change it. There are indeed six possible reasons why, if we did not take account of the reactions which I have just been discussing, we should have expected²⁵ labor’s share to have increased during the last century and to be clearly discernible in empirical studies.

1. The first has already been dealt with, namely, the enormous growth in the stock of assets in relation to labor supply (in the sense of the number of man-hours of standard productivity).

2. The second possible reason is that this century unions have been growing in aggregate membership and becoming expert in strike-threat techniques, while investors have, one would have thought, been adjusting their expectations to the use of these techniques only gradually. Hence, substantial temporary exploitation of investors could have occurred. As I have insisted, in the short run, during a period in which entrepreneurial anticipations are gradually being molded by experience of the emerging aggressive unionism, some redistribution at investors’ expense may well happen, everything depending upon how wisely entrepreneurial predictions of the future private use of coercive power cause early changes in the composition of the community’s stock of assets.²⁶

3. A yet stronger reason why we should expect an increase in “labor’s” proportion of aggregate income is the remarkable increase in government employment, and resort to various kinds of “welfare handouts” which form part of “labor’s share.” Public services tend to be labor intensive, while public servants are remunerated via direct transfers, that is, via taxation.²⁷ A very large part of the community’s income has come to be redistributed in this way, and much of it is in the form of what really amounts to relief work. For this part of the redistribution, however, the strike-threat is not the cause except in so far as government employees are permitted to use this method in order to force redistribution through taxation. The limitations to strike-threat power which were discussed in Chapters 1 and 10 do not, as we saw, apply to government employments. There is no clear limit to the “soaking” of taxpayers by governments when the bulk of the visible taxes are paid by a political minority, except the ability of entrepreneurs to export capital; and even that door may be obstructed or completely closed by exchange control (including “interest equalization”). Taxation by local governments (states, provinces, counties, municipalities) may of course divert some investment to other areas within a country. But because the actual use of strikes in the government sector has been increasing in recent years, in what I feel has been a largely unpredicted manner and degree, one would have expected “labor’s share” to have grown.²⁸ There is some evidence that this may be the explanation of a very small rise in that share which certain empirical investigations disclose.

4. A further reason why we should expect to find evidence of a larger share accruing to labor is the growing governmental exploitation of the provident otherwise than by overt taxation. Politicians often find it expedient to mulct those rentiers who have not correctly predicted the speed and duration of inflation, while the debasement of a currency implies a redistribution which seems likely on balance to cause the share of property to decline. This is because wage contracts are short term, and labor shortage at initial wage rates caused by inflation pulls up the money price of labor with a relatively short time lag, while bonds are long-term contracts. Of course, the losses borne by interest receivers are offset in some measure by the gains to residual claimants. But through the presence of the strike threat, what would otherwise be “restorative” entrepreneurial yields due to inflation (see p. 229), may be successively seizable through the exercise of union power. When the inflationary process is recognized as having been built into the economy, there may be no countervailing gains against the losses imposed on interest receivers. In these circumstances, however, the rentiers’ expectations must be brought into the reckoning. As soon as they come to anticipate inflation, market interest rates rise sufficiently to prevent their further exploitation. Nevertheless, over the years for which most of these statistical comparisons have been made, one would have expected the exploitation of the rentier class to have been reflected in the figures, and therefore to have caused some increase in labor’s share.²⁹

5. Another reason why one would have expected, ceteris paribus, to find evidence of a rising trend in labor’s relative share, is sometimes described as “intersector shifts” of labor. It has reference to transfers of consumer preference toward the outputs of more labor-intensive types of occupations.³⁰ For instance, labor’s share in agriculture is lower than it is in nonfarm occupations as a whole and there has been a shift—indeed a substantial shift—away from agriculture in most Western countries.³¹ Moreover, in an increasingly affluent society, we can observe a rising preference for “life-enrichment” activities, toward the service occupations generally, toward the constructional industries, and toward commerce (as distinct from physical manufacture). These occupations tend to be more labor intensive than the average, and in them the proportion of wages in relation to the value of output is well above the average. Growing mechanization may have reduced the significance of this trend, but it has by no means offset it.

6. The final reason for expecting to find that labor’s share has been increasing is what has been called the “ratchet” effect. It is concerned with the phenomenon noticed on page 227, namely, the expansion of labor’s percentage during periods of depression and unemployment. The “ratchet” effect is operative when that gain is not wholly offset during the shrinkage of the percentage (normally to be expected) as the flow of wages recovers and fuller employment is achieved. Because wage rates are more rigid downward than they are upward, this could mean a gradual increase in labor’s proportion (at the expense, of course, of a reduced rate of recovery in the aggregate wages flow) through the seizure of what I have called the “restorative” element in the residual yield. (See pages 229 and 253-255.) If such a redistribution does tend to occur, however, it must be classed as a special case of the situation in which investors have failed to perceive their exploitability. And this is relevant to another circumstance, also connected with cyclical effects, namely, eventual recourse to “disinflationary” policies which are not accompanied by government action to protect the wages flow³².” In such circumstances, investors may be exploitable by continued labor union pressures to raise wage rates, and labor’s share will tend to increase as long as entrepreneurial expectations underrate the probability of exploitation in this form.

In addition to the above-mentioned reasons for expecting evidence of actual transfers from yields to property in favor of yields to labor, there are five reasons for expecting the statistics to reflect merely apparent (that is, illusory) redistributions in the same direction,

1. Income statistics seldom include nonpecuniary yields to property, yet this kind of income has obviously been growing in importance as the general level of material well-being has been rising. An increasing proportion of real income has tended to be received in the form that Marshall called “gratifications” from investment in consumer capital goods. In the United States, this must be true in some degree of all income groups except perhaps the poorest; but the higher the income, the greater the importance which must be given to the yield from consumer durables. The propensity of well-to-do persons with expanding incomes to hold a large proportion of their assets in “luxury” property forms like mansions, country estates, mountain lodges, seaside dwellings, yachts, and so forth, as well as costly jewelry, antique furniture, valuable pictures, and the like, is an obvious manifestation of the phenomenon. No doubt the incentive is largely demand for status symbols, which the political trends of the last half century have encouraged. But any such “conspicuous investment” must be regarded as yielding a nonpecuniary income that normally exceeds the interest which, capitalized, represents the pecuniary value of the assets.³³ The tendency is likely to have been reinforced by the growing recourse to the strike threat. It is one of the ways in which attempts to avoid exploitation can cause assets to take on a less wage-multiplying form (see pp. 143-144). Moreover, for reasons connected with the growing tax burden over the present century, other yields to property have tended more and more to take the form of nonpecuniary benefits; and such benefits are seldom reckoned as part of “national income.” The remuneration of property in this “invisible” form must therefore tend to raise the apparent proportion of labor’s income to all income.

An important special case is the “gratifications” in which we mostly share in some degree from collectively owned assets the services of which are “free,” or sold to us at prices which are insufficient to cover interest and depreciation. The growing proportion of labor’s remuneration as a result of government employment is not balanced in income statistics by the yield to the growing stock of property which is owned by the state (that is, supposedly by the people). The illusory element from this cause can be regarded as unimportant if collective ownership of assets is judged to result in an egalitarian distribution of the services rendered by those assets. But as with most consumer durables, they tend to be in the nature of luxuries—things which one would think satisfy the priorities of the rich in income use rather than those of the poor. It is doubtful whether free preferences expressed in the market would lead to the people who form the category “labor” voluntarily paying for such things if their cost in interest and upkeep had to be met out of, say, handouts of equivalent value in the form of a “negative income tax.” On balance, then, it seems that income data must reflect an illusory redistribution in labor’s favor from this cause.³⁴

An offsetting factor may be the growing proportion of fringe benefits in relation to pure wage income for which the statisticians may not have been able to make full allowance. This is a rather recent phenomenon, and I do not think it can have had more than a negligible influence over the greater part of the periods which empirical studies of income distribution have covered. Such influence as it has had, however, may have led to an understatement of labor’s proportion.

2. A further consideration which cannot be ignored is the effect of corporation taxes on the form of income declaration. Whenever the managers themselves are the owners of a large proportion of the capital, minimization of the tax burden has an important consequence. If the income of such managers is classed as a yield to labor, their total tax is less than if it is classed as a yield to capital. Hence as taxation of corporations has increased, the remuneration of managers has been less likely to take the form of stock options or stock allocations; for the identical remuneration can be offered, with identical incentives, at a smaller sacrifice of income to the tax collectors. It becomes profitable to remunerate executives through commissions, bonuses or salary increases. Again, what would have been declared as “profit” in earlier periods has been increasingly declared as salaries—that is, yield to “labor.” And resort to the device of undistributed corporate profits as a reaction to inflation, while actually increasing, is a quite likely additional cause of an illusory rise in labor’s share.

3. A possibility which should be mentioned is that underdeclaration of profits due to attempts to minimize the rising taxation burden may create an unduly low declaration of income from property. Such underdeclaration could be inadvertent. Growing investment in research, including market research, pilot schemes, models and prototypes, or even advertising expenditures ought at times to be treated as investments, although actually treated as current costs.

4. Another very important reason why we should expect recorded statistics to show an illusory growth in labor’s share is the fact that the proportion of income accruing to small businesses has declined gradually in the course of technological progress. Services which were formerly remunerated by income returned as “profit” have been increasingly remunerated in the form of income returned as “wages.” For instance, last century we found a much larger proportion than we do today of people who were serving at shop counters, yet owned their own businesses. They gained a modest return which was described as “profit.” Today the same class of people will be wage-earning shop assistants, but enjoying considerably larger real incomes as “wages” than their forebears earned as “profits.” Again, housewives and other members of a family who perform services in the home will be remunerated by a share of, say, the husbands’ wage income. But their remuneration is not reported as a contribution to income. Last century and early this century far more women fell into this class than today, when their counterparts are largely performing paid work. Such women have been released from providing services for the family directly, and provide instead services for the community generally. Families are being increasingly served through the preparation and cooking of food in factories, while manufacturing provides, replaces, repairs and services washing machines, vacuum cleaners, dishwashers, central heating apparatus, mechanical can openers, etc. The labor employed in this field is remunerated. When it had been performed domestically no remuneration was recorded. The proportion of privately owned assets directed by the “self-employed” or in the form of unincorporated undertakings (industrial, agricultural and commercial) in the United States has declined substantially since the beginning of the century (from 41 percent in 1900 to 23 percent in 1956).³⁵ The consequences upon the description of incomes must have been far from negligible. Again, self-employment has always been important in agriculture (although agriculture has tended more and more to become a corporation-directed activity). Much of the real earnings of farm operators (including their families) does not get reported as income.³⁶ Hence because the relative share of agricultural income in national income has been declining, we have a particularly important example of the growing tendency for income of a kind which was earlier returned as profit to be returned as wages.

5. A factor of some importance is that in an area as vast as that of the United States, considered over a period in which average incomes and standards of living in different districts differed widely at the outset but which, with the passage of time, were gradually becoming less unequal, the shift of workers from areas of relatively low productivity and low earnings to areas of relatively high productivity and high earnings, must have meant that the money earnings of those who moved increased by a greater proportion than their real earnings increased. The relatively lowly paid regions are typically low cost-of-living areas. For the same kind of reasons the movement of labor from agriculture to higher paid industrial or urban pursuits must have meant that any increase in money earnings which the statistics record was substantially greater than any increase in real earnings. This is because of the relative cheapness of living on farms and the perquisites available in that occupation.³⁷

Through the operation of these eleven factors, six concerned with actual redistributions and five with illusory, we should have expected to find, from studies of income statistics, that during the last century at least, an indisputable and substantial redistribution of recorded income from investors to labor could be discerned. In fact we find nothing of the kind. Although some such studies have seemed at first to show that small transfers of aggregate income in the expected direction have indeed been experienced, further investigation into the validity of the methods or significance of the data appear nearly always to have established that the proportions in which aggregate income is divided between the two broad groups “property” and “labor,” far from having changed discernibly, have remained disconcertingly constant!

Theoretical economists and statisticians have long been fascinated with the ultimate constancy that is discovered, which Schumpeter (reviewing the course of empirical inquiry in this field) termed “a remarkable fact”³⁸ and other economists have described by words like “miracle,” “mystery,” “amazing,” “Medusa-like,” and so forth. Yet the fixity of the proportions which has prompted such descriptions might be, Solow has suggested, merely “an optical illusion” or a “mirage,”³⁹ or, as Samuelson has suggested, “an interesting coincidence.”⁴⁰ If my argument on pages 220 et seq. is valid, however, the constancy is neither a “miracle” nor a “mirage.” There is some justification, I think, for Samuelson’s phrase, “an interesting coincidence,” although there is more to it than mere coincidence.

The thesis presented above certainly does suggest that a fairly constant ratio will be established between the value of the efforts of men and the value of the services of the tools men make. My reasoning in these passages is simply a special exposition of the “classical” marginal productivity theory of the valuation of the services of men and of assets. On this issue, Bronfenbrenner has referred to what he calls the “considerable constancy” implied by “conventional marginal distribution theory” . . . “provided only that the elasticity of substitution between capital and labor is not well below unity,”⁴¹ and he is here, I think, touching on what is fundamental. My own argument in Chapter 10 can be interpreted as an explanation of why elasticities of substitution will be less than unity in the short period, but move in the direction of unity in the long period (see also pp. 233 et seq.). Reder has concluded (in a review of thought and empirical inquiry on the subject) that “the mechanisms of product and factor substitution have been such that whatever pressure unions have been able to bring to bear upon wage rates has been offset in so far as any effect upon relative shares is concerned.”⁴² If this explanation is acceptable, an important causal factor in the constancy cannot be appropriately described as merely coincidental. At the same time, we cannot hold that the elasticities and substitutions which appear to have been bringing about the balance over the periods investigated must necessarily be operative under all conceivable institutions or policies. One “control” imposed (by unions or government) on the valuation of the services of men and of assets appears always to set more or less countervailing reactions going; but the measure of the permanence of proportions which results is a chance phenomenon. There is no reason to expect an exact restoration of disturbed ratios. Indeed, during a time span in which strike-threat pressure in each period, as we have seen, has been greater than entrepreneurs had expected in the previous period, some evidence of redistribution in favor of labor would be a reasonable expectation. Hence if my suggested explanation of why no such redistribution has indeed occurred is sound (namely, that labor—in an entrepreneurial capacity—had to incur higher costs for the services of assets, while the stock of assets assumed a less wage-multiplying but less exploitable form), the rough balance in proportions brought about does seem to justify the word “coincidence.”

In a strike-free regime, the reduced risk involved in capital-economizing investments could have increased entrepreneurial demands for labor from the assets side by more than it increased demands for the services of assets from the side of the providers of labor. If this judgment is valid, we must accept that it is a coincidence—although an explicable coincidence—that, during the nineteenth and twentieth centuries, the strike-threat system has everywhere been permitted to prevent labor’s share from rising (as well as reducing the absolute wages flow) and has throughout just happened to maintain that share constant.

In my opinion, there can be few surviving “optical illusions” due to the pitfalls of statistical investigation. It is true that the researchers have been unable to discern the weights of the many heterogeneous factors which are unchallengeably the determinants of income distribution. And there are certainly illusory data. But these tend almost entirely to show a spurious rise in labor’s share (at least if the argument under headings (1) to (5) (pp. 230-233 is acceptable). It is indeed only after adjustments have been made to offset such “optical illusions” as can be identified that investigators have discovered the hardly changing long-term ratios. But, although it must certainly be stressed that “an appearance of inevitableness” in relative shares does not mean “an inevitableness” unqualified, the fact of the constancy found, under different definitions of the aggregates compared, cannot be described as a “mirage.” It is an undeniable reality.

APPENDIX TO CHAPTER 15
Note on a Recent Contribution

Just after this book was prepared for publication, I noticed an important article by Professors H. G. Johnson and P. Mieszkowski (hereinafter referred to as “the authors”),⁴³ which relates to the topics here discussed, and reaches conclusions which are similar, although arrived at by very different methods.

It begins with a rigorous geometric examination of a model in which labor is regarded as homogeneous and two commodities only are produced, one with capital-intensive methods and the other with labor-intensive methods. It is pointed out that, through the influence of unionization, “the allocation of factors among industries” and “the allocation of production and consumption among industries” will be rendered inefficient.⁴⁴Making abstraction of such repercussions it is argued, however, that if unionization occurs only in the capital-intensive sector, “unionized labor must gain, while nonunionized labor must also gain”; for “unionization is in effect a tax on the labor of the unionized industry, and therefore has the effect of shifting demand away from that industry.” Such a tax results in “a fall in the demand for and price of the services of capital and an increase in the demand for labor, from which both sections of labor may gain.”⁴⁵ For similar reasons, it is argued, if unionization occurs only in the relatively labor-intensive industries, the capitalists will gain and labor will lose in both sectors.

It is true that the end-products of capital-intensive activities can be rendered less preferred at the prices which result from a labor tax on them and hence the end-products of labor-intensive activities (on which no labor tax is levied) rendered more preferred. But this merely illustrates a particular case with very unrealistic assumptions under which labor’s relative share may be raised (the issue to which this chapter has been devoted).

If the labor tax is levied in capital-intensive activities only, and the reactions on resource allocation (which the authors stress on p. 543) are brought into the reckoning, then given the losses caused thereby in aggregate real income, any conceivable increase in labor’s absolute income is extremely difficult to imagine.

If we now drop the two commodities assumption and suppose that a very important source of demands for labor-intensive activities is the incomes of those who own assets—the capitalists (a highly realistic assumption), we can see that a labor tax on capital-intensive activities may cause labor’s relative share to fall as well as its absolute share. Capital-intensive activities contribute to the real wage flow just as labor-intensive activities do. And their outputs contribute to the source of demands for all noncompeting productive services, those of labor as well as those of assets.

In turning from geometric analysis to algebraic, with arithmetic illustrations, and using data based on empirical evidence, the authors draw attention explicitly to the limited practical relevance of their conclusions. Their methods here, they warn, “do not allow for the long-run effects of unions on the distribution of income and the real wage.” Their findings, they say, “err to the extent that the formation of unions changes the level of investment,”⁴⁶ Even so, they find that, given other plausible assumptions, any gains of unionized labor are largely or wholly at the expense of nonunionized labor.⁴⁷

Later, dropping the assumptions that labor is homogeneous and all labor in capital-intensive activities is unionized, the authors substitute the assumption that all blue-collar workers are unionized and all white-collar workers are nonunion. This change in assumptions is shown not to affect the conclusion that “unionized labor gains primarily at the expense of nonunion labor.” It is shown further that the tendency for a “tax on labor” (duress-imposed wage gates) to benefit unionized workers at the expense of nonunionized, may be limited by “a substitution of capital for labor in the union sector.”⁴⁸

I mentioned above the authors’ warning that “the formation of unions changes the level of investment.” In the long run, they say, “the level of capital formation will fall.”⁴⁹ Now it is true that labor costs raised by the strike threat must reduce prospective yields to investment in general. Hence, ceteris paribus, given any propensity to save, the rate of interest must fall. Exactly how it will affect the magnitude of achieved savings or dissavings (that is, the rate of net accumulation or accumulation of output-yielding assets) will depend upon a variety of considerations. But what is most important is not the magnitude of the savings flow (the “level of investment”) but its composition, it is the form assumed by the stock of assets in their replacement or net accumulation which matters most—the extent to which assets acquire more “wage-multi8plying” attributes, irrespective of whether new methods induced are capital-intensive or labor-intensive (or, alternatively, capital-economizing or labor-economizing).

If the penalization of investment by duress-imposed wage rates in an industry causes (through its bearing on end-product prices) demand for the output of that industry to fall, it will become unprofitable to replace fully (or maintain a previous rate of growth in) the stock of complementary assets. The workers remaining in the industry may well gain, but marginal workers will be laid-off and potential recruits to the industry will be forbidden access to the bargaining table. In my judgment, however, the vital consequence will be the reduction of the real value of labor’s earnings in noncompeting activities and the reduction of yields to previously invested capital in noncompeting activities, because the offer of outputs from the protected field for the non-competing inputs will contract. Labor in general must certainly suffer detriment. The authors cover this reality only through their warning (mentioned above) that their methods do not allow for reactions upon real wage rates.

Nevertheless, the authors reach the final conclusion that, for “a partially unionized economy, . . . most, if not all of the gains of union labor are made at the expense of nonunionized workers, and not at the expense of earnings of capital.”⁵⁰ Such “unionization of labor does not in fact benefit labor at the expense of capital.”⁵¹

Turning then to a wholly unionized economy, the authors find that even if all labor were unionized and the “bargaining power” of the unions happened to be equally spread (presumably meaning by this use of the term “bargaining power” that the threat to disrupt by strikes happened to reduce prospective yields to investment everywhere and in all activities by an equal proportion),⁵² unless the unions could somehow offset any monopsonistic purchase of labor, or unless the unions could thereby seize some share of the monopolistic gains of complementary parties, “the distribution of income (would) be essentially the same as the distribution in an economy in which unions (did) not exist.”⁵³ That is, the proportional shares of capital and labor in the reduced aggregate income would be more or less unaffected.

NOTES

¹ See S. Golden and H. J. Ruttenberg, The Dynamics of Industrial Democracy (New York; Harper, 1942), p. 151.

² M. Woll, Labor, Industry and the Government, p. 146, quoted in P. Sultan, Labor Economics (New York: Henry Holt, 1957), p. 383.

³ Ludwig von Mises, Human Action (New Haven: Yale University Press, 1949), Chapter 2, passim.

⁴ C. L. Schultze and L. Weiner, eds., The Behavior of Income Shares, National Bureau of Economic Research (Princeton: Princeton University Press, 1964).

⁵ Tibor Scitovsky, “A Survey of Some Theories of Income Distribution,” in ibid.

⁶ H. Gregg Lewis, “The Effects of Unions on Industrial Wage Differentials,” Aspects of Labor Economics: A Conference of the Universities—National Bureau Committee for Economic Research, A Report of the National Bureau of Economic Research (Princeton: Princeton University Press, 1962).

⁷ John T. Dunlop, “Comment,” Ibid., p. 343.

⁸ R. M. Solow holds, however, that if we could identify the part of the wages flow which is a yield to investment in human capital, it would materially affect our conclusions. “A Skeptical Note on the Constancy of Relative Shares,” American Economic Review, 1958, p. 630.

⁹ In the actual world, of course, we do not find such arrangements, for reasons which have already been discussed (see Chapter 6). Hence no serious difficulties arise for this reason.

¹⁰ Attempts to distinguish “wages” from “salaries” involve so much arbitrariness, it seems to me, that any significance of inferences relevant to the apparent constancy of “factor shares” is destroyed.

¹¹ ‘Nor can effective allowance be made for changes in the proportion of income from property generally accruing to the group of persons envisaged as “labor,” or (as we noticed above) the proportion that properly should be isolated as yield to labor in income returned as “profits.”

¹² Theoretically, it is essential to assume man-hours of unchanged average efficiency or productivity.

¹³ While this is true in each particular activity, the principle cannot be safely transferred to the whole economy without encountering the lurking danger of the fallacy of composition.

¹⁴ That is, in reducing the cost of services flowing into the replacement or net accumulation of assets, a cheapening of labor is enabling a given sacrifice of consumption by savers (i.e., a given flow of savings) to go further.

¹⁵ S. Lebergott, in The Behavior of Income Shares, pp. 53-100, especially p. 66.

¹⁶ Ibid., p. 66.

¹⁷ G. Garvy, in The Behavior of Income Shares, p. 96,

¹⁸ We saw in Chapter 7 that a contrived scarcity in one activity, in increasing labor’s share in that activity, creates an “incidental plenitude” of labor for the benefit of investors in other activities, so that labor’s share elsewhere will tend to be reduced and to restore therefore aggregate proportions.

¹⁹ W. J. Fellner has drawn attention to these important considerations, American Economic Review, Proceedings, 1953, p. 491.

²⁰ Under the relative austerity imposed on consumers by the restriction of inputs in general, different preferences will be expressed, and both the composition of the stock of assets and the form of skill acquisition will adjust to the expectations created by the changing demand situation. Because this kind of reaction can be reasonably expected in practice, the likelihood of any increase in labor’s percentage for such a reason is reduced.

²¹ Solow thinks that one assumption here is factually wrong. He suggests that improving investment in human capital through “education, training, public health, etc . . .” may have been sufficient to cause “the measurement in man hours” to underestimate “the rate at which the labor force grows as properly measured in efficiency units.” (Solow, op. cit., p. 630.)

²² The tendency for such a situation to raise labor’s percentage will not be weakened if resort to the shift system enables a plant to be worked by two or more sets of workers during the 24 hours. For such a capital economy is expressing an increased demand for complementary labor.

²³ Samuelson accepts that the stock of physical capital in the United States has grown six-fold since the beginning of the century, whereas population has only doubled. The number of man-hours has increased much less rapidly than population.

²⁴ This phenomenon forces one to ask again whether it is not really rather foolish to be concerned with percentage shares instead of with absolute shares. It seems indeed that, in a strike-threat age, labor in general is best off when its percentage share of aggregate income is least!

²⁵ Solow (op. cit., p. 619) has warned about the fuzzy use of the notion of “what one would ordinarily expect.” I hope that my use of the notion absolves me from suspicion of sloppy thinking on the issue.

²⁶ I do not include here the much less likely, but theoretically possible reason that, before the emergence of powerful unions, labor had been exploited for the benefit of investors through monopsony, while, as the exploitation was overcome through union resistance, a redistribution of income could have been expected. My reason for not again referring to this possibility here is that the argument presented in Chapter 8 has shown the improbability of more than negligible monopsonistic exploitation occurring in the actual world.

²⁷ See M. Reder, “Alternative Theories of Labor’s Share,” in M. Abramovitch, Allocation of Economic Resources (Stanford, Calif.: Stanford University Press, 1959), p. 196.

²⁸ I say “unpredicted” because strikes in governmental employments are illegal in most countries. But as the people of the United States have learned recently, illegality does not amount to much if law enforcement and the judiciary have been allowed to fall into political control.

²⁹In referring to the rentiers’ expectations, it is important to refer to the relative time-lag in the expectations of the low-income rentiers who, for other reasons also, are in an inferior position when it comes to escaping exploitation through inflation. Humanitarians should never forget that rentiers with incomes below the average include many relatively poor people who, largely in order to avoid having to rely upon public assistance or charity (for themselves or their dependants), invest in savings and loan associations, life insurance, endowment insurance, pension funds, etc. Income transfers at the expense of the real value of the savings of the thrifty poor seem to benefit mainly union members whose incomes are above the average of all incomes.

³⁰ This possibility must not be confused with the vastly more important shifts from work of low productivity to work of high productivity (noticed above, pp. 220-221), which would (in itself) increase only labor’s absolute (not its relative) share.

³¹ “Contrary to what one might expect, agriculture is a capital intensive industry utilizing relatively large amounts of capital (including land) per worker.” (D. Gale Johnson, “The Functional Distribution of Income in the United States,” Review of Economics and Statistics, 1954, p. 181.

³² I. e., not accompanied by governmental action (a) to permit or facilitate market-selected wage-rate adjustments or (b) to protect or raise the profitability of production (profit prospects) via the crude remedy of wage-rate “controls.”

³³ That part of the yield which exceeds interest is “consumers’ surplus.”

³⁴ Collectively-owned assets which produce an income for government, thereby lightening the tax burden, give rise to no problems.

³⁵ R. Goldsmith, A Study of Saving in the United States, quoted by I. B. Kravis, “Relative Income Shares in Fact and Theory,” American Economic Review, December, 1959, p. 920.

³⁶ See Reder, op. cit., p. 197.

³⁷ This point is made by Kravis (op. cit., p. 934) who, estimating from data in a U.S. Bureau of Labor Statistics publication of 1945, suggested that a dollar of nonfarm income was roughly equivalent to 70 percent of a dollar of farm income.

³⁸ Joseph A. Schumpeter, History of Economic Analysis (Oxford: Oxford University Press, 1954), p. 1042.

³⁹ Solow, op cit., pp. 618-619.

⁴⁰ Paul Samuelson, Economics (8th ed.; New York: McGraw-Hill Book Company, 1971), p. 719.

⁴¹ Martin Bronfenbrenner, “A Note on Relative Shares and the Elasticity of Substitution,” Journal of Political Economy, 1960, p. 287,

⁴² Reder, “Alternative Theories” p. 196.

⁴³ H. G. Johnson and P. Mieskowski, Quarterly Journal of Economics, November 1970, p. 539.

⁴⁴Ibid., p. 543.

⁴⁵Ibid., p. 547. A tax on unionized labor, the proceeds of which accrue to that labor, is only one way of envisaging the income-distribution consequences of any fixing of the price of a particular supply of labor under strike-threat duress.

⁴⁶Ibid., p. 548. My own analysis has stressed the composition rather than “the level” of investment, as I am about to reiterate.

⁴⁷ Ibid., pp. 554–558.

⁴⁸ Ibid., p. 559.

⁴⁹ Ibid., p. 548.

⁵⁰ Ibid., p. 560. In practice workers laid-off because of duress-imposed labor costs often find employment in less well-paid unionized occupations, as may “excluded” workers, never admitted to an activity they could enter under market freedom.

⁵¹ Ibid., p. 561.

⁵² Actually, the authors use the phrase, “If . . . the bargaining power of all unions is the same in all industries . . .” (Ibid., p. 561.)

⁵³ Ibid., p. 561.