Discussion of empirical findings

The empirical findings from our study of the U.S. input—output data as well as the results from other researchers suggest the following:

a The WRP curves display, more often than not, an “irregular” or what is the same “well-behaved” shape in the aggregate economy. Slightly convex in the case of a fixed capital model, and more linear and imperceptibly concave shape in the circulating capital model, b The same conclusion does not hold at the level of particular industries as we showed in Figures 5.9—5.14. Both curvatures (convex or concave) are possible independently of the employed (circulating or fixed capital) model.

c There is a clear trend in rising maximum real wage or productivity of labor for the aggregate economy and to individual industries with only a few exceptions indicating idiosyncratic characteristics of particular industries. Consequently, a rising maximum real wage amounts to falling unit costs and prices, as a result of rising productivity induced by technological change.

d For low profit rates, the WRP curves rotate upward induced by technical change. We just cannot see this effect in our graphs, simply because by construction and convenience in presentation, the curves share the same horizontal intercept, which is the standard ratio or maximum relative rate of profit, for p = 1. The maximum rate of profit, R, or what is the same thing, the output-capital ratio or “productivity” of capital over long stretches of time is expected to be falling. If the vertical intersection of a recent WRP curve is above that of the past years and, this is the usual case, we have technological progress, which potentially can be measured and compared with that of the curves of past years by the area under the respective curves.

e The WRP curves either do not cross at all, the usual case or, cross only once in the relevant (positive) region.

f Given all of the above, it follows for reswitching to be possible, the two WRP curves must be too close to each other and the more convex WRP curve must have both ends (the maximum real wage and maximum rate of profit) higher than the other WRP curve. The above requirements make reswitching a remote possibility.

Table 5.1 reports the maximum real wages and the maximum rates of profit for both circulating and fixed capital models for the U.S. economy and selected years. These findings may be of help in our judgment about the possibilities of reswitching, provided the approximate linear character of the WRP curves.

In Table 5.1, we observe that both models share the same maximum real wage (or productivity of labor), which increases over time indicating technological progress. This combined with falling, as expected from the classical analysis, maximum rate of profit limits the possibility of reswitching; this, however, does not preclude the case, for short time periods, the maximum rate of profit to be higher (as is the case of the years 2007 and 2014) than that of a previous year against which is compared. Under these circumstances, the old technique is inferior and is absolutely dominated by the more recent; so the case of reswitching may be excluded. Consequently, Samuelson’s (1966) “facts of life” do not include the general occurrence of “capital reversals” on which Sraffians have invested so much over the years in the hope that neoclassical economists would ultimately recognize the “truth” and turn to the strand of the classical theory of value and distribution that makes no use of the upsetting, for its social repercussions, labor theory of value.

The idea of switching is not necessarily supported by the empirically found near-linear shape of the WRP curves or the monotonic shape of the PRP curves. The above empirical findings do not, however, rule completely out the case of switching, which, time and again, has been found to be not only exceptional but, whenever it happens, it takes place at unusually high or

Table 5.1 Maximum real wage and rate of profit in circulating and fixed capital models

Circulating Capital Model Fixed Capital Model


Maximum Real Wage or Labor Productivity

Maximum Rate of Profit

Maximum Real Wage or Labor Productivity

Maximum Rate of Profit


























low rates of profits. Consequently, the rare occasions of switching do not cast doubt on the classical political economy theorization of price trajectories consequent upon changes in income distribution as Sraffian economists in the late 1970s hastened to claim (Steedman 1977).

From the Sraffian perspective, Kurz and Salvadori (1995, p. 450) were from the first if not the first that commented on the empirical findings of the 1970s and 1980s, all of which were supportive of the near linearity WRP curves. They characterized the empirical works “fundamentally mistaken” by arguing that the WRP curves because they were constructed from input—output tables of different years they did not allow for the effective choice of techniques to take place over the examined years. In other words, the so derived WRP curves represented the only technique applicable in a single year, whereas the theoretical relation “refers to technical knowledge at a given moment of time”. However, one cannot downplay the fact that when empirical data are given at quite distant times, say for every five years, as for example in the U.S. benchmark input—output tables, then apparently there is no real “choice” of alternative techniques from those employed in the past. The reason is that firms are forward-looking and, neither, choose techniques from a hypothetical book of blueprints of various past alternatives nor they just look at the distributional variables and switch to a new technique, which suits best to the new combination of factor payments and so forth.

The near linearity of WRP curves for the various economies, however, should not lead to the conclusion that the neoclassical theory escapes criticism. The marginal productivity theory of income distribution, the cornerstone of neoclassical economics, is based on the assumption of perfect competition in all markets, which allows for the perfect substitution of factors of production that leads to the choice of technique and the optimization behavior of the firm. However, the idea that firms freely choose between alternative techniques available in a book of blueprints specifying all possible input combinations for the production of a given flow of output, and select the most efficient combination, is meaningless in conditions of real (or classical) competition, a dynamic process of rivalry among units of capital over market shares (Shaikh 1978; Tsoulfidis and Tsaliki 2005). The idea is that in real competition, the so-called “choice” of technique is not realized in any smooth way, as is usually claimed in the neoclassical approach, according to which even a slight increase, for example, in wage, is enough to lead to the immediate substitution of labor for another relatively cheaper productive resource. In real competition, techniques change after the passage of relatively long stretches of time in order for their effects to become visible. Moreover, the substitution between factors of production is more limited than is usually thought, in the short period, at least, since techniques, which are actually in use, are not so sensitive to price changes. Consequently, between the neoclassical notion of perfect substitution and Leontief’s conception of fixed proportions (the famous “cooking recipe” analogy), we would say that the latter represents a closer to reality description of the economy.

Sraffian economists (not all) have silently adopted an approach to competition no different from perfect competition and along with a static time framework attempt to derive the reswitching results. It seems that followers of the Sraffian tradition would rather exclude these near linearity WRP, disturbing for their approach, curves and for this reason, they pose a Kuhnian kind of “protective belts” around their core proposition of “reswitching of techniques”. If reswitching is true, not only theoretically but also empirically, then the neoclassical theory is in big trouble and the Sraffian alternative is there waiting to become the new theory of value and distribution adopted by economists because of its logical consistency and, therefore, superiority. Because the required data are available, there is no choice for the Sraffians but to utilize them in testing the validity of their core theoretical proposition. The testing terrain for the Sraffians and their theory of value is the derivation of the exact shape of both the PRP trajectories and the WRP curves associated with them.

It is important to emphasize that the WRP curves that we derive for each year and industry are just the weighted average technique employed in each industry, where many firms or rather establishments according to the BEA and other official statistical services, or “units of capital” according to Marx, are activated. However, each one of the establishments or firms is employing its own technique. Usually, relatively few capitals or firms activated in the industry utilize the very advanced techniques and other less advanced capitals are stuck with their older and perhaps outdated technique, which probably should be abandoned any time soon; however, capitals are bound to carry on their old technologies because of their past investment decisions, which become compelling under the present circumstances. The weighted average technique is the one reflected in the WRP curves of each industry and not necessarily the dominant or the “best practice” in the sense that it attracts new investment and determines prices and profit rates (Salter 1969, ch. 2; Tsoulfidis 2015). This does not mean that the average techniques are necessarily by far too different, both quantitatively and qualitatively, from the dominant (especially in most manufacturing and service industries); thus, by utilizing the average technique, we do not get quite different results from those that we would have derived had we had the option of using the dominant techniques. However, there are industries such as in agriculture and mining where the “dominant” producers (or “regulating capitals”) happen to be the least efficient ones and over time, as demand increases, we may have the old “least efficient” producers to become the more efficient and the new average technique may even fall below the old, provided that there is only increase in demand and no technological change. In other industries, such as the high-tech ones, the most advanced firms (the dominant ones) set the pace for the whole industry and operating under these conditions becomes extremely difficult for the outdated technological firms to catch up (Tsaliki and Tsoulfidis 2015).

From the above discussion, it does not follow that the various other techniques, even the dominant ones, activated within an industry and their WRP curves are of an altogether different shape from the specific year’s weighted average. Moreover, if we could make a selection of dominant techniques over the years, we would not obtain any different WRP curves and frontiers from those derived from the available input—output data average technique based on benchmark years during which there is a sufficiently long time for technological change to take place. After all, the more advanced techniques of today, employed by a relatively limited number of establishments in the industry after a few years, will be diffused and they will become the new industry’s average. It follows, therefore, that if the shape (quasi-lineav) of the WRP curves in every year tested are quite similar to each other, one does not expect any different WRP curves had the choice of techniques been made over the years.

< Prev   CONTENTS   Source   Next >