("The Causes of the 1929-33 Great Collapse: A Marxian Interpretation," by James Devine)

[previous section] [end of this section] [section II]

[section III] [section IV ]

[appendices] [references] [notes] [short version] [new analysis of the future]

[[p. 120]]


I. Theoretical Framework.

As part of its effort to construct a Marxian interpretation of the Collapse's origins, this paper critically appropriates many different types of research, including neoclassical work. Such a method can produce eclectic rather than synthetic results. To avoid this pitfall, this section develops a general Marxian perspective, which will then be applied. Within this framework, it will be seen that many neoclassical authors excel at empirical work and description of superficial processes. What they lack, however, is theoretical depth.

These authors typically emphasize accidents or conjunctural institutions.2 This is partly accurate for any concrete situation or event. But while both structural causes and unique shocks to the system ("triggers") play a role, it is one-sided to ignore the societal structures affecting events. Given this, some theoretical principles are needed to help us discover which factors and which structures are most important in causing the event in question, the Collapse.

First, structural causes are more important than one-shot events: because of their greater persistence, structures help to shape and constrain individuals' actions and determine events again and again. Given this, how does one decide which social structures are most important?

A social structure is more likely to be determining rather determined to the extent to which it is durable, large, and world-encompassing.3 [[p. 121]] By these criteria, an analysis of capitalism and its laws of motion are most likely to produce an understanding of the origins of the Collapse.4

In terms of impact on the general historical process, long-term trends, and major events, capitalism (with other durable social structures) is more important than such one-shot events as the October 1929 stock-market Crash, which Christina Romer, among others, has brought back to the fore as a cause of the Collapse.5 First, Romer [1993: 31], is wrong to argue that since stock-market "bubbles are, almost by definition, inexplicable events, the [downward] stock market swing is legitimately viewed as an exogenous shock." Such bubbles (and their popping) are not mysterious but are a normal outcome of unfettered stock-market behavior, while the stock market itself is a commonplace accouterment of capitalism as a system. Second, and more importantly, the degree of structural stability of the world outside of Wall Street is central to the impact of such shocks. The degree of systemic stability differed drastically between 1929 and 1987 and thus was decisive in determining the impact of the stock-market crashes in those years: even though the magnitude of the Dow's decline was similar to that of 1929, the crash of 1987 had far fewer negative effects on the U.S. economy; its impact seemed limited to the brokerage industry. In fact, given the late-1920s fragility of world capitalism described below, other triggers could have substituted for the Crash in causing the Collapse. Finally, the dynamic nature of capitalism normally makes the existence of "shocks" the rule rather than the exception.

A similar logic applies to durable structures such as corporate law, which are constrained and shaped by the even more durable social structures such as capitalism. The laws of motion of capitalism are consistent only with a limited set of possible bodies of corporate law, while the capitalists actively push for an even smaller set, perceived to favor their interests. Part of the instability of capitalism in the late 1920s was due to that period's corporate law, which allowed pyramiding of corporations (e.g. Insull's utility schemes) and other forms of excessive leverage [Sobel, 1968: chs. 5,6].6 But the possibility of extreme leveraging is a determined outcome rather than a determining factor. Though post-1929 reforms helped prevent these problems for a long time, profit-seeking encouraged innovation and political lobbying to recreate them: in the 1980s, new forms of leverage (e.g., leveraged buy-outs using junk bonds) were invented. Further, as with stock-market crashes, the impact of excessive leveraging depends on the general stability of the economy.

Capitalism's laws of motion, however, cannot be the only determinant, even when complemented by corporate law. After all, we did not see crises of the 1930s magnitude before that time, even though capitalism and broadly similar corporate laws existed long before 1929.

Further, the less resilient structures and one-shot events cannot be reduced to mere epiphenomena of the more basic social structures, since they are relative autonomous. The set of possible corporate legal systems consistent with capitalism is not singular; thus, corporate lawyers can [[p. 122]] have an impact on history. Similarly, the speculative dynamic of the stock market, as studies of the 1929 and 1987 crashes suggest, cannot be simply explained by economic information about "fundamentals" (such as the dynamics of the profit rate). The structure of the stock market allows and encourages fads and speculative bubbles to affect its dynamics.

Because of this relative autonomy of the less durable structures from the more durable ones, and of one-shot events from "deeper" causes, these can and do play a role in determining the actual course of events. Specifically, they are crucial to the timing of economic crises. As argued below, the issue of timing can be pivotal, because the persistence of a capitalist expansion encourages the accumulation of imbalances that can make the ensuing recession more intense.

In sum, the 1929 Crash might be seen as analogous to a match that starts a fire that burns down a building. This chain of events can only occur if the building is already dry and flammable. But analogies are never enough: we have to move toward a concrete understanding of the causes of the Collapse. First, in section I.A, we need to describe capitalism and its crisis tendencies. As we shall see, capitalism in general is not sufficient to historical explanation, so more historically-specific institutions and events play a role, as developed in section I.B.

A. Over-Accumulation and Crisis.7

Crises are possible because of the unplanned nature of the capitalist system as a whole. But Marxist theory goes further, beyond disproportionality theory or Keynesianism, to predict that they are necessary or inevitable; this does not mean that crises cannot be delayed, but rather that forces that delay a crisis only tend to intensify it.8 Capitalism is an inherently dynamic system consisting of different parts (production, circulation, credit, and so forth) that develop unevenly. When the different parts get out of step, typically due to excessive accumulation, they suddenly adjust to get back into sync, which [after awhile] can allow further accumulation. It is this forcible adjustment which is an economic crisis. The initial maladjustment, as we shall see, is reflected in changes in the profit rate. The form of these changes can vary depending on the objective conditions encountered by accumulation.

Driving capitalism into over-accumulation and crisis are two structurally-based tensions, class antagonism and capitalist competition. First, the capitalists' domination and exploitation of workers in production and the resulting alienation results in persistent class antagonism. This antagonism, however, is not always expressed in overt conflict, because competition among workers typically mutes class issues. But despite the many divisions within the working class, class conflict appears (to different degrees) in almost all workplaces, due to the underlying class antagonism and the interconnectedness of the different workers' experiences and interests.

The reality of, and potential for, this conflict has an impact on management decisions: they are not simply making technical decisions about how to allocate resources but are also trying to solve a socio-political [[p. 123]] problem of how to induce workers to promote the owners' profit goals. This problem affects the management techniques used and the technologies introduced, as with Taylorization [cf. Braverman, 1974]. It also encourages capitalist expansion (to get greater power over labor) and/or capital flight.

While capitalist competition contributes to divisions within the working class, thus moderating societal stress, it is also the second basic structural tension of capitalism. Unlike the passive competition of vendors under simple commodity production, competition among capitalists is dynamic and aggressive [cf. Marx, 1849: sect. 5]. Each capitalist must worry about falling behind actual and potential rivals and so must actively expand -- invade old markets, create new ones, introduce new technologies and management strategies, and so forth. Each must accumulate to survive as a capitalist, rather than fall into the overworked petty-bourgeois fringe or even lower.

Both tensions encourage expansion of the system, to swallow almost the entire world. Competition also tends to disrupt established institutions, including stalemates in class relations. The latter encourages overt class conflict as businesses try to pass the costs of competition (e.g., those of a crisis) onto workers or onto other classes such as the petty bourgeoisie. Alternatively or in tandem, each capitalist tries to dump such costs onto other capitalists, intensifying capitalist competition.

Competitive accumulation also drives the business-cycle expansion, which is allowed and encouraged by the competition among banks in supplying credit. Such expansion complements -- and thus amplifies -- the results of multiplier-accelerator interaction and other reasonable mainstream explanations of instability.9 This regularly leads to aggregate over-investment and crisis ending a boom.

Here, "over-accumulation" has been replaced by "over-investment" because the emphasis is on excessive fixed capital. Circulating capital, unlike fixed capital, is not a persistent imbalance that can block accumulation for long.

The main effects of over-investment can be explained in terms of its impact on rate of profit. For Marxists, this rate gets as much attention as does the money supply for monetarists: in Thomas Michl's [1988] words, it is both the thermometer and the thermostat of capitalist accumulation, a key measure of the system's health -- from a capitalist perspective -- and an agent of causation, as capitalists react to its fall.

Before developing this rate's role in crises, we must address two dimensions of controversy about how the profit rate should be measured. Should one calculate it using value (labor hours) or price magnitudes? Second, to use Moseley's [1991] terms, should one use the "Marxian" profit rate or the "conventional" rate of profit? The former, among other things, includes the wages of unproductive labor as part of surplus-value or profits, while the latter uses the more empirical definition of profits. Because we want to understand the decisions of capitalists, this paper uses the profit rate calculated from price data following conventional definitions.10 Since capitalists make decisions at the level of empirical appearances [[p. 124]] rather than at that of socialized production, this profit rate is the one most likely to determine the rate of accumulation of fixed capital (fixed investment) and to be the proximate cause of crises.11

Thus, the following profit rate will be used:

r = R/K = (R/Y)(Y/Z)/(K/Z)


where R is conventional profit income, Y is national production, Z is full-capacity production, and K is the stock of fixed means of production ("capital").12 R/Y, Y/Z, and K/Z are the profit share, the rate of capacity utilization, and the capital-capacity ratio, respectively. These are rough price analogs of Marxian value categories, i.e., the rate of surplus-value, the speed of turnover of commodities (the inverse of commodity turnover time), and the value composition of capital, respectively.13

The tendency toward over-accumulation is expressed in at least three different ways: underconsumption, the high-employment profit squeeze, and the rising composition of capital. First, in the Grundrisse, Marx writes that individual capitals compete to push real wages down relative to productivity, possibly causing underconsumption [1857-8: 420]. Even though such cost-cutting is seen as good for individual profit production, it can hurt profits on the societal level as wage incomes constrain consumption, which in turn prevents the full realization of profits. As with realization problems in general, this depresses the profit rate by slowing the circulation of commodities and thus lowering capacity utilization (Y/Z).

Classical underconsumptionism posited that depression is normal for a capitalist economy, arising from a persistent tendency toward low consumer spending [Bleaney, 1976: 11]. Marx and many Marxists decisively criticized this theory and have been absolutely right to reject such universal stagnation tendencies [cf. Bleaney, 1976; Clarke, 1993].14 However, we should not reject the role of stagnant consumption in causing or encouraging crises under certain specific historical conditions. Unlike classical underconsumption theory, this paper (1) emphasizes forces endogenous to capitalism which drive it to over-expand rather than to stay mired in stagnation (see above); and (2) sees major periods (such as the 1960s) in which low consumption did not cause problems for capitalism.

Underconsumption forces can play a role in two cases [Devine, 1983]. First, underconsumption problems can be crucial during the period after the crisis (whatever its reason), in an "underconsumption trap."15 Falling wages and workers' consumption hurts profit rates if other elements of aggregate spending are prevented from rising enough to fully realize profits. Specifically, if capitalist accumulation is blocked by a mutually reinforcing combination of unused capacity, excessive debt, and pessimistic expectations, the competition to cut [[p. 125]] wages can contribute to turning a recession into a depression, as in the early 1930s (see section III.F).

Second, in the theory of "over-investment relative to consumption," the existence of stagnant workers' consumption requires a growing share of accumulation in the national product in order to realize surplus-value and profits. The structural tensions discussed above and rising profit rates can encourage such growth. Though in theory it is possible for growth to continue forever in this situation, accelerating accumulation implies that the economy becomes increasingly unstable and prone to collapse.16 The theory centers on the negative effects of excessively rising profit rates in an economic boom, and is developed for the 1920s in section III.C.

The other two major crisis theories emphasize over-accumulation relative to supply constraints as leading to the accumulation of persistent imbalances that depress profit rates. The second theory, emphasizing external constraints, appears when Marx writes (in chapter 25 of volume I of Capital) of over-accumulation relative to labor-power supplies, the high-employment profit squeeze: as the economy moves toward high employment, accumulation can pull up wages, lowering R/Y. Similar effects arise if capitalists spend too much on unproductive labor [cf. Moseley, 1991] or on raw materials (or as capitalism drives itself into ecological crises). Usually these three happen simultaneously to different degrees, as part of over-expansion relative to supply. [See Devine, 1987]

To many readers, the high employment profit squeeze is not a crisis theory at all but instead a theory of an automatic mechanism which allows the restoration of the normal conditions for capitalist exploitation and accumulation.17 However, if excessive accumulation is allowed to continue by countervailing forces such as government war spending or credit expansion (as during the 1960s), we can see persistent imbalances develop. This in turn can cause the forcible adjustment of the system -- the painful purgation of imbalances -- that defines a crisis. [In the 1970s, this accumulation of imbalances and the resulting falling rate of profit was reflected more in stagflation than in "old fashioned" recessions, partly because of credit expansion and public policy aimed at preventing recession.]

The third theory emphasizes capitalism's endogenous creation of supply constraints to accumulation. It stems from Marx's famous discussion, in volume III of Capital, of the tendency for the rate of profit to fall: over-accumulation results in excessive use of fixed capital, which raises the composition of capital and K/Z, depressing the profit rate and thus accumulation.18 This theory has been criticized largely using models that assume static rather than dynamic competition and, crucially, that real wages stay constant despite increases in labor productivity (the Okishio theorem). If real wages rise with [labor] productivity (so that the rate of surplus-value or R/Y is constant), as in several of Marx's presentations, this theory is more viable [cf. Laibman, 1992, ch. 6]. More generally, the theory works better to the extent that real wages rise relative to productivity. Further, as Dobb [1937] argued, the over-accumulation relative to labor-power supplies might actually induce excessive K/Y.

The high-employment profit squeeze and rising composition of capital can thus be seen as being potentially complementary phenomena, as in Armstrong, et al. [1991: ch. 11]. Finally, there is some evidence [[p. 126]] for rising K/Z contributing to falling profit rates in the U.S. during certain periods, e.g., before 1919 or so (see section III.D) and after 1972 (see section IV.C).

Over-investment, whatever its cause, leads to recession and/or financial crunch (as over-expansion of loans becomes obvious to creditors). Once a cyclical downturn hits, it spawns movements toward both recovery and continued stagnation. On the former, the downturn often reverses the forces that depressed the profit rate, restoring the conditions allowing accumulation. Rising unemployment depresses wages relative to productivity, allows production speed-ups, and lowers the demand for raw materials, encouraging any high-employment profit squeeze to end. Further, some fixed capital is scrapped or sold far below historical or reproduction cost, moderating or reversing any rise in the composition of capital (and in K/Z). Excessive debt can be purged via bankruptcy. So the recession [eventually] helps create conditions allowing the next boom. On the other hand, it is possible for the imbalances created by the expansion phase to block the normal aggressiveness of accumulation, as in the underconsumption trap. Moreover, a decrease in the system's social legitimacy and a resultant intensification of overt class conflict might discourage accumulation and thus recovery.  

B. The Institutional Environment.

The above formulation of crisis theory does not provide enough information to determine which of the major crisis theories predominates at any one time. Even given a basic cause, abstract tendencies determine neither the length and robustness of the prosperity period nor the timing and intensity of economic downturns. Similarly, the conflict between tendencies that result from a downturn means that the actual outcome depends on factors as yet unexplained. Thus, in explaining the empirical path of the economy over the cycle and longer periods, there is a role for historically-specific conditions and institutions. Most importantly, this social environment helps determine whether a downturn is merely a cyclical crisis (a recession followed by a recovery) or a deeper and more serious structural crisis,19 such as the Depression of the 1930s or the contemporary "Silent Depression" (see section IV.C). So we must move from abstract capitalism toward considering actually-existing capitalism. A contrast with the Marxist "Regulation" school sets the stage for this paper's theory.20

In this view, each historical era of capitalism's development has a dominant "regime of accumulation." For capitalist growth to be stable for decades, as after World War II in the core capitalist countries, the regime of accumulation must be complemented by an institutional "mode of regulation." Crises such as the Great Collapse occur because the mode of regulation is inadequate to stabilize the regime of accumulation. Since the start of the 20th century (until perhaps the 1970s), the regime of accumulation in the U.S. and other core capitalist countries was based on intensive accumulation, i.e., the Taylorization [[p. 127]] of production, rapid increases in labor productivity, and capital deepening. The rise of mass-production technology made the economy unstable, because it did not mesh well with pre-World War II "competitive regulation" which centered on regular wage-cutting. In order to regularize the production and realization of profits, mass-production technology requires the existence of stable mass consumer markets -- based on high and rising real wages. Because this could not exist under competitive regulation, the U.S. economy during the 1920s was ripe for a fall into Depression.21

Rather than criticize the Regulation school at length [cf. Brenner and Glick, 1991], four major points should be emphasized. First, we should avoid that school's tendency to ignore the international regime, since the world economy is crucial. Given the long-term growth of interdependence since the 18th century, it is increasingly central to the behavior of all national economies. Also, it must be explained why the U.S. was so crucial so that Regulationists such as Aglietta can devote almost all of his attention to it.22 Below, this paper emphasizes the structural need for a hegemonic power on the international level; this requirement seems to be part and parcel of the increasingly overt socialization of capitalism, which in the limit requires a world government.23 As Kindleberger [1986] and other "global Keynesians" argue, the U.S. failed to fulfill this role during the inter-War period, even though its size and financial impact made the U.S. economy the key link in the international chain. But Kindlebergerian analysis is insufficient: we need to bring in the classical-Marxist theories of imperialism. Of all of these theories, this paper emphasizes that of Nikolai Bukharin [1917] and what John Willoughby [1979] terms the "Rowthorn-Mandel Rivalry-Hegemony-Rivalry Cycle." This suggests not only that the absence of a hegemon destabilizes the system but also that the normal uneven development of the world system makes the existence or non-existence of a hegemon a temporary state of affairs. Capitalist development involves not only hegemony but hegemonic decline and wars of succession [cf. Goldstein, 1988: chs. 5, 6, 13, 14]. [We must also consider the possibility that problems may result when the hegemonic power becomes excessively strong, perhaps as in the 1990s. This concern is suggested not only by current events but by Temin's [1989] emphasis on the incorrect international policy regime of the late 1920s. The short-sighted use of power can be disastrous.]

Second, for the United States itself, Brenner and Glick [1991] argue convincingly that "intensive accumulation" is normal under full-blown capitalism and precedes the 20th century by decades and the notion of competitive regulation is poor. There is also much doubt concerning the hypothesis that Taylorism surged in practice or effectiveness (as opposed to in rhetoric) before the Collapse [Edwards, 1979: 97-104; Nelson, 1991]; the rise of "scientific" management instead seems to be a gradual long-term trend, applied in a general way that put into practice principles that predated Taylor. One point in favor of the Regulation view is the ratchet upward of the trend labor productivity growth rate after 1919 or so. But this shift needs explanation (see section III.A).

Instead of modes of regulation, this paper emphasizes the role of different long-term regimes of the relative abundance or scarcity of labor-power [cf. Dobb, 1963: 23]. "Labor scarcity" should not be interpreted in narrow economic terms; it simply means that real wages rise relative to productivity as accumulation [[p. 128]] progresses. This can be due to one or more of the following factors: (1) increasingly powerful (and widespread) working-class organization; (2) bottlenecks preventing labor-power from moving to where demand is growing or capital from moving to where the labor-power is abundant; (3) the normal increase in working-class needs encouraged by capitalist accumulation, which pushes workers to fight for higher wages [cf. Lebowitz, 1992: ch. 2]; or (4) barriers to labor-saving technical change. Relative labor abundance is the opposite of this, a period when capital "gets its way," at least in the short term.

The normal tendency for capital to over-accumulate appears differently in the two types of regimes [cf. Devine, 1983, 1987].24 In a "labor scarce" situation such as that of the U.S. in the 1960s, over-expansion leads to profit rates being squeezed by high labor and raw material costs plus a rising composition of capital. But in a labor-abundant economy such as that of the 1920s in the U.S., as accumulation proceeds, production of surplus value is relatively easy and the share and rate of profits rises in leading sectors (see sections III.A and B). This led to over-investment relative to demand (see section III.C). The underconsumption trap theory (see section III.F) also presumes that real wages fall relative to productivity.

Third, the long-term fall in the profit rate in the U.S. before the 1920s [Duménil et al., 1987] must be integrated into the story, though the reasons for this decline are beyond the scope of this paper. These authors' focus on the 1900-1929 period as a way to understand the 1920s and 1930s also seems a step forward. However, this paper tries to go beyond their excessively aggregative analysis in order to understand uneven development among sectors within the U.S. Thus, we should reject these authors' view, and that of Gordon et al. [1983] and Brenner and Glick [1991], in which the rise in the wage share of income was a cause of the Collapse. In the crudest possible terms, this paper's conclusion is that the problem in the late-1920s U.S. was low rather than high wages.

In more sophisticated terms, a long-term fall in the economy-wide profit rate (up to the end of the 1910s) encouraged an intense effort to boost profit rates, which was largely successful in the leading sectors (manufacturing, the corporate sector) until 1929. In that year, the latent problems of growing income inequality and slow growth of consumption that resulted from this effort came to the surface, encouraging the Collapse. This theory follows the classic falling-profit-rate-and-then-underconsumption sequence mentioned in section A , but there is a long delay associated with the "then."

Fourth, the Regulation school's (and the related Social Structure of Accumulation school's) emphasis on historically-specific institutions ("modes of regulation"), while not wrong, is one-sided. Institutions such as the state, banks, and the like, can and do stabilize capitalism, at least temporarily. But since the essential elements of capitalism that encourage crises are not abolished by such stabilization, temporary prosperity is purchased at the price of allowing the piling-up of imbalances which simply make the crisis more severe when it comes. [[p. 129]]

Moreover, such institutions reflect and involve the same basic tensions that encourage crises in the first place, i.e., class antagonism and capitalist competition. Thus, they are necessarily imperfect stabilizers. In fact, they can intensify crisis tendencies.

Turn now to more concrete analysis, which combines these elements, i.e., worldwide inter-state contention and, in the U.S., uneven development and labor abundance, profit-decline-induced over-investment relative to demand, and historically-specific institutions and events.

[introductory section] [top of this section] [next section]

[section III] [section IV ]

[appendices] [references] [notes] [short version] [a new analysis of the future]