A new theoretical framework links inflation and accumulation, with the Israeli experience as a case study. The focal point is the process of differential accumulation by the largest core firms.

1. INTRODUCTION

THE RISE AND DECLINE OF ISRAELI INFLATION - from two percent in the mid-1960s, to over 400 percent in the mid1980s, to less than ten percent in the late 1990s - remains a riddle for Israeli economists (Levitan and Pitterman, 1989, 437). Particularly mysterious, write Razin and Sadka ( 1993, 3) , is the fact that this rise and fall was counter-cyclical: the low inflation of the early years was associated with rapid economic growth, the high inflation of the 1970s and 1980s with stagnation, and the disinflation of the 1990s with an apparent "overheated" economy.

A large part of the riddle has to do with the methods used to solve it. Most of those writing on the subject willingly arrest themselves in the ahistorical straightjacket of orthodox macroeconomics. Their common starting point is that inflation comes from excess demand, with the main debate revolving around the specific source of this excess. The usual suspects are all "autonomous": greedy workers, foreign oil sheiks, and of course half witted, weak politicians. Fortunately, though, since inflation is created "by will," it can also be stopped at will. All it takes is visionary statesmen who will listen to professional economists and implement bold (albeit) orthodox policies.

The purpose of this paper is to offer an alternative way of understanding the inflationary experience of Israel, situating it within a broader theoretical framework of capital accumulation. Viewed from this perspective, inflation is neither the consequence of market imperfections, nor the result of misguided policies, but rather part of an ongoing restructuring of capitalist society. The central facet of this restructuring, we argue, is the process of differential accumulation, that is, the progressive redistribution of ownership claims in favor of dominant capital groups. This latter process can be sustained in a number of different ways, one of which is inflation. The implication is that in order to understand when, why and how inflation rises and falls, we first need to articulate the broader processes affecting accumulation in general, and differential accumulation in particular.

The second section of the paper outlines the basic concept of dif ferential accumulation and the two principal regimes through which it unfolds. The third section applies this framework to the Israeli case, examining some of its main historical contours: the rise of inflation during the 1970s and 1980s and the underlying political-economic forces sustaining it. The fourth section situates the inflationary process in the wider discussion of power and conflict. The final section examines the transition since the mid-1980s to a period of low inflation.

2. DIFFERENTIAL ACCUMULATION: CAPITAL AS POWER1

As implied by its name, capital is the central institution of capitalism. It is therefore surprising that no one really knows exactly what it is. Most people think of capital as something "real," in the sense of machines, buildings, or more broadly as a technological process. But that is too simple. All of these material "things" existed well before capitalism, as well as in parallel systems such as communism and fascism. Moreover, capital can exist and expand with little or no machinery - consider the case of Microsoft, whose highly "nonmechanized" operations nonetheless generate one of the world's fastest rates of capitalization.

The key to the enigma of capital is anchored in the relationship between income and productivity. For business people, capital is simply the capitalization of future profit expectations (Veblen, 1904, 1923). Mainstream economists, despite the legacy of the Cambridge Controversy, continue to assume that such earnings come from the productivity of machines and structures owned by the firm. But then, how should we reconcile this belief with something such as the recent Asian crisis, where capital values were buried under an avalanche of new capital goods? Arguably, the root cause of this crisis was "overinvestment" (understood in a business sense). Encouraged by cheap loans and ever growing exports, Asian companies went on augment ing their productive capacity well beyond what could be sold at profit able prices. Eventually, the discrepancy triggered a loss of confidence, culminating in massive decumulation. In other words, a rise in the "quantity of machines" led to a fall in the "quantity of capital." An opposite example is offered by the "capitalization" of the former Soviet Union, where rapid amalgamation of private wealth was achieved through a massive destruction of productive capacity.

How is this possible? How could more machines lead to less capital, and fewer machines to more capital? The answer is simple: capital has to do not with machines and their utilization, but with profits, and the relationship between profit and production is complex and potentially nonlinear. Up to a point, profit increases with output, but then, if output growth turns into glut, the relationship could become negative, leading to business ruin. More generally, profit is determined not only by how much is produced and sold, but also by the manifold political institutions of distribution. The conventional wisdom that each factor of production gets paid what it contributes at the margin is nothing but a well-defended myth, as the high priests of neoclassical economics reluctantly admit (Samuelson, 1962; Ferguson, 1969). For a start, no one has ever been able to observe individual factor contributions. Moreover, even if each factor of production had its own intrinsic productivity - a very big if - the only way this could lead to a proportionate distribution of income is through perfectly competitive markets, which do not exist and perhaps never did (Braudel, 1985). Markets usually operate with some principal actors who can affect production and prices. More importantly, markets never exist in a political vacuum. In the real world, market outcomes as well as their very existence, are political phenomena (Polanyi, 1944).

Accumulation, understood as the progressive capitalization of expanding profit, is a manifestation of numerous power arrangements and institutions. What makes profit possible is not productivity per se, but the strategic control of productivity for business ends. From this perspective, the study of accumulation is the study of politics at large, the deciphering of social power under capitalism. This, of course, is not new. Marx devoted much of his life to examining this link, as did many of his followers. Yet, the literature on power and accumulation remains besieged by a fundamental dilemma: it views social power and capital accumulation as related though essentially distinct phenomena. Indeed, when it comes to measurement, both Marxist and neoclassical analysis count accumulation in the "materialistic" units of labor time and utils, respectively. Power can still influence accumulation, of course, but only from the "outside." In contrast, our approach seeks to integrate both power and productivity into the very definition of capital.

According to orthodox economics, capitalists are driven by the desire to maximize their hedonic pleasure, and the vehicle for doing so is profit maximization. We begin from a different starting point. Capitalists, we argue, are driven by the quest for power, not utility, and the way to achieve it is not by maximizing profit, but by "beating the average." The principal drive of capitalists, operating through their corporate organs, is to have their profit and capitalization expand faster than those of other owners; they seek not absolute accumulation, but differential accumulation.

Capitalist power is exercised over society as a whole. This is what is generally meant by the "power of capital" (Gill and Law, 1989) . But that is not enough. The overall power of capital depends on its internal dynamics. If capitalist power were indeed as defused as the neoclassicists suggest, the likely result would have been the disappearante of profit and possibly of capitalism itself. Capitalist power, like other forms of power, is based on exclusion. In other social systems, exclusion is usually the upshot of rigid, largely static customs. Under capitalism it is dynamically created through the formation and ref ormation of "distributional coalitions," to borrow the term coined by Olson (1982). The upshot is that the accumulation of capital in general depends on the accumulation of capital at the center. It is dominant capital, the large amalgamations at the core of the process, that are crucial. The periphery of capital, comprising the many capitals outside the core, is in fact a constant threat to the general power of capital. Only to the extent that dominant capital is able to retain and augment its exclusive power against other lesser capitals, can the overall power of capital be sustained.

In sum, power is both the end and the means of accumulation. On the one hand, profit is sought not for the sake of consumption, but as a code of power, reflected in the constant drive of business to "beat the average" and exceed the "normal rate of return." On the other hand, profit can be obtained and retained only through the formation of exclusive coalitions, preventing others from accessing the same sources of power. The dual role of power suggests that in measuring accumulation, our units should be relative, not absolute, and that our focus should be not only on capital as a whole, but also and perhaps more so on its internal organization.

The locus of capital accumulation, we argue, is a differential process, a process by which the large corporate coalitions at the core of the political economy emerge, grow, and sometimes disintegrate and decline. The principal indicator for this process is the growth of profit and capitalization of the largest corporate groups relative to the economy's average. The process unfolds through what we call regimes of differential accumulation. These regimes are useful in understanding the links between the process of accumulation on the one hand, and the general contours of capitalist development on the other. We turn to examine them now.

Regimes of Differential Accumulation: An Analytical Framework

By definition, the level of profit is the product of profit per employee and the number of employees. This suggests two basic ways for corporations to expand their control over profit: they can either increase their breadth of accumulation by enlarging the size of their organization measured in terms of employment, or they can augment their depth of accumulation by raising their profit per unit of organization, measured in profit per employee. The depth of accumulation itself can be further decomposed as a product of three components: the markup expressed as the profit share of sales, the price level measured as the ratio between nominal sales and output volume, and output per employee (assuming all output is sold). Conceptually, we can write:2

Enlarge 200% Enlarge 400%

Now, our claim in the previous section was that in order to understand the evolution of mature capitalism, we should focus on the largest blocs of capital, as well as on capital at large. The key process, we argue, is the extent to which these core groups of dominant capital are able to "beat the average" and achieve a differential rate of accumulation. This can be examined by expressing every component of equation (1) in relative terms, i.e., as a ratio between a typical dominant capital firm and the average firm in the corporate universe:3

Enlarge 200% Enlarge 400%

Now, looking at a financial asset as the present value of a profit stream, its rate of accumulation is given by the rate of growth of profit less the rate of growth of the discount factor. When we measure differential accumulation, though, the discount factor, which is common to all assets, drops out. The result is to make the rate of differential accumulation (DA) - in our case, the extent to which a typical core firm accumulates faster than the average firm - depend solely on the difference between their respective profit growth rates. Using equation (1), this can he further decomposed, such that:4

Enlarge 200% Enlarge 400%

This framework for differential accumulation can be used at various levels of analysis, from the individual core corporation, through corporate coalitions, to the corporate core as a whole. At every level, it enables us to examine accumulation and power not as distinct processes that need to be related, but as two sides of the same process of capitalist development. The abstract crystallization of power appears on the left-hand side of our equations in the form of differential accumulation, while the concrete political-economic processes through which this power is formed and reformed appear on the right-hand side, as components of breadth and depth.

Viewed as broad social regimes, depth and breadth involve dif ferent and often conflicting processes. Consider first the breadth route. Dominant capital groups could achieve differential employ ment growth either by adding new, "green-field" capacity faster than the average, or through amalgamation (which raises their own employment but not total employment). The preconditions for a sustained breadth regime is overall employment growth, or expansion into new territory where the core groups do not yet operate. Otherwise, added capacity is bound to become excessive, while the supply of new acquisition targets dries out. Politically, a breadth regime necessitates some measure of openness and stability, in the absence of which growth is undermined and capital mobility runs into barriers. The combination of high growth, capital mobility and relative political stability works to mitigate both pricing power and the pressure to use it. The result is that a breadth regime is commonly characterized by relatively low inflation.

A depth regime is far more conflictual. It involves either differential growth of output per employee (so-called "labor productivity") which necessitates the erecting of barriers to prevent competitors from using the same technology; differential inflation, which is self propelling and potentially destabilizing; or differential markup growth, which implies competitive redistribution and hence a heightened capital-labor conflict. Seen from a Veblenian perspective, all three mechanisms are based on "strategic limitation" by the core groups, who are able to politically "sabotage" society's productive effort for their own differential gain. The broad economic feature of the depth route is therefore stagflation - stunted growth, combined with relatively high inflation.

Generalized as regimes, breadth and depth are unlikely to take place simultaneously. Breadth involves a competitive capitalist race to grab new profit streams in the context of economic and employment growth. It thus weakens the ability and resolve of large firms to act in concert, openly or tacitly, in pursuit of capital-labor redistribution, which is the main venue of depth. Similarly, differential accumulation through depth, because it thrives on limitation and "sabotage," ends up hampering growth and spoiling the benign political climate necessary for unfettered corporate amalgamation. It thus undermines the very preconditions of breadth.

Between the two regimes, breadth is the path of least resistance. Although growth may loosen business cooperation among the core groups and hence erode their pricing power, this is commonly tolerated in the expectation that such power will eventually be restored through amalgamation. In addition, growth tends to obscure the underlying capital-labor conflict, thus helping to consolidate the social hegemony of business institutions. Differential expansion via breadth is not always feasible, though, and when circumstances undermine green-field investment and/or amalgamation, the core reluctantly gravitates toward depth.

Compared with breadth, the depth route is far more contentious in methods, risky in application, and uncertain in outcomes. But the differential gains, at least in the short run, are potentially huge. The reason is that, contrary to breadth, where expansion is counted in employees, depth is counted in the much more "flexible" monetary units of profit. Unlike the rate of growth of the former, that of the latter has no inherent limit. The spoils, therefore, are commensurate with the risk, and when dominant capital is pushed to the wall, with no breadth left, the temptation and imperative to dive into depth are difficult to resist.

This framework is not deterministic. We do not need to have either differential breadth or differential depth. Dominant capital may be unable to achieve a positive rate of differential accumulation, in which case one or both regimes will have been operating in "reverse," with the core companies trailing rather than beating the average. But if dominant capital does expand differentially, our framework suggests its expansion will likely occur through one of these regimes rather than both.

3. THE ISRAELI CASE

Turning to the Israeli case, our purpose is first to relate the country's broad political-economic development with the differential profitability of its dominant capital, and then use this framework as a basis for understanding the rise and decline of inflation. The task is hardly simple, though. The main problem is lack of disaggregated data, particularly on the income side. For instance, the national accounts contain no specific breakdown of capital income by category (such as profit, rent and interest), there are no sectoral income accounts, and there is no systematic tracking of production and finances by corporate size. Fortunately, though, these hurdles are not insurmountable. We bypass them by using roundabout estimates, which although inaccurate, are nonetheless sufficiently robust for our purpose of tracing overall trends.

This section begins with a birds-eye decomposition of differential accumulation into breadth and depth. We define Israel's dominant capital as comprising the five largest corporate groups during the mid-1980s: Bank Hapoalim, Bank Leumi, Israel Discount Bank Holding (IDBH), Koor and Clal. The origin of these firms goes back to the early part of the century, and the history of their emergence, consolidation and ongoing restructuring is intimately intertwined with Israel's capitalist development (Frenkel and Bichler, 1984; Rowley, Bichler and Nitzan, 1988).

We approximate three ratios based on equation (2): differential profit, differential employment (breadth) and differential profit per employee (depth). For differential profit, the numerator is given by overall net profit of the core divided by five.5 The denominator, av erage profit per firm, is approximated indirectly. Although there are no national accounting data for overall profit, existing data can be used to derive a useful if rough proxy. This is computed by subtracting from net national income the sum of wages and salaries, corporate taxes and income tax paid by unincorporated businesses. The result is a measure of non-labor income, which includes in addition to net profit also the net income of unincorporated business, rent and interest. Because profit is the most volatile component, its fluctuations (although not levels) are likely to dominate and hence correlate with those of non-labor income as a whole. This aggregate measure of non-labor income is then divided by the overall number of industrial establishments - a proxy for the number of firms yielding a rough index for the economy's average net profit per firm.6

Next, consider breadth measured by differential employment. Ideally, this could be observed directly by dividing average employment in the core by average employment per firm in the corporate universe. Neither of these figures is available, though, and both measures must therefore be approximated indirectly. Here too, we rely on data for industrial establishments. Specifically, we use the ratio between aggregate employment by all large establishments (each with 300 or more employees), and the average employment per establishment in the industrial sector as a whole. Although this ratio is limited to the industrial sector and is based on establishments rather than on firms, its temporal movements should provide a good proxy for changes in differential employment by the core.7 Finally, the depth side, measured by differential profit per employee, is given by dividing differential profit by differential employment.

Figure 1 plots all three indices on a logarithmic scale (smoothed and normalized). The data point to the early 1970s as a watershed in Israeli differential accumulation. Until then, differential depth was dormant, with the core's profit per employee expanding roughly in line with the economy's average. Most of the action was in differential breadth, where the core firms expanded their employment relative to the averages - partly through differential absorption of newly proletarianized workers, but mostly through mergers and acquisitions. Although prior data are unavailable, it is likely that breadth was the major source of differential accumulation through much of the 1950s and 1960s. Rapid economic growth, ushered in first by massive immigration, then by German Holocaust compensation, and finally by the incorporation of the Palestinians into the Israeli economy, meant that the "boundaries" of the corporate universe were constantly expanding. This, combined with the fact that during this initial phase the large companies were still relatively small, implied an ample supply of takeover targets. Inflation in that period was relatively low, suggesting that the struggle over profit per employee was muted. Whatever differential accumulation occurred in that period must hence be attributed largely to breadth.

The situation changed drastically in the early 1970s. Pent-up demand from the earlier immigration wave was receding, the inflow of cheap capital had ended, and the proletarianization of the Palestinians was coming to a close. Moreover, with the core groups already having gained control over much of the economy, further amalgamation was becoming increasingly difficult. And as the breadth regime was reaching a dead end, pressures to shift to depth were building up.

The Inflationary Period of the 1970s and 1980s

Because differential employment growth was negligible during the late 197fls and 1980s, our analysis of that period can safely focus on depth only, so that:

Enlarge 200% Enlarge 400%

The third component of depth - differential growth of output per employee - cannot be examined with available data, although that should matter little for the period at hand. The reason is that during the late 19'70s and 1980s, GDP per capita grew by a meagre 1.4 percent annually, compared with nearly 6 percent during the previous quarter century, suggesting that differential growth in output per employee, if there was any, must have been negligible and can be ignored. For our purpose here, therefore, we ,can simplify further, approximating differential accumulation as a sum of two main components only:

Enlarge 200% Enlarge 400%

(5) DA == Differential Markup Growth + Differential Inflation

This equation, too, cannot be observed directly since there are no separate data on the core's markup and rate of inflation. It can be deciphered, however, if only tentatively, by roundabout means.

Figure 2 shows the relationship between the overall rate of consumer price inflation and our depth index of differential profit per employee. The two series moved tightly together until the late 1980s, and then broke apart. The positive correlation of the earlier period may have both direct and indirect causes. The obvious, direct link is that when overall inflation is on the rise, dominant capital's own rate of inflation rises even faster, culminating in differential inflation (and vice versa when overall inflation is falling). But inflation can also of fect the overall rate of differential accumulation indirectly, even in the absence of differential inflation. This will happen if inflation is positively correlated with differential markup growth, so when inflation rises the core groups are able to raise their profit share of sales faster than the average (and vice versa when inflation falls) . Of course, both mechanism may operate simultaneously, and as our analysis below suggests this seems to have been the case in Israel.

Differential Inflation: Armament and Finance

At first sight, the notion that differential inflation can enhance differential accumulation seems counter-intuitive. If the large core groups raise their prices faster than the average on the depth side, will this not "price them out of the market" by undermining their differential employment growth on the breadth side? The answer is, not necessarily. If these groups were equally present in all sectors, then this might indeed be the case. But in the Israeli reality of the 1970s and 1980s, there were two key sectors where the core groups were probably far more important than elsewhere in the economy: armament and finance.

In both sectors, dominant capital enjoyed something akin to a "near monopoly" - not so much that it literally dominated the entire business, but in the sense of facing little profit competition. Moreover, since the "product" or "service" of these sectors had no close substitutes, their providers confronted practically no intersectoral competition. Under such circumstances, differential inflation could boost depth with a minimal or no impact on breadth.

Consider first the military sector. Figure 3 correlates the overall rate of inflation (based on the GDP deflator), with the relative price of military procurement (the so-called "real price"), measured as the ratio of the GDP deflator for military procurement to the overall GDP deflator. The evidence from the chart is rather remarkable. Until the early 1974s, military price inflation was lower than overall price inflation, causing the relative price of military procurement to fall. One possible reason is mismeasurement, but there could be another, substantive explanation: during that period, military production was not yet privatized, so there was greater concern with productivity improvements, as well as an effort to translate productivity gains into lower relative prices.

Enlarge 200% Enlarge 400%

Enlarge 200% Enlarge 400%

This honeymoon ended by the mid-1970s. With the core groups rapidly displacing state companies as major players in the field, the relative price of military procurement not only stopped falling, but became closely correlated with inflation. The intensification of the Israeli-Arab conflict made the military sector a captured market for the core firms, and with little foreign competition (military imports were financed by U. S. aid), the road for differential inflation was wide open. Since the late 1980s, the situation has reversed. Overall inflation dropped sharply (for reasons to which we return later), putting downward pressure on military price inflation. Indeed, the downward pressure in the military sector was much stronger than elsewhere in the economy: militarization came under growing public scrutiny and calls for domestic spending cuts mounted. Under these political circumstances, the relative pricing power of the core firms deteriorated rapidly, contributing, as the data suggest, to negative differential inflation in that period.

A similar picture emerges from the financial sector. Here, too, the core groups reined supreme, with a government "license" against domestic competition, complemented by capital controls which ef fectively eliminated external competition. Before examining this case, however, an important clarification is in order. The notion of "price" in the financial sector is somewhat ambiguous. Note that financial intermediaries sell not a product, but a service: a "linking" between those who have capital and those who wish to use it.

Seen in this light, financial intermediation is one of the most standardized services, having changed little over the past millennia. The cost of providing it, though, has dropped a great deal. The spread of universal accounting standards, asset markets and new communication technologies must have contributed to a huge productivity improvement. On the face of it, then, the real cost of intermediation should have trended down. Moreover, since we are dealing with relative prices, this tendency should have been independent of inflation.

But that is not what happened in Israel. In Figure 4, we contrast the rate of inflation with the interest rate spread between lending and deposit rates. Expressed in percentage terms, the latter index measures the proportionate "cut" financial institutions take from the money they move between savers and borrowers. In Israel, this spread, instead of remaining fixed or falling, has actually moved up and down with inflation.

Note that according to orthodox theory, such correlation is fully justified. After all, the income of the financial intermediary has to be protected against inflation, so if overall prices rise so must the price of intermediation. But there is a catch here. This is because the nominal amount of deposits and lending is already linked to inflation: the higher the inflation the more money is changing hands. The service provided by the intermediary, on the other hand, is independent of inflation; its inputs and cost are associated not with the nominal amounts of money being moved, but with the number of transactions. The latter tends to rise with inflation, though not by much, particularly in a highly "indexed" economy such as Israel (wage payments, as well as most government transfers and other regular payments, continued to be made once a month regardless of the rate of inflation) . In order for the real price of intermediation to remain unchanged, all it takes therefore is a fixed interest spread.

Yet as we have seen, in Israel the interest rate spread has been tightly correlated with inflation. The consequence is that instead of remaining stable as it should have, real income from intermediation rose more than tenfold (!) during the 1975-1985 period (approximation based on interest rate spreads and the volume of outstanding credit measured in constant prices, both from the Central Bureau of Statistics). Remarkably, this rise occurred while the volume of intermediation remained relatively stable. It was only in the late 19$0s, when inflation started to decline, that intermediation activity began picking up again - but then, instead of increasing, the real income from intermediation actually went down! In other words, we have here a classic case of "Veblenian sabotage," with the income of absentee owners rising as they tighten their grip through "industrial limitation" and stagflation, and falling when their power loosens, growth resumes and inflation subsides.

Enlarge 200% Enlarge 400%

To summarize. The core groups, by dominating the military and financial sectors, were able to enjoy differential inflation during the 1970s and early 1980s. Faced with no meaningful rivalry, they kept their inflation in these sectors higher than the average, without any repercussions on the breadth side. In this way, even with no significant changes in differential markups, inflation for them was a principal engine of differential accumulation.

Because of the leading role of the military and finance in the Israeli economy at the time, differential inflation by the core firms was not only a consequence of overall inflation, but also one of its principal causes. During the 1974s and 1984s, it fuelled an upward price spiral as other income groups tried to keep up. And when these two sectors nearly collapsed - first the financial sector in the early 1980s, and then the military sector a few years later - their lower differential inflation was a major force pulling the overall rate of inflation down. We turn to this issue now.

Differential Redistribution: Wage Income and Capital Gains

The other component of depth - differential markup growth - is considerably more difficult to assess, since there are no markup data for either dominant capital or the average firm. But here too, a roundabout inquiry can still be instructive. Changes in markup are affected largely (although not solely) by the redistributional struggle between workers and owners (the other key participants being gov ernment and foreign owners). The analysis of differential changes in markups therefore requires that we examine this struggle at both the level of dominant capital and that of the average firm.

As illustrated in Figure 5, labor generally lost from inflation.8 When inflation started rising in the early 1970s, the wage share of GDP fell sharply. Some of the lost ground was recovered during the late 1970s, but the gains in pre-tax wages were largely eaten away by higher effective tax rates. And as inflation continued to soar during the early 1980s, labor again was losing. It was only since the mid-1980s, when inflation began to recede, that the net wage share started to rise. Inverted, this relationship suggests that the overall income share of capital was positively related to inflation, although this in itself does not tell us about the markup, which measures specifically the profit share of income. Insight into the latter can be gained by looking at the relationship between inflation and the stock market.

Enlarge 200% Enlarge 400%

From an investor's viewpoint, inflation affects equity prices in two conflicting ways: it raises expected profit, but it also raises the rate of interest used to discount it down to present value, leaving the net impact depending on the balance of the two. During the Gilded Age of the Israeli stock market, between the early 1970s and early 1980s, the first effect clearly dominated. Rising inflation brought soaring interest rates, but the stock market rose even faster, suggesting a spectacular, if only implicit, rise in expected profit. There are two possible reasons for this: one is that investors expected inflation to generate massive earning gains through redistribution, the other is that they were simply manipulated into buying increasingly "overvalued" stocks. We consider each reason in turn.

Figure 6 contrasts the evolution of CPI inflation with that of an index of redistribution between owners and workers. The index is computed as a ratio between the monthly capital gain from investment in equities (calculated as first differences between monthly observations of the general stock market index) and the average monthly pre-tax wage rate. Movements in this index reflect the relative benefit from owning assets as opposed to owning labor power. The chart suggests that until the early 1990s, equity investors tended to benefit relative to workers when inflation was rising, and suffer when it was falling. In other words, inflation helped redistribute earnings not only in favor of capital income in general, but also in favor of profit in particular. Between the late 1970s and mid-1980s, capital gains rose roughly five times faster than the average wage rate. From the investors' viewpoint, at least some of this relative appreciation seems justified - first, given the negative correlation between the wage share and inflation shown in Figure 5, and, second, since the wage bill is larger than total profit, so when it falls, the relative effect on profit is much larger.

The next question concerns specifically the markup of dominant capital. Was the core's own profit share positively correlated with inflation, and if so, was the redistributional impact greater than in the rest of the business sector? In other words, was inflation helping the core groups more than the average firm, hence generating differential markup growth? In principle, inflation could be associated with a differential decline, rather than growth, in the markup (for instance, if workers in the core firms are better able to maintain or even raise their wage share of sales relative to workers in smaller companies, the profit share in the core will fall relative to the average) . Although we cannot know for sure, this was probably not the case in Israel. The reason is that since the early 1970s, dominant capital accounted for an increasing share of the stock market's overall capitalization, with the three largest banks alone making up 44 percent of the total by 1982, up from a mere seven percent a decade earlier (Bejsky et al, 1986, 61). In other words, if the rising stock market indeed reflected a general redistribution in favor of profit as we argue, a growing share of this redistribution must have been enjoyed by the large core firms. We can therefore tentatively conclude that until the mid-1980s, dominant capital benefited from both components of equation (5) - differential inflation as well as differential redistribution.

The major caveat in this analysis is that during this Gilded Age, stocks became increasingly "overvalued" with investors "discounting" an income redistribution far in excess of what was possible in practice. That in itself of course would be hardly unique to Israel; equity prices everywhere are subject to "hype cycles" of excessive optimism and pessimism, driven primarily by self fulfilling expectations about asset prices rather than their underlying profit (Nitzan, 1995; 1996). What makes the Israeli bubble interesting for our purpose here is, first, the mechanism by which hype was inflated and maintained; and, second, the impact this had on differential accumulation.

Enlarge 200% Enlarge 400%

Until 1982, investor hype was fuelled through a system popularly known as "regulation" ( Visut, in Hebrew) . The largest players were the core groups, whose game plan enjoyed the tacit backing of the government and the central bank (Bejsky et al., 1986). Their scheme, gradually articulated into a full-fledged conspiracy, was to convince the mass of middle- and working-class retail investors that they could legitimately expect a pre-determined real capital gain, and that this gain would rise with inflation. And for a long time, the promise was kept and "everybody" seemed to be gaining. But then the gains were hardly even, and what on the surface seemed like a societal panacea was in fact a cover-up for massive redistribution.

The process assumed two principal forms. One was an ongoing re-rating of relative stock prices through "financial alchemy" typical of mature capitalism with fully vendible capital. Thus, in addition to rigging the overall market, the large players (as well as smaller operators) also manipulated individual stocks, generating spectacular three-digit price gains per month, usually without any connection to the company's actual or expected performance.

The other redistributional mechanism worked through public offerings (POs). According to received economic doctrine, the stock market is where entrepreneurs (who make "real investment") meet savers (who supply the funds). Their PO marriage contract, in principle at least, is for the former to take the money of the latter, put it to productive use, and once the investment comes to fruition, split the spoils. Since POs are based on expectations for additional profit, it should be correlated positively with net investment (leaving replacement investment to be financed through depreciation allowances). Intuitive as it may sound, though, this explanation must be missing something. Between 1973 and 1982, for instance, net investment dropped from 21 percent of GDP to less than seven percent, yet the ratio of POs to GDP, instead of falling, rose tenfold! (POs estimates based on first differences in market capitalization deflated by stock prices) .

Impossible? Not really. While POs are indeed driven by expectations for more profit, that in itself does not mandate new productive capacity. In fact, under certain conditions, such as those prevailing in Israel at the time, the only way to increase profit expectations is to curtail productive capacity. During the 1970s and early 1980s, differential accumulation depended on rapid increases (and expected increases) in depth, which in return required limitations on breadth. Any attempt to plow PO funds back into production would have therefore spelled an excess-capacity disaster, and the possible collapse of differential accumulation altogether. It was much better to channel the money back into the financial market.

The eventual 1983 collapse of the stock market demonstrated the extent to which the process was driven by hype, but that hype was itself a powerful mechanism of redistribution. Established capitalists - mostly the core groups - had the market on their side. Systematically rising stock prices acted as an irresistible magnet, luring fresh liquidity from "aspiring capitalists" into POs. This enabled the core firms not only to redistribute existing ownership titles through stock rigging, but also to create and sell new ones through massive dilution. When the crash finally came, dominant capital was bailed out by the government, whereas most of the less fortunate investors found out they owned plenty of "excessive expectations."

From Capital Accumulation to Inflation

So far we have considered the effect of inflation on accumulation, but causality also runs in reverse, with accumulation fueling inflation. Conventional economic thinking explains inflation as a monetary phenomenon of rising liquidity, the famous "too much money chasing too few commodities." This, however, is more a description than an explanation. If the price level is given by the ratio between the total quantity of money and the total volume of commodities, inflation is by definition a process of rising liquidity, and therefore cannot be simultaneously explained by it. The interesting question is why liquidity is rising. The monetarist answer is essentially statist: It is all the policy makers' fault. They are the ones controlling the money supply; they are the ones who allow it to grow excessively; so they are responsible.

There are two problems with this view. The first, common to all statist explanations, is excessive emphasis on human agency. According to this logic, the decline of inflation in the western countries since the mid-1980s was largely the result of governments and central bankers "getting their act together." But then, why did they not come to their senses earlier, and when they finally did see the light, why did it happen collectively? More generally, if inflation results from the whim of policy makers, why does this whim move in long-term cycles? Unless we can answer these questions, inflation remains "indeterminate."

The second problem is that standard definitions of the "quantity of money," such as M1, M2 and so on, depend not only on how much of it is "supplied" by the authorities, but also on how much is "demanded" by economic units. What standard theory tends to overlook is that the latter is predicated on the expectation of return. With the exception of cash, every financial asset - from bank deposits, to bank loans, to bonds to stocks - is valued proportionately to its anticipated future earnings. Over the long run, the expansion of such assets, or capital, can be sustained only with a parallel growth of expected capital income. This, then, is the missing prelude in the monetarist explanation. Instead of:

Enlarge 200% Enlarge 400%

Moreover, the very definition of money as a liability of the financial sector is far too narrow for any "money-based" explanation of inflation. If bank deposits are a source of excess demand, why not bonds or equities? Surely, the latter are similarly liquid and could be used in much the same way to bid up the prices of goods and services.

Figure '7 illustrates the significance of financial assets for Israeli inflation. In it, we chart the annual rate of CPI inflation against the stock market's capitalization, expressed as a percent of GDP and lagged two years. Given that the overall value of shares is a major component of the so-called "financial assets held by the public," the tight correlation between the two series existing until the late 1980s is telling. Because we are using a lagged series, this correlation suggests that during that period, the rising stock market was not only a consequence of inflation, but also one of its major fuels. Thus, if inflation is to be attributed to "wrong policy," the blame should be leveled against the policy of "regulating" the stock market. With net investment needs falling, rising stock prices and the tidal wave of POs flooded the economy with fresh liquidity, propping up inflation. (The situation has changed markedly since the early 1990s, when the resumption of growth helped absorb the liquidity of a soaring market, leaving little fuel for stroking inflation.)

4. INTERPOLATION: FROM CONFLICT INFLATION TO ACCUMULATION

According to mainstream economic theory, inflation is a real monster. "Like diseases, inflation shows different levels of severity," write Samuelson, Nordhaus and McCallum (1988, 240-242) in their introductory textbook, ranging all the way from the "moderate" to the "galloping" to the "hyper." The first strain of the disease is associated only with the moderate cost of "inefficiency"; the second generates "serious economic distortions"; whereas the third, which the authors label alternatively as a "deadly virus" and a "cancer," has practically "nothing good" about it.

Enlarge 200% Enlarge 400%

Limited to the macroeconomic dimension, these characterizations fail to account for the redistributive aspect of inflation in general, and for its role in differential accumulation in particular. From a relative viewpoint, inflation cannot be labeled as universally "good" or "bad": whatever it does to harm the well-being of some, will, by definition and for the same reason, advance the interests of others.

Distributive explanations of inflation are not new, of course. In Latin America, for instance, rapid inflation has often been associated with "structural" and "tug-of war" theories. According to Hirschman (1985), however, this literature remains problematic. It shows that inflation indeed causes redistribution, but fails to explain why the same result cannot be achieved by other, more effective methods. Hirschman's criticism is relevant for our purpose, both for its insight and shortcomings.

The Latin American variety of structural inflation theories emphasized the role of production bottlenecks. Tug-of war explanations are similar, only that instead of economic structure, they focus on the role of aggressive social attitudes. Combining the two, Hirschman offers a synthesis in which an aggressive society, when encountering structural bottlenecks, becomes prone to inflationary spirals, leading to both redistribution and restructuring.

The standard monetarist claim against this thesis is that in the long-run distribution is determined only by "real" economic forces, not by monetary phenomena, and that inflationary redistribution is therefore largely ephemeral. Hirschman agrees that the initial impact of inflation may differ from its final consequences - though that in itself does not make it neutral. The reason is that the monetarist conception of "real" economic forces, because of their belief in perfect competition, is far too narrow. In a non-competitive environment, the balance of political power and strategic economic positions are as "real" as technology and physical resources. In Brazil, for instance, inflation was a principal vehicle of industrialization: combined with a fixed exchange rate, it undermined resource exports while encouraging machinery imports, thus contributing to an irreversible redistribution from the primary to the secondary sector.

Yet, even then, the riddle remains. Although inflation does redistribute income and alters structure, it is hardly the most "efficient" way of doing so. In fact, most other forms of redistribution (such as universal suffrage, shorter work week, pension payments, public medical care, government contracts and subsidies) are far more durable, and once achieved, are more difficult to reverse. Why, then, given the availability of "superior" methods of redistribution, do economic agents continue to use inflation as a vehicle for redistribution?

There are two common answers, both of which are unsatisfactory according to Hirschman. One is that economic agents simply do not recognize this "inefficiency" of inflation. Such misconceptions, though, cannot persist for long, argues Hirschman, since agents quickly learn that inflationary gains are easy to lose. The other explanation is that agents are caught in a "prisoner's dilemma," where infra-group dynamics force them to act as "rational fools," as Amartya Sen put it. This, however, is valid only in a competitive context. When the social arena is instead dominated by large corporations, labor unions and the state, a negotiated redistributional outcome is usually feasible without inflation. The puzzle of "why inflation" therefore persists.

The answer, according to Hirschman, lies not so much in the outcome, as in the struggle itself. His explanation, somewhat reminiscent of our own power theory of accumulation, is that the real purpose of a tug-of war inflation is not higher standards of living, but the very inflicting of damage:

In a situation of high social tension and group antagonism, such behavior could come to be engaged in less for its normally expected material result - additional real income - than because it is enjoyed for its own sake. This means inverting the usual mean-ends relationship; the gratification of intergroup hostility that is obtained by the achievement of a highly inflationary price or wage rise can be the real benefit, to the point where it would not matter if inflation eroded, totally and in short order, the gains achieved. (Hirschman, 1985, 71.)

Inflation, according to Hirschman, is in fact a relatively peaceful valve for social tensions which could otherwise culminate in an open civil war: "With inflation," he states, "each group is able to engage in conflictive behavior and to demonstrate its power and its antagonism to the other groups. From this point it is not far to the conclusion that such demonstration is an important function of inflation, and perhaps its real motive" (71-72).

For all its insights, Hirschman's analysis has one significant weakness: it makes inflation an extra-capitalist phenomenon. After all, inflation is neither new nor unique to Latin America. Arguing that it is the result of vicious people playing an endless "I'll-show you" game against one another, and that this is a substitute for civil war, is hardly a satisfactory theory. Hirschman does not provide evidence that the distribution of income in Latin America was largely unaltered by inflation, nor should we expect this to be the case elsewhere. For example, is it reasonable to say that the redistributional effects of German inflation in the 1920s could have been achieved more effectively by bargaining? Or can we characterize Israeli inflation of the 1970s and 1980s as a substitute for civil war? More generally, if inflation is merely a weapon in an innate conflict, why does it not persist indefinitely, but rather move in long cycles? In our view, Hirschman is right to emphasize predatory instincts, but his insight can be made much more useful as part of a broader theory of capitalism. The theory of differential accumulation could provide such a framework.

Our analysis in this paper argues that inflation is one of several means of differential accumulation. In the Israeli case, inflation started to rise when the conditions underwriting the earlier breadth regime were no longer sustainable. With proletarianization and foreign aid drying up, differential accumulation by dominant capital could continue only through an inflationary expansion of depth. In principle, differential accumulation could be sustained even without a major redistribution from labor to capital, provided that the core could maintain its own rate of inflation faster than the average. The Israeli experience suggests that such differential inflation was indeed paramount, particularly in the military and finance sectors, but that the rate of differential redistribution was probably positive as well.

In this context, politicians and central bankers operate more as moderators than as "policy makers." According to Brenner and Galai (1984), during the 1974s both the government and the banks were in fact forced into regulating the prices of financial assets - the former in order to finance its (exogenously) growing deficit, the latter in order to keep their capital adequacy at internationally accepted levels. From a differential accumulation perspective, though, there was nothing inevitable about this process. As the breadth regime was coming to a close by the late 1960s, the core groups faced two possible scenarios: a shift to depth or collapsing differential accumulation. As it turned out, the first scenario materialized, but there is nothing deterministic about this outcome. A combination of global circumstances and deliberate domestic choices enabled a "smooth" shift into depth, so as to continue differential accumulation by other means. Israel's increasing involvement in the Middle East conflict justified higher military budgets and capital controls, which in turn led to the financial explosion in the bond and stock markets. And with this process in motion, the next logical step was to allow banks to regulate the stock market, so as to prevent the process from getting out of hand. The effect on macroeconomic indicators, as well as on most of the citizenry, may have been negative, but the impact on differential accumulation was clearly positive.

The government's other important role was to maintain the institution of indexation. As we emphasized earlier, inflation is an openended process, in that prices could rise without bound. But as inflation rises so does the risk of "total loss," particularly for workers (consider the impact on the purchasing power of a fixed wage when annual inflation runs at 300 or 400 percent). Although such losses could spell enormous gains for other social groups (as attested by the hyper-inflation of Germany and Austria before the Second World War, or in Russia during the 1990s), they also create systemic risk for the social structure as whole. This is where indexation comes in.

Mainstream economists correctly point out that indexation is a major factor allowing inflation to continue. They are wrong, though, in suggesting that removing indexation will make inflation go away. Inflation exists not because of indexation, but the other way around. The main latent function of indexation is to protect capitalists, which it does indirectly by partly sheltering workers from the vagaries of inflation. This point is often misunderstood and deserves further elaboration. Critiques on the left commonly accuse the Israeli government of fully indexing its obligations to capitalists, while keeping wages only partly indexed. Although correct, this observation is of only secondary significance. The more important service for capital is that indexation smoothes the impact on wages of inflation. The reason is that inflation can never be fully controlled - neither by capitalists nor by governments. Hence, if inflation is to be used as a means of differential accumulation, it needs indexation in order to prevent unanticipated deviations from pushing workers too close to the brink.

Of course, full indexation will make inflation useless as a means of redistribution. So once in place, the trick is to make indexation only partial, and this is where the state comes into the picture. Its role is to manipulate both the extent and timing of indexation, based on the degree to which inflation deviates from its "desirable" path, and on the threat it posses for social stability. This also suggests that the most effective way of ending inflation is not to eliminate indexation, but rather to impose full indexation, as illustrated by the Brazilian Piano Real of 1994.

5. THE END OF INFLATION: TOWARD A NEW BREADTH REGIME

Although inflation could support differential accumulation indefinitely, at least technically, the process is politically unsustainable. In contrast to breadth, which usually comes together with growth and rising standards of living, depth is commonly accompanied by stagnation, and hence by heightened social conflict. For dominant capital, this conflict appears as growing inability to "manage" macroeconomic variables. Moreover, as inflation rises so does the risk that secondary players, for instance large stock-market speculators, might out-maneuver dominant capital.

In the Israeli context, the first cracks in the depth regime appeared in 1982 with the collapse of the "free stocks" (small capitalization shares, whose prices were less tightly rigged), to be followed nine months later by the crash of the banking shares. The consequence was a sharp fall in the ratio of capital gains to wages (Figure 6). To prevent a decummulation meltdown, the government stepped in with a massive bailout plan, which meant further liquefaction of the system and even faster inflation. But capitalist confidence was seriously shaken, and with the middle class badly hurt and no longer willing to hold shares, that was merely sufficient to keep the capital gain/ wage ratio at its new lower level. In parallel, the U. S. elite began demanding an end to the depth regime. One reason was that high spending provided Israeli military contractors with enough of a threshold to enter into main battle systems development. Their U. S. counterparts found this unacceptable, and Israel was duly forced to curtail its domestic military budget (as opposed to arms imports), and abandon its grandiose plans for its own fighter aircraft. Another reason was that the U. S. Administration became concerned that inflation may bring serious social instability. To highlight its resolve, Washington dispatched a team of economists, led by Herbert Stein and Stanley Fischer, to draft a ten-point stabilization program, known as "Herb's 10 Points." The program, secured by a $1.5 billion standby facility to ward off currency speculators, followed the standard IMF template of devaluation, tight monetary and fiscal policy and, of course, a significant wage erosion. Although the plan was presented as a mere "recommendation," the Israeli government was hardly in a position to ignore it.

The result was the 1985 New Economic Plan (NEP). In the decade following the stabilization plan, Israeli economists have often congratulated the government for finally taking a bold stand, swallowing the bitter medicine they recommended all along. From the perspective of differential accumulation, however, the plan was hardly "bold." With both military spending and the stock market in tatters, dominant capital faced a grave risk. For decades, consensus among the elites was bolstered by an ongoing process of differential accumulation - first through breadth and then through a concerted manipulation of depth. But with a melting stock market and the consequent collapse of differential accumulation, this consensus broke down.

It was from this point onward, when it could no longer stand as the central pillar of differential accumulation, that inflation turned into a "public enemy." The most pressing need was to stabilize the profit of dominant capital, and since inflation could no longer do the job, the onus fell on wages. In order to do so, though, it was first necessary to create the proper crisis atmosphere and a consequent need for "sacrifice." Paradoxically, this was not easy. The problem was that for years, workers were repeatedly told that they were the big winners from inflation (through their "excessive" wage demands). But if so, why would they give up this perk now? Clearly, the "cost" had to be re-articulated, and as it turned out there was no shortage of volunteers. Kleiman (1984), for instance, identified five heavy burdens: the wasteful activity associated with frequent re-pricing of products (such as putting new pricing stickers on goods and making new price lists) ; the cost of learning new prices and the loss of human capital embedded in (ephemeral) knowledge of nominal values; the risk of capital loss from financial speculation; the "acute sense of frustration" associated with heightened uncertainty; and last but not least, the "sorrow of lost opportunities" and the cost of mistaken decisions. According to Kleiman's own arithmetic, over the five year period ending in 1982, the societal burden associated with these losses amounted to 3-4 percent of GDP, which, in his opinion was what the public (that is, workers) should be willing to forego in order to have inflation stop dead in its tracks.

And so by the mid-1980s, the preconditions for a regime shift were all in place: a differential accumulation crisis of unprecedented proportions, mounting pressures from the United States to change course, and an ideological front, erected with impressive academic support, on the necessity of "sacrifice." The clear target was labor. The breadth route was closed, and so was differential inflation. The only means of preventing further drops in differential accumulation was differential redistribution through wage erosion.

The centerpiece of the 1985 NEP was price and wage controls. The reason is that while dominant capital knew what it wanted, it could not achieve it on its own. It needed an external moderator, and this is where the state again became crucial. A particular regime of differential accumulation is based on a broad consensus, common habits of thinking and similar modes of behavior. Since these conventions often tend to acquire their own inertia, a transition from one regime to another often requires outside "intervention."

The elite does not see it in these terms, of course. Given the antagonistic appearance of differential accumulation, it rarely acknowledges, even to itself, the cooperative aspects of its endeavor. The Israeli upward price spiral was propagated by a complex set of open and tacit collusions among the big "price makers." This could not be freely admitted, of course, so the blame was duly assigned to "inflationary expectations." But then what are these expectations if not the consensus of the elite itself? To change expectations therefore is to change the outlook of the elite, and this is when an external actor like the government becomes handy.

This logic is evident in a recent assessment of the NEP written by David Brodet ( 1995), who, as an official of the Ministry of Trade and Industry at the time, was responsible for wage and price controls. Orthodox macroeconomic policies, he writes, are necessary but insufficient for taming galloping inflation. The main problem is inertia, and that can be broken only with direct controls. Specifically, agents have to be "convinced" that prices can move down as well as up, which is why he implemented the so-called "pair policy," in which for every price the ministry raised, another was lowered.

But then changing the "price mentality," as Brodet puts it, was largely a means to an end. In our view, the ultimate goal was to alter the "profit mentality." Specifically, price controls were designed, if only implicitly, to show that from now on, redistribution is to be achieved not through price inflation but through wage deflation. And this was the NEP's Achilles' Heel.

Unlike inflation, which leaves prices unbounded on the upside, wage deflation has definite limits, particularly when combined with a highly contractionary "stabilization" policy. And, indeed, the economic recovery of 1985-87 was short lived, and inflation, while low by previous standards, remained at two-digit levels. The government responded by jacking up real interest rates, and growth collapsed once more. Of course, the official culprit was again labor. According to Fischer (1995, 595), things were working fine until workers started to protest their falling income, and then ` just as it was about to win the war, the government signed a most generous contract with the Histadrut [Confederation of Labor Unions] ." Workers again started to spend, he laments, leaving the central bank with little choice but to hit the monetary breaks and kill the recovery.

Seen from a differential accumulation perspective, though, macroeconomic tightening was a dead end from the very start. The substitute for depth is breadth, and indeed, it was only with the immigration boom of the 1990s that things began to move. The "miracle" that followed had little to do with bold politicians and correct policy, however. It was mostly anchored in the collapse of the Soviet Union, which unleashed an influx of over 800,000 new immigrants, an accompanying investment boom, and most importantly, the prospects for joining the globalization process. Together, these processes have finally convinced the Israeli elite (as well as many foreign investors) that a new breadth phase was in the making.

However, such a new phase, if it is to unfold, will have to rely increasingly on expansion outside of Israel. The domestic corporate scene remains highly concentrated, there is little prospect for a new immigration wave, and most of the population has already been proletarianized. For Israel's dominant capital, it is now the outside world where the opportunity for breadth lies. To realize this promise both regionally and globally, the core groups need stability, and so in an ironic historical twist, the same corporations which previously supported the war economy are now the principal defenders of peace.