Why do firms give up the benefit of market transactions in pursuit of further,
presumably
more expensive internalization?
Nitzan Bichler - 2012 - Capital as Power
?
?
?
1926=100
www. bnarchives. net
Average Employment per Firm **
? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? Breadth 337
sixteenfold, whereas overall employment has grown only threefold. As a result of this difference, average employment per firm has dropped by over 75 per cent (note the logarithmic scale). 3
The process wasn't always even. During the two decades between the mid- 1920s and mid-1940s, the number of firms remained relatively stable, first because of the Great Depression, and subsequently due to the Second World War. Consequently, changes in overall employment during that period were reflected more or less fully in the average size of firms, which fell throughout the Depression only to rise rapidly thereafter.
In the longer run, however, this early pattern proved an aberration. As noted, capitalism is subject to strong centrifugal forces, one of which is the inability of business enterprise to control the overall number of independent capitalists on the scene. And indeed, after the war, the number of firms started multiplying again, while their average size trended down more or less contin- uously. Since large-firm employment has increased over the same period, we can safely conclude that overall employment growth served to boost the differential breadth of dominant capital.
The indirect impact of employment growth, operating through depth, is more complex and harder to assess. On the one hand, the multiplicity of small firms keeps their profit per employee low - partly by precluding cooperation and pricing discretion and partly by undermining formal political action. This fact bears positively on the differential depth of dominant capital. At the same time, unruly growth in the number of small firms can quickly degen- erate into excess capacity, threatening to unravel cooperation within domi- nant capital itself. The balance between these conflicting forces is difficult if not impossible to determine.
All in all, then, green-field growth is no panacea for dominant capital. Although the process boosts its differential breadth, it has an indeterminate and possibly negative effect on differential depth. One way to counteract this latter threat is to scare the underlying population with 'overheating' and educate 'policy makers' about the benefits of 'balanced growth' - a
to the IRS, in 2004 roughly 12 per cent of all active corporations had no assets (zero book value) - a bit less than in 1935, when 13 per cent of corporations were in the same position. The long-term stability of this ratio means that the attendant bias need not concern us here.
3 Our measurements here are not strictly comparable: we contrast the number of corporations with overall non-agricultural private employment (that also includes proprietorships and partnerships), rather than with corporate employment only (for which data are not publicly available). However, we can assess the accuracy of this comparison indirectly, by looking at the share of corporations in nonfarm private GDP (using data from the U. S. Bureau of Economic Analysis, Tables 1. 3. 5 and 1. 14). This share rose from 61 per cent in 1929 to 67 per cent in 2007 - a 10 per cent increase. Now, if we assume that relative employment trends roughly track relative GDP trends, the implication is that, over the entire period, corporate employment rose by only 10 per cent more than overall nonfarm employment. Compared to the threefold rise in nonfarm private employment reported in the text, this bias is too small to affect the overall results.
? 338 Accumulation of power
stagnationary recipe that has been applied with considerable success over the past half-century. 4 But the more fundamental solution to the problem is corporate amalgamation.
Mergers and acquisitions
A mystery of finance
Our discussion of amalgamation begins with Figure 15. 2. The chart plots a 'buy-to-build' indicator, expressing the dollar value of mergers and acquisi- tions as a per cent of the dollar value of gross fixed investment. In terms of our own categories, this index corresponds roughly to the ratio between internal and external breadth. (The data sources and method of computing this index are described in the Data Appendix to the chapter. )
The chart illustrates two important processes - one secular, the other cyclical. Secularly, it shows that, over the longer haul, mergers and acquisi- tions indeed have become more important relative to green-field investment
1,000. 0
100. 0
10. 0
1. 0
0. 1
? ? ? ? ? log scale
? ? ? 1999
218%
? ? ? ? ? ? ? ? ? ? ? ? trend growth rate: 3. 4% per annum
? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? Buy-to-Build Indicator
? ? ? ? ? ? 1896
0. 3%
(mergers & acquisitions as a per cent
of gross fixed private domestic investment)
www. bnarchives. net
? ? ? ? ? ? ? ? ? ? ? ? ? 1880 1900
1920 1940 1960 1980 2000 2020
Figure 15. 2 US accumulation: internal vs external breadth Source: See Data Appendix to the chapter.
? 4 For a neo-Marxist analysis of this long-term policy bias, see Steindl (1979).
Breadth 339
(Proposition 3 in Chapter 14). At the end of the nineteenth century, money put into amalgamation was equivalent to less than 1 per cent of green-field investment; a century later, the ratio surpassed 200 per cent. The trend growth rate indicated in the chart suggests that, year in, year out, mergers and acqui- sitions grew 3. 4 percentage points faster than new capacity.
Now, whereas employment associated with new capacity is added by small and large firms alike, amalgamation increases mostly the employment ranks of dominant capital. The net effect of this trend, therefore, is a massive contri- bution to the differential accumulation of large firms. 5
The reasons for this tendency are not at all obvious. Why do firms decide to merge with, or take over other firms? Why has their urge to merge grown stronger over time? And what does this process mean for the broader political economy?
These are not straw-man questions. Mergers baffle the experts. 'Most mergers disappoint', writes The Economist, 'so why do firms keep merging? ' (Anonymous 1998). And the textbooks offer no clear answer. According to one influential manual, mergers remain one of the 'ten mysteries of finance', a riddle for which there are many partial explanations but no overall theory (Brealey et al. 1992: Ch. 36).
The efficiency spin
Needless to say, amalgamation is a real headache for mainstream economics, whose models commonly rely on the assumption of atomistic competition. Alfred Marshall (1920) tried to solve the problem by arguing that firms, however large, are like trees in the forest: eventually they lose their vitality and die out in competition with younger, more vigorous successors.
On its own, though, the forest analogy was not entirely persuasive, if only because incorporation made firms potentially perpetual. So, for the sceptics, Marshall had to offer an add-on. Even if large firms failed to die, he said, and instead grew into a corporate caste, the attendant social inconvenience was still tolerable - because, first, such a caste tended to be benevolent and, second, the political costs were outweighed by the greater economic efficiency of large-scale business enterprise.
The rigorous spin on this justification was provided by Ronald Coase (1937), who stated that the size of firms is largely a matter of 'transaction costs'. Inter-firm transactions, he asserted, are the most efficient since they are subject to market discipline. Unfortunately, such transactions are not free, and therefore they make sense only if their efficiency gains exceed the
5 The effect on relative employment growth is probably somewhat smaller than implied by the dollar figures. First, amalgamated companies often end up shedding some workers, and second, merger and acquisition data include divestitures that reduce rather than raise employment (though only if the acquirer isn't part of dominant capital). Correcting for these qualifications, though, isn't likely to alter the overall trend.
? 340 Accumulation of power
extra cost of carrying them through; otherwise, they should be internalized as intra-firm activity. Using such a calculus, one can then determine the proper 'boundary' of the firm, which, according to Coase, is set at the precise point where 'the costs of organizing an extra transaction within the firm become equal to the costs of carrying out the same transaction by means of an exchange on the open market or the costs of organizing in another firm' (p. 96).
The ideological leverage of this theory proved immense. It implied that if companies such as General Electric, Cisco or Exxon decided to 'internalize' their dealings with other firms by swallowing them up, then that must be socially efficient; and it meant that their resulting size - no matter how big - was necessarily 'optimal' (for instance, Williamson 1985; 1986). In this way, the nonexistence of perfect competition was no longer an embarrassment for neoclassical theory. To the contrary, it was the market itself that determined the right 'balance' between the benefits of competition and corporate size - and what is more, the whole thing was achieved automatically, according to the eternal principles of marginalism.
But then, there is a little glitch in this Nobel-winning spin: it is irrefutable. The problem is, first, that the cost of transactions (relative to not transacting) and the efficiency gains of transactions (relative to internalization) cannot be measured objectively; and, second, that it isn't even clear how to identify the relevant transactions in the first place. This measurement limbo makes marginal transaction costs - much like marginal productivity and marginal utility - unobservable; and with unobservable magnitudes, reality can never be at odds with the theory. 6
For instance, one can use transaction costs to claim that the historical emergence of 'internalized' command economies such as Nazi Germany or the Soviet Union proves that they were more efficient than their market predecessors. The obvious counterargument, which may well be true, is that that these systems were imposed 'from above', driven by a quest for power rather than efficiency. But then, can we not say the exact same thing about the development of oligopolistic capitalism? 7
In fact, if it were only for efficiency, corporations should have become smaller, not larger. According to Coase's theory, technical progress, particu- larly in information and communication, reduces transaction costs, making the market look increasingly appealing and large corporations ever more cumbersome. And, indeed, using this very logic, Francis Fukuyama (1999) has announced the 'death of the hierarchy', while advocates of the 'E-Lance Economy' (as in freelance) have argued that today's corporate behemoths are
6 The literature on 'measuring' transaction costs is reviewed sympathetically by Wang (2007) and Macher and Richman (2008). For a critical assessment, see Buckley and Chapman (1997).
7 For more on the contrast between power and efficiency arguments here, see Knoedler (1995).
? Breadth 341
anomalous and will soon be replaced by small, 'virtual' firms (Malone and Laubacher 1998). So far, though, these predictions seem hopelessly wrong: amalgamation has not only continued, but accelerated, including in the so-called high-technology sector, where transaction costs have supposedly fallen the most.
From efficiency to power
From an efficiency perspective, the relentless growth of large firms is indeed puzzling.
Why do firms give up the benefit of market transactions in pursuit of further, presumably more expensive internalization? Are they not inter- ested in lower costs?
From a power viewpoint, though, the riddle is more apparent than real. Improved technology certainly can reduce the minimum efficient scale of production (MES); and, indeed, today's largest establishments (plants, head offices, etc. ) often are smaller than they were a hundred years ago. However, firms are not production entities but business units; and given that they can own many establishments, their boundary need not depend on production as such. The real issue with corporate size is not efficiency but differential profit, and the key question therefore is whether amalgamation helps firms beat the average - and if so, how?
The conventional wisdom here is that mergers and acquisitions are a disci- plinary form of 'corporate control'. According to writers such as Manne (1965), Jensen and Ruback (1983) and Jensen (1987), managers are often subject to conflicting loyalties, and this conflict may compromise their commitment to profit maximization (the so-called principal-agent problem). The threat of takeover puts these managers back in line, forcing them not only to improve efficiency, but also to translate such efficiency into higher profit and rising shareholders' value.
This argument became popular during the 1980s. The earning yield on US equities fell below the yield on long-term bonds for the first time since the 1940s, and that drop gave corporate 'raiders' the academic justification (if they needed one) for launching the latest and longest merger wave. The logic of the argument, however, was and remains problematic. Mergers may indeed be driven by profit, but that in itself has little to do with productivity gains. Neither is there much evidence that mergers are prompted by inefficiency, or that they make the combined firms more efficient. 8 Indeed, as we suggested in Chapter 12 and argue further below, the latent function of mergers in this regard is not to boost efficiency but to tame it - a task that they achieve by keeping a lid on overall capacity growth. Moreover, there is no clear indica- tion that mergers make the amalgamated firms more profitable than they
8 See for example, Ravenscraft and Scherer (1987), Caves (1989), Bhagat, Shleifer and Vishny (1990) and Kaplan (2000).
? 342 Accumulation of power
were separately - although here the issue is somewhat more complicated and requires some explication.
Two points are worth noting. First, there is a serious methodological diffi- culty. Most attempts to test the effects of mergers on profitability are based on comparing the performance of merged and non-merged companies. 9 While this method may offer some insight in the case of individual firms, it is misleading when applied to dominant capital as a whole. Looking at the amalgamation process in its entirety, the issue is not how it compares with 'doing nothing' (that is, with not amalgamating), but rather how it contrasts with the alternative strategy of green-field investment. Unfortunately, such a comparison is impossible since the very purpose of mergers and acquisitions is to avoid creating new capacity. In other words, amalgamation removes the main evidence against which one can assess its business success.
Perhaps a better, albeit 'unscientific', way to tackle the issue is to answer the following hypothetical question: What would have happened to the prof- itability of dominant capital in the United States if, instead of splitting its investment one third for green-field and two thirds for mergers and acquisi- tions, it were to plough it all back into new capacity? As Veblen (1923) correctly predicted, such a 'free run of production' is not going to happen, so we cannot know for sure. But then the very fact it hasn't happened, together with the century-long tendency to move in the opposite direction, from green- field to amalgamation, already suggests what the answer may be. 10
The second important point concerns the meaning of 'profitability' in this context. Conventional measures, such as earnings-to-price ratio, return on equity, or profit margin on sales, relevant as they may be for investors, are too narrow as indicators of capitalist power - particularly when such power is vested in and exercised by corporations rather than individuals. A more appropriate measure for this power is the distribution and differential growth of profit (and of capitalization more broadly), and from this perspective mergers and acquisitions make a very big difference. By fusing previously distinct earning streams, amalgamation contributes to the organized power of dominant capital, regardless of whether or not it augments the more conven- tional rates of return. In our view, this 'earning fusion', common to all mergers, is also their ultimate reason.
And indeed, by gradually shifting the emphasis from building to buying, from colliding to colluding and from a certain measure of public oversight to increasingly capitalized government, corporate capitalism in the United States and elsewhere has been able not only to lessen the destabilizing impact of green-field cycles pointed out by Marx, but also to reproduce and consoli- date on an ever-growing scale. Instead of collapsing under its own weight,
9 For instances of this approach, see Ravenscraft (1987), Ravenscraft and Scherer (1989), Scherer and Ross (1990: Ch. 5) and Healy, Palepu and Ruback (1992).
10 See also footnote 12 in Chapter 12, for the accumulation consequences of Japan's anti- merger/all-for-growth attitude.
? Breadth 343
capitalism seems to have grown stronger. The broader consequence of this shift has been creeping stagnation (Proposition 3 in Chapter 14); and yet, as Veblen already suggested a century ago, the large accumulators have learned to 'manage' this stagnation for their own ends.
Patterns of amalgamation
Merger waves
Now, this general rationale for merger does not in itself explain the concrete historical trajectory of corporate amalgamation. Mergers and acquisitions grow, but not smoothly, and indeed the second feature evident in Figure 15. 2 is the cyclical pattern of the series (Proposition 4 in Chapter 14).
Looking at the past century, we can identify four amalgamation 'waves'. The first wave, occurring during the transition from the nineteenth to the twentieth century, is commonly referred to as the 'monopoly' wave. The second, lasting through much of the 1920s, is known as the 'oligopoly' wave. The third, building up during the late 1950s and 1960s, is nicknamed the 'conglomerate' wave. And the fourth wave, beginning in the early 1980s, does not yet have a popular title, but based on its all-encompassing nature we can safely label it the 'global' wave.
This wave-like pattern remains something of a mystery. Why do mergers and acquisitions have a pattern at all? Why are they not erratic? Or, alterna- tively, why do they not grow continuously or in line with green-field invest- ment? So far, most attempts to answer these questions have approached the issue from the micro perspective of the firm, which is precisely why they run into a dead end.
Tobin's Q
One of the more famous explanations is based on the work of Tobin and Brainard (1968; 1977), whom we have already met in Chapter 10. The under- pinnings of the argument cannot be simpler: capitalists are rational, and rational actors buy what is cheap. Their yardstick is Tobin's Q: the ratio between the capitalized value of an asset on the market and the price of building it from scratch. When Tobin's Q is smaller than 1, existing capacity is cheaper, so the firm will buy it from others. And when Tobin's Q is greater than 1, new capacity is cheaper, so the firm will construct it anew.
For the corporate universe as a whole, Tobin's Q is the ratio of overall market capitalization to the replacement cost of all outstanding assets; and if individual rationality extends to aggregate rationality, we should observe the buy-to-build indicator moving inversely with Tobin's Q: the less expensive existing assets are relative to newly produced ones, the greater the proportion of 'financial' to 'real' investment should be, and vice versa.
The explanation seems sensible enough, only that reality refuses to abide. Figure 15. 3 shows the relationship between the two magnitudes: our own
344 Accumulation of power
1,000. 0 3. 5
? ? ? ? ? ? ? ? ? log scale
? ? ? ? ? ? ? Buy-to-Build Indicator
? ? ? ? (mergers & acquisitions as a per cent of gross fixed capital formation, left)
? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? Tobin's Q
(ratio of market value of stocks and bonds to the current cost of fixed assets, right)
? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? www. bnarchives. net
? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? 100. 0
10. 0
1. 0
3. 0 2. 5 2. 0 1. 5 1. 0 0. 5 0.
www. bnarchives. net
Average Employment per Firm **
? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? Breadth 337
sixteenfold, whereas overall employment has grown only threefold. As a result of this difference, average employment per firm has dropped by over 75 per cent (note the logarithmic scale). 3
The process wasn't always even. During the two decades between the mid- 1920s and mid-1940s, the number of firms remained relatively stable, first because of the Great Depression, and subsequently due to the Second World War. Consequently, changes in overall employment during that period were reflected more or less fully in the average size of firms, which fell throughout the Depression only to rise rapidly thereafter.
In the longer run, however, this early pattern proved an aberration. As noted, capitalism is subject to strong centrifugal forces, one of which is the inability of business enterprise to control the overall number of independent capitalists on the scene. And indeed, after the war, the number of firms started multiplying again, while their average size trended down more or less contin- uously. Since large-firm employment has increased over the same period, we can safely conclude that overall employment growth served to boost the differential breadth of dominant capital.
The indirect impact of employment growth, operating through depth, is more complex and harder to assess. On the one hand, the multiplicity of small firms keeps their profit per employee low - partly by precluding cooperation and pricing discretion and partly by undermining formal political action. This fact bears positively on the differential depth of dominant capital. At the same time, unruly growth in the number of small firms can quickly degen- erate into excess capacity, threatening to unravel cooperation within domi- nant capital itself. The balance between these conflicting forces is difficult if not impossible to determine.
All in all, then, green-field growth is no panacea for dominant capital. Although the process boosts its differential breadth, it has an indeterminate and possibly negative effect on differential depth. One way to counteract this latter threat is to scare the underlying population with 'overheating' and educate 'policy makers' about the benefits of 'balanced growth' - a
to the IRS, in 2004 roughly 12 per cent of all active corporations had no assets (zero book value) - a bit less than in 1935, when 13 per cent of corporations were in the same position. The long-term stability of this ratio means that the attendant bias need not concern us here.
3 Our measurements here are not strictly comparable: we contrast the number of corporations with overall non-agricultural private employment (that also includes proprietorships and partnerships), rather than with corporate employment only (for which data are not publicly available). However, we can assess the accuracy of this comparison indirectly, by looking at the share of corporations in nonfarm private GDP (using data from the U. S. Bureau of Economic Analysis, Tables 1. 3. 5 and 1. 14). This share rose from 61 per cent in 1929 to 67 per cent in 2007 - a 10 per cent increase. Now, if we assume that relative employment trends roughly track relative GDP trends, the implication is that, over the entire period, corporate employment rose by only 10 per cent more than overall nonfarm employment. Compared to the threefold rise in nonfarm private employment reported in the text, this bias is too small to affect the overall results.
? 338 Accumulation of power
stagnationary recipe that has been applied with considerable success over the past half-century. 4 But the more fundamental solution to the problem is corporate amalgamation.
Mergers and acquisitions
A mystery of finance
Our discussion of amalgamation begins with Figure 15. 2. The chart plots a 'buy-to-build' indicator, expressing the dollar value of mergers and acquisi- tions as a per cent of the dollar value of gross fixed investment. In terms of our own categories, this index corresponds roughly to the ratio between internal and external breadth. (The data sources and method of computing this index are described in the Data Appendix to the chapter. )
The chart illustrates two important processes - one secular, the other cyclical. Secularly, it shows that, over the longer haul, mergers and acquisi- tions indeed have become more important relative to green-field investment
1,000. 0
100. 0
10. 0
1. 0
0. 1
? ? ? ? ? log scale
? ? ? 1999
218%
? ? ? ? ? ? ? ? ? ? ? ? trend growth rate: 3. 4% per annum
? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? Buy-to-Build Indicator
? ? ? ? ? ? 1896
0. 3%
(mergers & acquisitions as a per cent
of gross fixed private domestic investment)
www. bnarchives. net
? ? ? ? ? ? ? ? ? ? ? ? ? 1880 1900
1920 1940 1960 1980 2000 2020
Figure 15. 2 US accumulation: internal vs external breadth Source: See Data Appendix to the chapter.
? 4 For a neo-Marxist analysis of this long-term policy bias, see Steindl (1979).
Breadth 339
(Proposition 3 in Chapter 14). At the end of the nineteenth century, money put into amalgamation was equivalent to less than 1 per cent of green-field investment; a century later, the ratio surpassed 200 per cent. The trend growth rate indicated in the chart suggests that, year in, year out, mergers and acqui- sitions grew 3. 4 percentage points faster than new capacity.
Now, whereas employment associated with new capacity is added by small and large firms alike, amalgamation increases mostly the employment ranks of dominant capital. The net effect of this trend, therefore, is a massive contri- bution to the differential accumulation of large firms. 5
The reasons for this tendency are not at all obvious. Why do firms decide to merge with, or take over other firms? Why has their urge to merge grown stronger over time? And what does this process mean for the broader political economy?
These are not straw-man questions. Mergers baffle the experts. 'Most mergers disappoint', writes The Economist, 'so why do firms keep merging? ' (Anonymous 1998). And the textbooks offer no clear answer. According to one influential manual, mergers remain one of the 'ten mysteries of finance', a riddle for which there are many partial explanations but no overall theory (Brealey et al. 1992: Ch. 36).
The efficiency spin
Needless to say, amalgamation is a real headache for mainstream economics, whose models commonly rely on the assumption of atomistic competition. Alfred Marshall (1920) tried to solve the problem by arguing that firms, however large, are like trees in the forest: eventually they lose their vitality and die out in competition with younger, more vigorous successors.
On its own, though, the forest analogy was not entirely persuasive, if only because incorporation made firms potentially perpetual. So, for the sceptics, Marshall had to offer an add-on. Even if large firms failed to die, he said, and instead grew into a corporate caste, the attendant social inconvenience was still tolerable - because, first, such a caste tended to be benevolent and, second, the political costs were outweighed by the greater economic efficiency of large-scale business enterprise.
The rigorous spin on this justification was provided by Ronald Coase (1937), who stated that the size of firms is largely a matter of 'transaction costs'. Inter-firm transactions, he asserted, are the most efficient since they are subject to market discipline. Unfortunately, such transactions are not free, and therefore they make sense only if their efficiency gains exceed the
5 The effect on relative employment growth is probably somewhat smaller than implied by the dollar figures. First, amalgamated companies often end up shedding some workers, and second, merger and acquisition data include divestitures that reduce rather than raise employment (though only if the acquirer isn't part of dominant capital). Correcting for these qualifications, though, isn't likely to alter the overall trend.
? 340 Accumulation of power
extra cost of carrying them through; otherwise, they should be internalized as intra-firm activity. Using such a calculus, one can then determine the proper 'boundary' of the firm, which, according to Coase, is set at the precise point where 'the costs of organizing an extra transaction within the firm become equal to the costs of carrying out the same transaction by means of an exchange on the open market or the costs of organizing in another firm' (p. 96).
The ideological leverage of this theory proved immense. It implied that if companies such as General Electric, Cisco or Exxon decided to 'internalize' their dealings with other firms by swallowing them up, then that must be socially efficient; and it meant that their resulting size - no matter how big - was necessarily 'optimal' (for instance, Williamson 1985; 1986). In this way, the nonexistence of perfect competition was no longer an embarrassment for neoclassical theory. To the contrary, it was the market itself that determined the right 'balance' between the benefits of competition and corporate size - and what is more, the whole thing was achieved automatically, according to the eternal principles of marginalism.
But then, there is a little glitch in this Nobel-winning spin: it is irrefutable. The problem is, first, that the cost of transactions (relative to not transacting) and the efficiency gains of transactions (relative to internalization) cannot be measured objectively; and, second, that it isn't even clear how to identify the relevant transactions in the first place. This measurement limbo makes marginal transaction costs - much like marginal productivity and marginal utility - unobservable; and with unobservable magnitudes, reality can never be at odds with the theory. 6
For instance, one can use transaction costs to claim that the historical emergence of 'internalized' command economies such as Nazi Germany or the Soviet Union proves that they were more efficient than their market predecessors. The obvious counterargument, which may well be true, is that that these systems were imposed 'from above', driven by a quest for power rather than efficiency. But then, can we not say the exact same thing about the development of oligopolistic capitalism? 7
In fact, if it were only for efficiency, corporations should have become smaller, not larger. According to Coase's theory, technical progress, particu- larly in information and communication, reduces transaction costs, making the market look increasingly appealing and large corporations ever more cumbersome. And, indeed, using this very logic, Francis Fukuyama (1999) has announced the 'death of the hierarchy', while advocates of the 'E-Lance Economy' (as in freelance) have argued that today's corporate behemoths are
6 The literature on 'measuring' transaction costs is reviewed sympathetically by Wang (2007) and Macher and Richman (2008). For a critical assessment, see Buckley and Chapman (1997).
7 For more on the contrast between power and efficiency arguments here, see Knoedler (1995).
? Breadth 341
anomalous and will soon be replaced by small, 'virtual' firms (Malone and Laubacher 1998). So far, though, these predictions seem hopelessly wrong: amalgamation has not only continued, but accelerated, including in the so-called high-technology sector, where transaction costs have supposedly fallen the most.
From efficiency to power
From an efficiency perspective, the relentless growth of large firms is indeed puzzling.
Why do firms give up the benefit of market transactions in pursuit of further, presumably more expensive internalization? Are they not inter- ested in lower costs?
From a power viewpoint, though, the riddle is more apparent than real. Improved technology certainly can reduce the minimum efficient scale of production (MES); and, indeed, today's largest establishments (plants, head offices, etc. ) often are smaller than they were a hundred years ago. However, firms are not production entities but business units; and given that they can own many establishments, their boundary need not depend on production as such. The real issue with corporate size is not efficiency but differential profit, and the key question therefore is whether amalgamation helps firms beat the average - and if so, how?
The conventional wisdom here is that mergers and acquisitions are a disci- plinary form of 'corporate control'. According to writers such as Manne (1965), Jensen and Ruback (1983) and Jensen (1987), managers are often subject to conflicting loyalties, and this conflict may compromise their commitment to profit maximization (the so-called principal-agent problem). The threat of takeover puts these managers back in line, forcing them not only to improve efficiency, but also to translate such efficiency into higher profit and rising shareholders' value.
This argument became popular during the 1980s. The earning yield on US equities fell below the yield on long-term bonds for the first time since the 1940s, and that drop gave corporate 'raiders' the academic justification (if they needed one) for launching the latest and longest merger wave. The logic of the argument, however, was and remains problematic. Mergers may indeed be driven by profit, but that in itself has little to do with productivity gains. Neither is there much evidence that mergers are prompted by inefficiency, or that they make the combined firms more efficient. 8 Indeed, as we suggested in Chapter 12 and argue further below, the latent function of mergers in this regard is not to boost efficiency but to tame it - a task that they achieve by keeping a lid on overall capacity growth. Moreover, there is no clear indica- tion that mergers make the amalgamated firms more profitable than they
8 See for example, Ravenscraft and Scherer (1987), Caves (1989), Bhagat, Shleifer and Vishny (1990) and Kaplan (2000).
? 342 Accumulation of power
were separately - although here the issue is somewhat more complicated and requires some explication.
Two points are worth noting. First, there is a serious methodological diffi- culty. Most attempts to test the effects of mergers on profitability are based on comparing the performance of merged and non-merged companies. 9 While this method may offer some insight in the case of individual firms, it is misleading when applied to dominant capital as a whole. Looking at the amalgamation process in its entirety, the issue is not how it compares with 'doing nothing' (that is, with not amalgamating), but rather how it contrasts with the alternative strategy of green-field investment. Unfortunately, such a comparison is impossible since the very purpose of mergers and acquisitions is to avoid creating new capacity. In other words, amalgamation removes the main evidence against which one can assess its business success.
Perhaps a better, albeit 'unscientific', way to tackle the issue is to answer the following hypothetical question: What would have happened to the prof- itability of dominant capital in the United States if, instead of splitting its investment one third for green-field and two thirds for mergers and acquisi- tions, it were to plough it all back into new capacity? As Veblen (1923) correctly predicted, such a 'free run of production' is not going to happen, so we cannot know for sure. But then the very fact it hasn't happened, together with the century-long tendency to move in the opposite direction, from green- field to amalgamation, already suggests what the answer may be. 10
The second important point concerns the meaning of 'profitability' in this context. Conventional measures, such as earnings-to-price ratio, return on equity, or profit margin on sales, relevant as they may be for investors, are too narrow as indicators of capitalist power - particularly when such power is vested in and exercised by corporations rather than individuals. A more appropriate measure for this power is the distribution and differential growth of profit (and of capitalization more broadly), and from this perspective mergers and acquisitions make a very big difference. By fusing previously distinct earning streams, amalgamation contributes to the organized power of dominant capital, regardless of whether or not it augments the more conven- tional rates of return. In our view, this 'earning fusion', common to all mergers, is also their ultimate reason.
And indeed, by gradually shifting the emphasis from building to buying, from colliding to colluding and from a certain measure of public oversight to increasingly capitalized government, corporate capitalism in the United States and elsewhere has been able not only to lessen the destabilizing impact of green-field cycles pointed out by Marx, but also to reproduce and consoli- date on an ever-growing scale. Instead of collapsing under its own weight,
9 For instances of this approach, see Ravenscraft (1987), Ravenscraft and Scherer (1989), Scherer and Ross (1990: Ch. 5) and Healy, Palepu and Ruback (1992).
10 See also footnote 12 in Chapter 12, for the accumulation consequences of Japan's anti- merger/all-for-growth attitude.
? Breadth 343
capitalism seems to have grown stronger. The broader consequence of this shift has been creeping stagnation (Proposition 3 in Chapter 14); and yet, as Veblen already suggested a century ago, the large accumulators have learned to 'manage' this stagnation for their own ends.
Patterns of amalgamation
Merger waves
Now, this general rationale for merger does not in itself explain the concrete historical trajectory of corporate amalgamation. Mergers and acquisitions grow, but not smoothly, and indeed the second feature evident in Figure 15. 2 is the cyclical pattern of the series (Proposition 4 in Chapter 14).
Looking at the past century, we can identify four amalgamation 'waves'. The first wave, occurring during the transition from the nineteenth to the twentieth century, is commonly referred to as the 'monopoly' wave. The second, lasting through much of the 1920s, is known as the 'oligopoly' wave. The third, building up during the late 1950s and 1960s, is nicknamed the 'conglomerate' wave. And the fourth wave, beginning in the early 1980s, does not yet have a popular title, but based on its all-encompassing nature we can safely label it the 'global' wave.
This wave-like pattern remains something of a mystery. Why do mergers and acquisitions have a pattern at all? Why are they not erratic? Or, alterna- tively, why do they not grow continuously or in line with green-field invest- ment? So far, most attempts to answer these questions have approached the issue from the micro perspective of the firm, which is precisely why they run into a dead end.
Tobin's Q
One of the more famous explanations is based on the work of Tobin and Brainard (1968; 1977), whom we have already met in Chapter 10. The under- pinnings of the argument cannot be simpler: capitalists are rational, and rational actors buy what is cheap. Their yardstick is Tobin's Q: the ratio between the capitalized value of an asset on the market and the price of building it from scratch. When Tobin's Q is smaller than 1, existing capacity is cheaper, so the firm will buy it from others. And when Tobin's Q is greater than 1, new capacity is cheaper, so the firm will construct it anew.
For the corporate universe as a whole, Tobin's Q is the ratio of overall market capitalization to the replacement cost of all outstanding assets; and if individual rationality extends to aggregate rationality, we should observe the buy-to-build indicator moving inversely with Tobin's Q: the less expensive existing assets are relative to newly produced ones, the greater the proportion of 'financial' to 'real' investment should be, and vice versa.
The explanation seems sensible enough, only that reality refuses to abide. Figure 15. 3 shows the relationship between the two magnitudes: our own
344 Accumulation of power
1,000. 0 3. 5
? ? ? ? ? ? ? ? ? log scale
? ? ? ? ? ? ? Buy-to-Build Indicator
? ? ? ? (mergers & acquisitions as a per cent of gross fixed capital formation, left)
? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? Tobin's Q
(ratio of market value of stocks and bonds to the current cost of fixed assets, right)
? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? www. bnarchives. net
? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? 100. 0
10. 0
1. 0
3. 0 2. 5 2. 0 1. 5 1. 0 0. 5 0.
