Tangible assets are bought and sold on the market and
therefore
have a universal price.
Nitzan Bichler - 2012 - Capital as Power
On top of each of the Microsoft bars, we denote the per cent ratio of Microsoft relative to GM.
The two sets of bars on the left present data on the 'material' operations of the two firms. In terms of relative employment, depicted by the first set, GM is a giant and Microsoft is a dwarf. In 2005, GM had 335,000 workers, 5. 5 times more than Microsoft's 61,000. The second set of bars denotes the respective dollar value of the companies' plant and equipment, measured in historical cost. In line with our concession, we assume that these dollar values are proportionate to the 'productive capacity' of the two companies. According to these statistics, in 2005, GM's 'productive capacity', standing at $78 billion, was 33 times larger than Microsoft's, whose capital goods were worth a mere $2. 3 billion.
The two sets of bars on the right show the companies' respective capital- ization. Here the picture is exactly the opposite, with Microsoft being the giant and GM the dwarf. In 2005, Microsoft's equity had a market value of $283 billion, nearly 26 times GM's $11 billion. And even if we take the sum of debt and market value (which supposedly stands as the total claim on a
500
400
300
200
100
0
Fiction, mirror or distortion? 173
? ? ? ? ? ? ? GM
Microsoft
? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? 18%
2,583%
64%
? ? ? ? ? ? 3%
www. bnarchives. net
? ? ? ? ? ? ? ? ? ? ? Employees (000)
Plant and Equipment ($bn)
Market Value ($bn)
Market Value and Debt ($bn)
Figure 10. 1 General Motors versus Microsoft, 2005
Note: The per cent figures indicate, for any given measure, the size of Microsoft relative to GM.
Source: Compustat through WRDS (series codes: data29 for employees; data8 for net plant and equipment; data24 for price; data54 for common shares outstanding; data181 for total liabilities).
company's productive assets), the GM total of $475 billion was only 55 per cent greater than Microsoft's $306 billion - a far cry from its relatively huge work force and massive plant and equipment.
The usual response to such a discrepancy, from Alfred Marshall onward, points to 'technology' and 'human capital'. This is the 'knowledge economy', tells us Peter Drucker (1969: Ch. 12). Obviously, Microsoft's dispropor- tionate market value must be due to its superior know-how, packed as 'intan- gibles'. And since intangibles are not included in the plant and equipment category of corporate balance sheets on the one hand yet bear on capitaliza- tion on the other, we end up with a market value that deviates, often consid- erably, from the tangible capital stock.
This is a popular academic claim, and for a good reason: it is entirely reversible and totally irrefutable. To illustrate, simply consider the reverse assertion - namely that GM has more know-how than Microsoft. Since nobody knows how to quantify technology, how can we decide which of the two claims is correct?
174 Capitalization
Tobin's Q: adding intangibles
The discrepancy between capitalization and real assets is by no means limited to individual firms or particular time periods. In fact, it appears to be the rule rather than the exception. Consider Figure 10. 2, which plots the so-called Tobin's Q ratio for US corporate sector from 1932 to 2006. 3 In this figure, Tobin's Q measures the ratio between corporate capitalization and capital goods: for each year, the series takes the market value of all outstanding corporate stocks and bonds and divides it by the current replacement cost of corporate fixed assets (plant and equipment). Since both magnitudes are denominated in current prices, the ratio between them is a pure number.
3. 5
3. 0
2. 5
2. 0
1. 5
1. 0
0. 5
0. 0
1930 1940 1950
1960 1970 1980 1990
2000 2010 2020
? ? ? ? ? ? ? Tobin's Q
(ratio of the market value of corporate equities & bonds to the current cost of corporate fixed assets)
2. 8
? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? mean = 1. 24
? ? ? ? ? ? ? ? ? ? ? ? 0. 6
? ? ? www. bnarchives. net
? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? Figure 10. 2 Tobin's Q in the United States
Note: The market value of corporate equities and bonds is net of foreign holdings by US residents.
Source: U. S. Bureau of Economic Analysis through Global Insight (series codes: FAPNREZ for current cost of corporate fixed assets). The market value of corporate equities and bonds splices series from the following two sources. 1932-1951: Global Financial Data (market value of corpo- rate stocks and market value of bonds on the NYSE). 1952-2007: Federal Reserve Board through Global Insight (series codes: FL893064105 for market value of corporate equities; FL263164003 for market value of foreign equities held by US residents; FL893163005 for market value of corporate and foreign bonds; FL263163003 for market value of foreign bonds held by US residents).
3 The Q-ratio was proposed by James Tobin and William Brainard (1968; 1977) as part of their analysis of government stabilization and growth policies.
? Fiction, mirror or distortion? 175
Here too we uphold our theoretical concession. We assume that Fisher's symmetry between real assets and capitalization, although failing the materi- alistic test, can still hold in nominal space. Now, if this assumption were true to the letter, Tobin's Q should have been 1. One dollar's worth of 'real assets' would create a definite future flow of money income, and that flow, once discounted, would in turn generate one dollar's worth of market capitaliza- tion. The facts, though, seem to suggest otherwise.
There are two evident anomalies. First, the historical mean value of the series is not 1, but 1. 24. Second, the actual value of Tobin's Q fluctuates heavily - over the past 75 years it has oscillated between a low of 0. 6 and a high of 2. 8. Moreover, the fluctuations do not look random in the least; on the contrary, they seem fairly stylized, moving in a wave-like fashion. Let's inspect these anomalies in turn.
Why is the long-term average of Tobin's Q higher than 1? The conventional answer points to mismeasurment. To reiterate, fixed assets consist of plant and equipment; yet, as we have already seen in the case of Microsoft vs GM, capitalization accounts for more than just plant and equipment. And since Tobin's Q measures the ratio between the whole and one of its parts, plain arithmetic tells us the result must be bigger than 1. But, then, how much bigger? Even if we accept that there is mismeasurement here, the question remains as to why Tobin's Q should average 1. 24, rather than 1. 01 or 20 for instance. And here, too, just like in the case of Microsoft vs GM, the answer is elusive.
To pin down the difficulty, let's examine the structure of a balance sheet a bit more closely. Recall that corporations have two types of assets: tangible and intangible. 4 According to the neoclassical system of classification, tan- gible assets consist of capital goods - machines, structures and recently also software. Intangible assets, by contrast, represent knowledge, technology, organization, goodwill and other metaphysical entities. Mainstream econo- mists consider both types of assets productive, and the accountants concur - but with a reservation. Although tangible and intangible assets are both 'real', they cannot always be treated in the same way.
The reason is prosaic.
Tangible assets are bought and sold on the market and therefore have a universal price. Since the market is assumed to know all, this price is treated as an objective quantity and hence qualifies for inclusion in the balance sheet. By contrast, most intangible assets are produced by the firm itself. They are generated through internal R&D spending, in-house advertisement expenditures and sundry other costs associated with the likes of 'corporate re-engineering' and 'structural re-organisation'. These are not arm's-length transactions. They are not subject to the universalizing discipline
4 For mainstream analyses of intangibles, see for example Lev (2001) and Corrado, Hulten and Sichel (2006).
? 176 Capitalization
of the market, and therefore the intangible assets they generate lack an 'objective' price. And items that do not have an agreed-upon quantity, no matter how productive, cannot make it into the balance sheet. The best the accountants can do is to list them as current expenditures on the income statement.
There are two exceptions to the rule, though. One exception is when companies purchase pre-packaged intangibles directly through the market - for instance, by acquiring a franchise, patent, trademark, or copyright. The other is when one corporation acquires another at a price that exceeds the acquired company's book value. Since the merger does not create new tangible assets, the accountants assume that the premium must represent the intangible assets of the new formation. They also assume that since this premium is determined by the market, it must be objective. And given that the intangibles are objectively measured, the accountants feel safe enough to include them in the balance sheet.
So all in all we have three categories: (1) tangible assets that are included in the balance sheet, (2) intangible assets that are included in the balance sheet, and (3) intangible assets that are not included in the balance sheet. Now, as noted, fixed assets comprise only the first category, whereas capitalization reflects the sum of all three, and according to the conventional creed it is this mismatch that explains why the long-term average of Tobin's Q differs from 1.
The historical rationale goes as follows. Over the past several decades, US- based corporations have undergone an 'intangible revolution'. Their economy has become 'high-tech', with knowledge, information and commu- nication all multiplying manifold. As a consequence of this revolution, the growth of tangible assets decelerated, while that of intangible assets acceler- ated. And how do we know the extent of this divergence? Simple, say the neoclassicists. Subtract from the market value of firms the market price of their fixed assets, and then assume that, since the market knows all, the differ- ence equals the quantity of intangibles.
Using this standard method, a recent study of the S&P 500 companies esti- mates that, over the past thirty years, the ratio between their market value and the book value of their tangible assets has risen more than fourfold: from 1. 2 in 1975 to 5 in 2005 (Cardoza et al. 2006). The increase implies that in 1975 intangibles amounted to 17 per cent of the total assets, whereas in 2005 they accounted for as much as 80 per cent. Much of this increase is attributed to the growth of out-of-balance-sheet intangibles, whose share of market capitalization during the period is estimated to have risen from 15 to 65 per cent.
Conclusion: the 1. 24 mean value of Tobin's Q is hardly a mystery. It is simply another 'measure of our ignorance' - in this case, our inability to measure intangibles directly. Fortunately, the problem can be circumvented easily by indirect imputation. And, indeed, looking at Figure 10. 2, we can see that much of the increase in Tobin's Q occurred over the past couple of
Fiction, mirror or distortion? 177
decades - coinciding, as one would expect, with the upswing of the 'intangible revolution'.
This rationale may sound soothing to neoclassical ears, but accepting it must come with some unease. To begin with, the neoclassicists don't really 'measure' intangibles; rather, they deduce them, like the ether, as a residual. Moreover, according to their own imputations, this residual accounts for as much as 80 per cent of total market value. To accept this magnitude as a fact is to make the 'material' basis of the theory (shaky as it is) account for no more than 20 per cent of market capitalization - hardly an impressive achievement for a theory that calls itself 'mainstream'. Moreover, the impu- tation method itself doesn't seem very robust. Given that the quantity of intangibles is equal to the difference between market value and tangible assets, oscillations in market value imply corresponding variations in intan- gible assets. But, then, why would the quantity of a productive asset, no matter how intangible, fluctuate - and often wildly - even from one day to the next?
Boom and bust: adding irrationality
The solution to the latter riddle is to invoke irrationality. In this augmented neoclassical version, capitalized market value consists of not two compo- nents, but three: in addition to tangible and intangible assets, it also includes an amount reflecting the excessive optimism or pessimism of investors. And this last component, goes the argument, serves to explain the second anomaly of Tobin's Q - namely its large historical fluctuations.
This irrationality rationale is illustrated in Figure 10. 3. To explain it, let's backtrack and refresh the basics of rational economics. During good times, goes the argument, capitalist optimism causes investors to plough back more profits into productive assets. During bad times, the process goes in reverse, with less profit earmarked for that purpose. As a result, the growth of 'real' assets tends to accelerate in an upswing and decelerate in a downswing. This standard pattern is illustrated by the thick line in the figure. The data measure the rate of change of the current cost of corporate fixed assets (the denomi- nator of Tobin's Q), with the series smoothed as a 10-year moving average in order to accentuate its long-term pattern. According to the figure, the US corporate sector has gone through two very long 'real' accumulation cycles (measured in price terms), the first peaking in the early 1950s, the second in the early 1980s.
The vigilant reader will note that the accumulation process here reflects only the tangible assets - for the obvious reason that the intangible ones cannot be observed directly. But this deficiency shouldn't be much of a concern. Since neoclassical economists view intangible and tangible assets as serving the same productive purpose, they can assume (although not prove) that their respective growth patterns, particularly over long periods of time,
178 Capitalization 30
per cent
25 20 15 10
5
0 -5 -10 -15 -20
Current Cost of Corporate Fixed Assets (annual % change)
Hypothetical Market Value
of Corporate Equities & Bonds based on Neoclassical Orthodoxy Augmented by 'Market Abberations' * (annual % change)
1970 1980 1990 2000 2010 2020
? ? ? ? ? www. bnarchives. net
? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? 1930 1940 1950 1960
Figure 10. 3 The world according to the scriptures
* The value for each year is computed in two steps: first, by calculating the deviation of the growth rate of the (smoothed) 'real' series from its historical mean; and, second, by adding 2. 5 times the value of the deviation to the historical mean.
Note: Series are smoothed as 10-year moving averages.
Source: U. S. Bureau of Economic Analysis through Global Insight (series codes: FAPNREZ for current cost of corporate fixed assets).
are more or less similar. 5 So all in all, we could take the thick line as repre- senting the overall accumulation rate of 'real' capital, both tangible and intan- gible (denominated in current dollars terms to bypass the impossibility of 'material' quantities).
Now this is where irrationality comes in. In an ideal neoclassical world - perfectly competitive, fully transparent and completely informed - Fisher's 'capital wealth' and 'capital value' would be the same. Capitalization on the stock and bond markets would exactly equal the dollar value of tangible and intangible assets. The two sums would grow and contract together, moving
5 If, as neoclassicists seem to believe, the trend growth rate of intangibles is faster than that of tangibles, then the overall growth rate of 'real' assets (tangible and intangible) would gradu- ally rise above the growth rate of tangible assets only illustrated in Figure 10. 3. However, since the cyclical pattern would be more or less the same, this possibility has no bearing on our argument.
? Fiction, mirror or distortion? 179
up and down as perfect replicas. But even the neoclassicists realize that this is a mere ideal.
Ever since Newton, we know that pure ideas may be good for predicting the movement of heavenly bodies, but not the folly of men. Newton learned this lesson the hard way after losing plenty of money in the bursting of the 'South Sea Bubble'. Two centuries later he was joined by no other than Irving Fisher, who managed to sacrifice his own fortune - $10 million then, $100 million in today's prices - on the altar of the 1929 stock market crash.
So just to be on the safe side, neoclassicists now agree that, although capi- talization does reflect the objective processes of the 'real economy', the picture must be augmented by human beings. And the latter, sadly but truly, are not always rational. Greed and fear cloud their vision, emotions upset their calculations and passion biases their decisions - distortions that are further amplified by government intervention and regulation, lack of trans- parency, insider trading and other such unfortunate imperfections. All of these deviations from the pure model lead to irrationality and end in mis- priced assets.
But not all is lost. Convention has it that there is nonetheless order in the chaos, a certain rationality in the irrationality. The basic reason is that greed tends to operate mostly on the upswing, whereas fear usually sets in in the downswing. 'We tend to label such behavioural responses as non rational', explains Alan Greenspan (2008), 'But forecasters' concerns should be not whether human response is rational or irrational, only that it is observable and systematic'. The regularity puts limits on the irrationality; limits imply predict- ability; and predictability helps keep the faith intact and the laity in place.
The boundaries of irrationality are well known and can be recited even by novice traders. The description usually goes as follows. In the upswing, the growth of investment in productive assets fires up the greedy imagination of investors, causing them to price financial assets even higher. To illustrate, during the 1990s developments in 'high-tech' hardware and software suppos- edly made investors lose sight of the possible. The evidence: they capitalized information and telecommunication companies, such as Amazon, Ericsson and Nortel, far above the underlying increase in their 'real' value. A similar scenario unfolded in the 2000s. Investors pushed real-estate capitalization, along with its various financial derivatives and structured investment vehi- cles, to levels that far exceeded the underlying 'actual' wealth. The process, which neoclassicists like to think of as a 'market aberration', led to undue 'asset-price inflation'. Naturally, the capitalization created by such 'bouts of insanity' is mostly 'fake wealth'. It represents 'fictitious value' and leads to inevitable 'bubbles'. 6 But there is nonetheless a clear positive relationship here: the irrational growth of 'fake wealth', although excessive, moves in the same direction as the rational growth of 'real wealth'.
? 6 For a typical analysis of 'bubbles', complete with the above jargon, see Janszen (2008).
180 Capitalization
The process is said to invert during a bust. This is where fear kicks in. The 'real' economy decelerates, but investors, feeling as if the sky is falling, bid down asset prices far more than implied by the 'underlying' productive capacity. A famous illustration is offered by the Great Depression. During the four years from 1928 to 1932, the dollar value of corporate fixed assets contracted by 20 per cent, while the market value of equities collapsed by an amplified 70 per cent (we have no aggregate figures for bonds). A similar 'undershooting' occurred during the 1997 Asian financial crisis, with market value contracting by 50 per cent in many cases, against a growth slowdown or a very moderate decline in the dollar value of the 'real' capital stock. Yet here, too, the relationship is clear: the irrational collapse of 'fictitious value', however exaggerated, moves together with the rational deceleration of 'productive wealth'.
This bounded irrationality is illustrated by the thin line in Figure 10. 3. Note that this series is a hypothetical construct. It describes what the growth of capitalization might look like when neoclassical orthodoxy is augmented by 'irrationality' and 'market aberrations'. The value for each year in the hypothetical series is computed in two steps. First, we calculate the deviation of the growth rate of the (smoothed) 'real' series from its historical mean (so if the smoothed growth rate during the year is 8 per cent and the historical mean rate is 6. 7 per cent, the deviation is 1. 3 per cent). Second, we add 2. 5 times the value of the deviation to the historical mean (so in our example, the hypothetical smoothed growth rate would be 2. 5 * 1. 3 + 6. 7 = 9. 95 per cent). The coefficient of 2. 5 is purely arbitrary. A larger or smaller coefficient would generate a larger or smaller amplification, but the cyclical pattern would remain the same.
This simulation solves the riddle of the fluctuating Tobin's Q. It shows how, due to market imperfections and investors' irrationality, the growth of capitalization overshoots 'real' accumulation on the upswing, therefore causing Tobin's Q to rise, and undershoots it on the downswing, causing Tobin's Q to decline.
And so everything falls into place. Tobin's Q averages more than 1 due to an invisible, yet very real intangible revolution. And it fluctuates heavily - admittedly because the market is imperfect and humans are not always rational - but these oscillations are safely bounded and pretty predicable. Capitalization indeed deviates from the 'real' assets, though in the end it always reverts back to the 'fundamentals'.
Or does it?
The gods must be crazy
It turns out that while the neoclassical priests were busy fortifying the faith, the gods were having fun with the facts. The result is illustrated in Figure 10. 4 (where both series again are smoothed as 10-year moving averages). The thick line, as in Figure 10. 3, shows the rate of change of corporate fixed assets
Fiction, mirror or distortion? 181
measured in current replacement cost. But the thin line is different. Whereas in Figure 10. 3 this line shows the rate of growth of capitalization stipulated by the theory, here it shows the actual rate of growth as it unfolded on the stock and bond markets. And the difference couldn't have been starker.
The gyrations of capitalization, instead of amplifying those of 'real' assets, move in exactly the opposite direction. It is important to note that we are dealing here not with short-term fluctuations of the business cycle, but with very long-term waves of roughly 30-year duration. Furthermore, the pattern seems anything but accidental. In fact, it is rather systematic: whenever the growth rate of 'real' assets decelerates, the growth rate of capitalization accel- erates, and vice versa. 7
This reality puts the world on its head. One could perhaps concede that 'real' assets do not have a material quantum - yet pretend, as we have agreed
25
20
15
10
5
0
-5
-10
1930 1940
1950 1960
1970 1980
1990 2000 2010 2020
? ? ? ? per cent
? ? Market Value of Corporate Equities & Bonds (annual % change)
? ? ?
The two sets of bars on the left present data on the 'material' operations of the two firms. In terms of relative employment, depicted by the first set, GM is a giant and Microsoft is a dwarf. In 2005, GM had 335,000 workers, 5. 5 times more than Microsoft's 61,000. The second set of bars denotes the respective dollar value of the companies' plant and equipment, measured in historical cost. In line with our concession, we assume that these dollar values are proportionate to the 'productive capacity' of the two companies. According to these statistics, in 2005, GM's 'productive capacity', standing at $78 billion, was 33 times larger than Microsoft's, whose capital goods were worth a mere $2. 3 billion.
The two sets of bars on the right show the companies' respective capital- ization. Here the picture is exactly the opposite, with Microsoft being the giant and GM the dwarf. In 2005, Microsoft's equity had a market value of $283 billion, nearly 26 times GM's $11 billion. And even if we take the sum of debt and market value (which supposedly stands as the total claim on a
500
400
300
200
100
0
Fiction, mirror or distortion? 173
? ? ? ? ? ? ? GM
Microsoft
? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? 18%
2,583%
64%
? ? ? ? ? ? 3%
www. bnarchives. net
? ? ? ? ? ? ? ? ? ? ? Employees (000)
Plant and Equipment ($bn)
Market Value ($bn)
Market Value and Debt ($bn)
Figure 10. 1 General Motors versus Microsoft, 2005
Note: The per cent figures indicate, for any given measure, the size of Microsoft relative to GM.
Source: Compustat through WRDS (series codes: data29 for employees; data8 for net plant and equipment; data24 for price; data54 for common shares outstanding; data181 for total liabilities).
company's productive assets), the GM total of $475 billion was only 55 per cent greater than Microsoft's $306 billion - a far cry from its relatively huge work force and massive plant and equipment.
The usual response to such a discrepancy, from Alfred Marshall onward, points to 'technology' and 'human capital'. This is the 'knowledge economy', tells us Peter Drucker (1969: Ch. 12). Obviously, Microsoft's dispropor- tionate market value must be due to its superior know-how, packed as 'intan- gibles'. And since intangibles are not included in the plant and equipment category of corporate balance sheets on the one hand yet bear on capitaliza- tion on the other, we end up with a market value that deviates, often consid- erably, from the tangible capital stock.
This is a popular academic claim, and for a good reason: it is entirely reversible and totally irrefutable. To illustrate, simply consider the reverse assertion - namely that GM has more know-how than Microsoft. Since nobody knows how to quantify technology, how can we decide which of the two claims is correct?
174 Capitalization
Tobin's Q: adding intangibles
The discrepancy between capitalization and real assets is by no means limited to individual firms or particular time periods. In fact, it appears to be the rule rather than the exception. Consider Figure 10. 2, which plots the so-called Tobin's Q ratio for US corporate sector from 1932 to 2006. 3 In this figure, Tobin's Q measures the ratio between corporate capitalization and capital goods: for each year, the series takes the market value of all outstanding corporate stocks and bonds and divides it by the current replacement cost of corporate fixed assets (plant and equipment). Since both magnitudes are denominated in current prices, the ratio between them is a pure number.
3. 5
3. 0
2. 5
2. 0
1. 5
1. 0
0. 5
0. 0
1930 1940 1950
1960 1970 1980 1990
2000 2010 2020
? ? ? ? ? ? ? Tobin's Q
(ratio of the market value of corporate equities & bonds to the current cost of corporate fixed assets)
2. 8
? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? mean = 1. 24
? ? ? ? ? ? ? ? ? ? ? ? 0. 6
? ? ? www. bnarchives. net
? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? Figure 10. 2 Tobin's Q in the United States
Note: The market value of corporate equities and bonds is net of foreign holdings by US residents.
Source: U. S. Bureau of Economic Analysis through Global Insight (series codes: FAPNREZ for current cost of corporate fixed assets). The market value of corporate equities and bonds splices series from the following two sources. 1932-1951: Global Financial Data (market value of corpo- rate stocks and market value of bonds on the NYSE). 1952-2007: Federal Reserve Board through Global Insight (series codes: FL893064105 for market value of corporate equities; FL263164003 for market value of foreign equities held by US residents; FL893163005 for market value of corporate and foreign bonds; FL263163003 for market value of foreign bonds held by US residents).
3 The Q-ratio was proposed by James Tobin and William Brainard (1968; 1977) as part of their analysis of government stabilization and growth policies.
? Fiction, mirror or distortion? 175
Here too we uphold our theoretical concession. We assume that Fisher's symmetry between real assets and capitalization, although failing the materi- alistic test, can still hold in nominal space. Now, if this assumption were true to the letter, Tobin's Q should have been 1. One dollar's worth of 'real assets' would create a definite future flow of money income, and that flow, once discounted, would in turn generate one dollar's worth of market capitaliza- tion. The facts, though, seem to suggest otherwise.
There are two evident anomalies. First, the historical mean value of the series is not 1, but 1. 24. Second, the actual value of Tobin's Q fluctuates heavily - over the past 75 years it has oscillated between a low of 0. 6 and a high of 2. 8. Moreover, the fluctuations do not look random in the least; on the contrary, they seem fairly stylized, moving in a wave-like fashion. Let's inspect these anomalies in turn.
Why is the long-term average of Tobin's Q higher than 1? The conventional answer points to mismeasurment. To reiterate, fixed assets consist of plant and equipment; yet, as we have already seen in the case of Microsoft vs GM, capitalization accounts for more than just plant and equipment. And since Tobin's Q measures the ratio between the whole and one of its parts, plain arithmetic tells us the result must be bigger than 1. But, then, how much bigger? Even if we accept that there is mismeasurement here, the question remains as to why Tobin's Q should average 1. 24, rather than 1. 01 or 20 for instance. And here, too, just like in the case of Microsoft vs GM, the answer is elusive.
To pin down the difficulty, let's examine the structure of a balance sheet a bit more closely. Recall that corporations have two types of assets: tangible and intangible. 4 According to the neoclassical system of classification, tan- gible assets consist of capital goods - machines, structures and recently also software. Intangible assets, by contrast, represent knowledge, technology, organization, goodwill and other metaphysical entities. Mainstream econo- mists consider both types of assets productive, and the accountants concur - but with a reservation. Although tangible and intangible assets are both 'real', they cannot always be treated in the same way.
The reason is prosaic.
Tangible assets are bought and sold on the market and therefore have a universal price. Since the market is assumed to know all, this price is treated as an objective quantity and hence qualifies for inclusion in the balance sheet. By contrast, most intangible assets are produced by the firm itself. They are generated through internal R&D spending, in-house advertisement expenditures and sundry other costs associated with the likes of 'corporate re-engineering' and 'structural re-organisation'. These are not arm's-length transactions. They are not subject to the universalizing discipline
4 For mainstream analyses of intangibles, see for example Lev (2001) and Corrado, Hulten and Sichel (2006).
? 176 Capitalization
of the market, and therefore the intangible assets they generate lack an 'objective' price. And items that do not have an agreed-upon quantity, no matter how productive, cannot make it into the balance sheet. The best the accountants can do is to list them as current expenditures on the income statement.
There are two exceptions to the rule, though. One exception is when companies purchase pre-packaged intangibles directly through the market - for instance, by acquiring a franchise, patent, trademark, or copyright. The other is when one corporation acquires another at a price that exceeds the acquired company's book value. Since the merger does not create new tangible assets, the accountants assume that the premium must represent the intangible assets of the new formation. They also assume that since this premium is determined by the market, it must be objective. And given that the intangibles are objectively measured, the accountants feel safe enough to include them in the balance sheet.
So all in all we have three categories: (1) tangible assets that are included in the balance sheet, (2) intangible assets that are included in the balance sheet, and (3) intangible assets that are not included in the balance sheet. Now, as noted, fixed assets comprise only the first category, whereas capitalization reflects the sum of all three, and according to the conventional creed it is this mismatch that explains why the long-term average of Tobin's Q differs from 1.
The historical rationale goes as follows. Over the past several decades, US- based corporations have undergone an 'intangible revolution'. Their economy has become 'high-tech', with knowledge, information and commu- nication all multiplying manifold. As a consequence of this revolution, the growth of tangible assets decelerated, while that of intangible assets acceler- ated. And how do we know the extent of this divergence? Simple, say the neoclassicists. Subtract from the market value of firms the market price of their fixed assets, and then assume that, since the market knows all, the differ- ence equals the quantity of intangibles.
Using this standard method, a recent study of the S&P 500 companies esti- mates that, over the past thirty years, the ratio between their market value and the book value of their tangible assets has risen more than fourfold: from 1. 2 in 1975 to 5 in 2005 (Cardoza et al. 2006). The increase implies that in 1975 intangibles amounted to 17 per cent of the total assets, whereas in 2005 they accounted for as much as 80 per cent. Much of this increase is attributed to the growth of out-of-balance-sheet intangibles, whose share of market capitalization during the period is estimated to have risen from 15 to 65 per cent.
Conclusion: the 1. 24 mean value of Tobin's Q is hardly a mystery. It is simply another 'measure of our ignorance' - in this case, our inability to measure intangibles directly. Fortunately, the problem can be circumvented easily by indirect imputation. And, indeed, looking at Figure 10. 2, we can see that much of the increase in Tobin's Q occurred over the past couple of
Fiction, mirror or distortion? 177
decades - coinciding, as one would expect, with the upswing of the 'intangible revolution'.
This rationale may sound soothing to neoclassical ears, but accepting it must come with some unease. To begin with, the neoclassicists don't really 'measure' intangibles; rather, they deduce them, like the ether, as a residual. Moreover, according to their own imputations, this residual accounts for as much as 80 per cent of total market value. To accept this magnitude as a fact is to make the 'material' basis of the theory (shaky as it is) account for no more than 20 per cent of market capitalization - hardly an impressive achievement for a theory that calls itself 'mainstream'. Moreover, the impu- tation method itself doesn't seem very robust. Given that the quantity of intangibles is equal to the difference between market value and tangible assets, oscillations in market value imply corresponding variations in intan- gible assets. But, then, why would the quantity of a productive asset, no matter how intangible, fluctuate - and often wildly - even from one day to the next?
Boom and bust: adding irrationality
The solution to the latter riddle is to invoke irrationality. In this augmented neoclassical version, capitalized market value consists of not two compo- nents, but three: in addition to tangible and intangible assets, it also includes an amount reflecting the excessive optimism or pessimism of investors. And this last component, goes the argument, serves to explain the second anomaly of Tobin's Q - namely its large historical fluctuations.
This irrationality rationale is illustrated in Figure 10. 3. To explain it, let's backtrack and refresh the basics of rational economics. During good times, goes the argument, capitalist optimism causes investors to plough back more profits into productive assets. During bad times, the process goes in reverse, with less profit earmarked for that purpose. As a result, the growth of 'real' assets tends to accelerate in an upswing and decelerate in a downswing. This standard pattern is illustrated by the thick line in the figure. The data measure the rate of change of the current cost of corporate fixed assets (the denomi- nator of Tobin's Q), with the series smoothed as a 10-year moving average in order to accentuate its long-term pattern. According to the figure, the US corporate sector has gone through two very long 'real' accumulation cycles (measured in price terms), the first peaking in the early 1950s, the second in the early 1980s.
The vigilant reader will note that the accumulation process here reflects only the tangible assets - for the obvious reason that the intangible ones cannot be observed directly. But this deficiency shouldn't be much of a concern. Since neoclassical economists view intangible and tangible assets as serving the same productive purpose, they can assume (although not prove) that their respective growth patterns, particularly over long periods of time,
178 Capitalization 30
per cent
25 20 15 10
5
0 -5 -10 -15 -20
Current Cost of Corporate Fixed Assets (annual % change)
Hypothetical Market Value
of Corporate Equities & Bonds based on Neoclassical Orthodoxy Augmented by 'Market Abberations' * (annual % change)
1970 1980 1990 2000 2010 2020
? ? ? ? ? www. bnarchives. net
? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? 1930 1940 1950 1960
Figure 10. 3 The world according to the scriptures
* The value for each year is computed in two steps: first, by calculating the deviation of the growth rate of the (smoothed) 'real' series from its historical mean; and, second, by adding 2. 5 times the value of the deviation to the historical mean.
Note: Series are smoothed as 10-year moving averages.
Source: U. S. Bureau of Economic Analysis through Global Insight (series codes: FAPNREZ for current cost of corporate fixed assets).
are more or less similar. 5 So all in all, we could take the thick line as repre- senting the overall accumulation rate of 'real' capital, both tangible and intan- gible (denominated in current dollars terms to bypass the impossibility of 'material' quantities).
Now this is where irrationality comes in. In an ideal neoclassical world - perfectly competitive, fully transparent and completely informed - Fisher's 'capital wealth' and 'capital value' would be the same. Capitalization on the stock and bond markets would exactly equal the dollar value of tangible and intangible assets. The two sums would grow and contract together, moving
5 If, as neoclassicists seem to believe, the trend growth rate of intangibles is faster than that of tangibles, then the overall growth rate of 'real' assets (tangible and intangible) would gradu- ally rise above the growth rate of tangible assets only illustrated in Figure 10. 3. However, since the cyclical pattern would be more or less the same, this possibility has no bearing on our argument.
? Fiction, mirror or distortion? 179
up and down as perfect replicas. But even the neoclassicists realize that this is a mere ideal.
Ever since Newton, we know that pure ideas may be good for predicting the movement of heavenly bodies, but not the folly of men. Newton learned this lesson the hard way after losing plenty of money in the bursting of the 'South Sea Bubble'. Two centuries later he was joined by no other than Irving Fisher, who managed to sacrifice his own fortune - $10 million then, $100 million in today's prices - on the altar of the 1929 stock market crash.
So just to be on the safe side, neoclassicists now agree that, although capi- talization does reflect the objective processes of the 'real economy', the picture must be augmented by human beings. And the latter, sadly but truly, are not always rational. Greed and fear cloud their vision, emotions upset their calculations and passion biases their decisions - distortions that are further amplified by government intervention and regulation, lack of trans- parency, insider trading and other such unfortunate imperfections. All of these deviations from the pure model lead to irrationality and end in mis- priced assets.
But not all is lost. Convention has it that there is nonetheless order in the chaos, a certain rationality in the irrationality. The basic reason is that greed tends to operate mostly on the upswing, whereas fear usually sets in in the downswing. 'We tend to label such behavioural responses as non rational', explains Alan Greenspan (2008), 'But forecasters' concerns should be not whether human response is rational or irrational, only that it is observable and systematic'. The regularity puts limits on the irrationality; limits imply predict- ability; and predictability helps keep the faith intact and the laity in place.
The boundaries of irrationality are well known and can be recited even by novice traders. The description usually goes as follows. In the upswing, the growth of investment in productive assets fires up the greedy imagination of investors, causing them to price financial assets even higher. To illustrate, during the 1990s developments in 'high-tech' hardware and software suppos- edly made investors lose sight of the possible. The evidence: they capitalized information and telecommunication companies, such as Amazon, Ericsson and Nortel, far above the underlying increase in their 'real' value. A similar scenario unfolded in the 2000s. Investors pushed real-estate capitalization, along with its various financial derivatives and structured investment vehi- cles, to levels that far exceeded the underlying 'actual' wealth. The process, which neoclassicists like to think of as a 'market aberration', led to undue 'asset-price inflation'. Naturally, the capitalization created by such 'bouts of insanity' is mostly 'fake wealth'. It represents 'fictitious value' and leads to inevitable 'bubbles'. 6 But there is nonetheless a clear positive relationship here: the irrational growth of 'fake wealth', although excessive, moves in the same direction as the rational growth of 'real wealth'.
? 6 For a typical analysis of 'bubbles', complete with the above jargon, see Janszen (2008).
180 Capitalization
The process is said to invert during a bust. This is where fear kicks in. The 'real' economy decelerates, but investors, feeling as if the sky is falling, bid down asset prices far more than implied by the 'underlying' productive capacity. A famous illustration is offered by the Great Depression. During the four years from 1928 to 1932, the dollar value of corporate fixed assets contracted by 20 per cent, while the market value of equities collapsed by an amplified 70 per cent (we have no aggregate figures for bonds). A similar 'undershooting' occurred during the 1997 Asian financial crisis, with market value contracting by 50 per cent in many cases, against a growth slowdown or a very moderate decline in the dollar value of the 'real' capital stock. Yet here, too, the relationship is clear: the irrational collapse of 'fictitious value', however exaggerated, moves together with the rational deceleration of 'productive wealth'.
This bounded irrationality is illustrated by the thin line in Figure 10. 3. Note that this series is a hypothetical construct. It describes what the growth of capitalization might look like when neoclassical orthodoxy is augmented by 'irrationality' and 'market aberrations'. The value for each year in the hypothetical series is computed in two steps. First, we calculate the deviation of the growth rate of the (smoothed) 'real' series from its historical mean (so if the smoothed growth rate during the year is 8 per cent and the historical mean rate is 6. 7 per cent, the deviation is 1. 3 per cent). Second, we add 2. 5 times the value of the deviation to the historical mean (so in our example, the hypothetical smoothed growth rate would be 2. 5 * 1. 3 + 6. 7 = 9. 95 per cent). The coefficient of 2. 5 is purely arbitrary. A larger or smaller coefficient would generate a larger or smaller amplification, but the cyclical pattern would remain the same.
This simulation solves the riddle of the fluctuating Tobin's Q. It shows how, due to market imperfections and investors' irrationality, the growth of capitalization overshoots 'real' accumulation on the upswing, therefore causing Tobin's Q to rise, and undershoots it on the downswing, causing Tobin's Q to decline.
And so everything falls into place. Tobin's Q averages more than 1 due to an invisible, yet very real intangible revolution. And it fluctuates heavily - admittedly because the market is imperfect and humans are not always rational - but these oscillations are safely bounded and pretty predicable. Capitalization indeed deviates from the 'real' assets, though in the end it always reverts back to the 'fundamentals'.
Or does it?
The gods must be crazy
It turns out that while the neoclassical priests were busy fortifying the faith, the gods were having fun with the facts. The result is illustrated in Figure 10. 4 (where both series again are smoothed as 10-year moving averages). The thick line, as in Figure 10. 3, shows the rate of change of corporate fixed assets
Fiction, mirror or distortion? 181
measured in current replacement cost. But the thin line is different. Whereas in Figure 10. 3 this line shows the rate of growth of capitalization stipulated by the theory, here it shows the actual rate of growth as it unfolded on the stock and bond markets. And the difference couldn't have been starker.
The gyrations of capitalization, instead of amplifying those of 'real' assets, move in exactly the opposite direction. It is important to note that we are dealing here not with short-term fluctuations of the business cycle, but with very long-term waves of roughly 30-year duration. Furthermore, the pattern seems anything but accidental. In fact, it is rather systematic: whenever the growth rate of 'real' assets decelerates, the growth rate of capitalization accel- erates, and vice versa. 7
This reality puts the world on its head. One could perhaps concede that 'real' assets do not have a material quantum - yet pretend, as we have agreed
25
20
15
10
5
0
-5
-10
1930 1940
1950 1960
1970 1980
1990 2000 2010 2020
? ? ? ? per cent
? ? Market Value of Corporate Equities & Bonds (annual % change)
? ? ?
