Putting Power Back Into Growth Theory

Fix, Blair Working Paper Putting Power Back Into Growth Theory Working Papers on Capital as Power, No. 2014/05 Provided in Cooperation with: The Bichler & Nitzan Archives Suggested Citation: Fix, Blair (2014) : Putting Power Back Into Growth Theory, Working Papers on Capital as Power, No. 2014/05, Forum on Capital As Power - Toward a New Cosmology of Capitalism, s.l., http://bnarchives.yorku.ca/438/ This Version is available at: https://hdl.handle.net/10419/157868 Standard-Nutzungsbedingungen: Die Dokumente auf EconStor dürfen zu eigenen wissenschaftlichen Zwecken und zum Privatgebrauch gespeichert und kopiert werden. Sie dürfen die Dokumente nicht für öffentliche oder kommerzielle Zwecke vervielfältigen, öffentlich ausstellen, öffentlich zugänglich machen, vertreiben oder anderweitig nutzen. Sofern die Verfasser die Dokumente unter Open-Content-Lizenzen (insbesondere CC-Lizenzen) zur Verfügung gestellt haben sollten, gelten abweichend von diesen Nutzungsbedingungen die in der dort genannten Lizenz gewährten Nutzungsrechte. Terms of use: Documents in EconStor may be saved and copied for your personal and scholarly purposes. You are not to copy documents for public or commercial purposes, to exhibit the documents publicly, to make them publicly available on the internet, or to distribute or otherwise use the documents in public. If the documents have been made available under an Open Content Licence (especially Creative Commons Licences), you may exercise further usage rights as specified in the indicated licence. https://creativecommons.org/licenses/by-nc-nd/4.0/ ! ! No. 2014/05 Putting Power Back Into Growth Theory Blair Fix December 2014 * Winner of the 2014 RECASP Essay Prize * http://www.capitalaspower.com/?p=1692 WORKING PAPERS ON CAPITAL AS POWER Putting Power Back Into Growth Theory DRAFT Blair Fix* York University [email protected] December 26, 2014 Abstract Neoclassical growth theory assumes that economic growth is an atomistic process in which changes in distribution play no role. Unfortunately, when this assumption is tested against real-world evidence, it is systematically violated. This paper argues that a reality-based growth theory must reject neoclassical principles in favour of a power-centered approach. Building on Nitzan and Bichler’s Capital as Power hypothesis, I argue that hierarchy formation is an integral part of the growth process. I hypothesize that the role of capital accumulation (through profit) is to facilitate hierarchy formation by legitimizing the authority of capitalists. * Blair Fix is a PhD student in the Environmental Studies program at York University. He is the author of the book Rethinking Economic Growth Theory from a Biophysical Perspective. His research challenges neoclassical theories of growth and aims to construct a new theory that combines biophysical analysis with radical political economy. 1 Introduction Neoclassical growth theory is the dominant approach to understanding economic growth. The canonical neoclassical model – the Solow-Swan model – treats growth as an atomistic process that takes place under conditions of perfect competition. Thus, it assumes that concentrated power plays no role in the growth process. Moreover, neoclassical growth theory assumes that changes in economic distribution have no effect on growth. The aim of this paper is twofold: first, to empirically demonstrate that these assumptions are false; and secondly, to begin formulating a growth theory that does explain reality. This paper provides evidence that a three-way link exists between distribution, corporate employment concentration, and the growth of energy consumption. Building on the Capital as Power framework proposed by Nitzan and Bichler (2009), I argue that hierarchy plays a central role in growth, and that the role of capital accumulation (through profit) is to facilitate hierarchy formation by legitimizing capitalist authority. 1 1 INTRODUCTION 2 1.1 Neoclassical Growth Theory Neoclassical growth theory is a logical extension of the neoclassical theory of the firm. The latter treats the firm as a ‘black box’ – all that is known are inputs (labor and capital) and outputs (goods and services). Neoclassical microeconomics posits the existence of a ‘production function’ – essentially a formula – that can explain how the quantities of a firm’s inputs are related to the quantities of its outputs. Neoclassical macroeconomics applies the same line of thinking to the entire economy: it is the economy that becomes a black box, described only by inputs and outputs. It is posited that a unique aggregate production function exists that can quantitatively explain how capital and labor inputs relate to total economic output. Although there are many varieties of neoclassical growth theory, the Solow-Swan model (Solow 1956; Swan 1956) has become the canonical approach (Acemoglu 2008). At its core is a Cobb-Douglas production function (Eq. 1) in which capital (K), labor (L) and ‘technical progress’ (A) mix together to create material output (Y).1 Y=ALβKα(1) While there are numerous problems with production functions,2for the present discussion, I am concerned with the following two implicit assumptions contained within the Solow-Swan model: (1) distribution is unrelated to growth; and (2) large institutions (i.e. concentrated power) are unimportant to growth I begin with distribution. A central tenet of neoclassical distribution theory is that one’s income is proportional to one’s marginal productivity. When applied to neoclassical growth theory, marginal productivity theory predicts that the exponents αand β (in Eq. 1) should be equal to capital and labor’s share of income, respectively. In neoclassical growth theory, it has become standard practice to assume that these exponents are constant. This tradition has its roots in the work of Cobb and Douglas (1928), who showed that fixed exponents could be used to model historical production. The constancy of ‘factor shares’ was later formalized by Nicholas Kaldor (1957) who put forward a list of six stylized facts about economic growth, one of which was the historical tendency for capital and labor income shares to remain approximately constant over time (about 1/3 and 2/3 respectively). By assuming that returns to labor and capital are constant over time, neoclassical growth theory assumes that changes in distribution do not affect growth. In order for the neoclassical aggregate production function to be valid, two more assumptions are required: (1) all firms (and the economy as a whole) must experience constant returns to scale; and (2) the economy must be perfectly competitive. Constant returns to scale is a property of a production function Y(K, L)such that increases in the scale of capital and labor inputs yield a corresponding increase in total output. Stated mathematically, this becomes: 1In many ways the “technical progress” term is a fudge-factor that adjusts for the empirical inaccuracy of the Solow-Swan model. Without this term, the Solow-Swan model fails to explain a large portion of historical growth (Ayres and Warr 2009). 2For a small sample of the literature critiquing production functions, see Felipe and Fisher 2003; Felipe and Holz 2001; Fisher 1969; Robinson 1953; Shaikh 1974; Shaikh 2005 1 INTRODUCTION 3 Y(cK, cL) = cY (K, L)(2) While, in principle, the production function of individual firms can have either constant, increasing, or decreasing returns to scale, in order to maintain compatibility with the marginal productivity theory of distribution, one must assume constant returns to scale. This assumption arose from an ‘adding up’ problem that neoclassical economists first faced when formulating the marginal productivity theory of distribution. Joan Robinson summarizes this issue: “How do we know that, if each factor is paid its marginal product, the total product is disposed of without residue, positive or negative?” (1934, p. 398). Using a theorem developed by the mathematician Leonhard Euler, theologian Philip Wicksteed (1894) formulated an elegant solution to the ‘adding up’ problem. He showed that if one assumes that production exhibits constant returns to scale, then Euler’s theorem could be used to ‘prove’ that each factor receives payment in exact accordance with its marginal productivity, thus solving the ‘adding up’ problem.3As a result of this theorem, neoclassical growth theory (which maintains compatibility with neoclassical distribution theory) is forced to assume that all firms have constant returns to scale. This means that size is assumed to be neither an advantage nor disadvantage: all firms, large or small, are given the same production function. We now turn our attention to the assumption of perfect competition. Neoclassical theory predicts that under conditions of perfect competition, markets will allocate resources in the most efficient manner possible (given existing patterns of distribution).4Here, perfect competition is used specifically to mean that firms are price-takers: they have no market power that allows them to dictate the price of their output. However, as Coase (1937) noted, this results in a paradox. While the firm is the basic unit of production in neoclassical theory, the theorized efficiency of perfect competition implies that firms should not exist. According to neoclassical logic, the most efficient form of production should occur when competition is most atomistic – when every individual is self-employed. Given this paradox, why does neoclassical theory continue to assume perfect competition? Steve Keen (2001) argues that it is because without it, the most basic tenet of neoclassical theory – the equilibrium-seeking price mechanism – cannot be justified. Before explaining the problem, let us first review the neoclassical explanation of the market. Arthur Salter says it best: ... the normal economic system works itself. For its current operation it is under no central control, it needs no central survey. Over the whole range of human activity and human need, supply is adjusted to demand, and production to consumption, by a process that is automatic, elastic, and responsive. (1921, p. 15) In neoclassical theory, the equilibrium-seeking quality of the free market is explained by the forces of supply and demand. Market equilibrium occurs at the inter- 3As used by Wicksteed, Euler’s Theorem states that if Y=f(a, b, c, ...)is a production function with factors of production a, b, c, ... that exhibits constant returns to scale, then Y=a∂Y ∂a +b∂Y ∂b +c∂Y ∂c +.... That is, ouput Yis guaranteed to be the sum of the quantity of each factor times its marginal productivity. 4This is known as the first fundamental theorem of welfare economics: under conditions of perfect competition, market equilibrium is Pareto-efficient. It is impossible to make any one individual better off without making at least one individual worse off. 2 MEASURING ECONOMIC GROWTH 4 section of supply and demand curves. These theoretical curves, in turn, are explained in terms of the behavior of individual consumers and producers. The downward sloping demand curve is explained by the law of diminishing marginal utility: a consumer will derive a decreasing amount of pleasure from each additional unit of consumption. The upward sloping supply curve is explained by the law of increasing marginal costs: each additional unit of production is assumed to become costlier to produce.5As long as all firms are price-takers (meaning there is perfect competition), the resulting equilibrium quantity of production is such that the market price equals both the marginal utility of the buyer and the marginal cost of the supplier. This elegant explanation of the price mechanism underlies all other aspects of neoclassical theory. Yet without perfect competition, it fails to function. If firms have even the slightest market power, then the equilibrium price will diverge from marginal costs, causing the theory to break down. Steve Keen summarizes: “Unless perfect competition rules, there is no supply curve” (2001, p. 101). Thus, the assumption of perfect competition is central to the internal consistency of neoclassical theory. When applied to neoclassical growth theory, this leads to the conclusion that the optimal growth path should be through atomistic competition: large firms should play no preferential role (indeed, they should not exist). This amounts to an implicit assumption that concentrated power plays no role in growth. To summarize, neoclassical growth theory assumes that distribution and institutional size play no role in economic growth. In the following sections, I demonstrate that these assumptions are directly contradicted by empirical evidence. First, however, I review a major problem (that goes largely unnoticed) facing all economic growth theorists: the accepted measure of economic scale – ‘real’ GDP – is irreconcilably flawed. 2 Measuring Economic Growth There is nothing more important to a theory of economic growth than the ability to objectively measure the scale of the economy. Without such a measure, a scientific theory of growth is impossible. Unfortunately, the accepted measure of economic growth – ‘real’ GDP – is plagued by fundamental epistemological difficulties that render it fundamentally subjective. Any act of measurement begins with an act of reduction. The observer must find a suitable unit for reducing the qualities of the universe to a single quantity. The choice of unit crucially affects this mapping of quantity onto quality. Thus, the concept of economic growth is only meaningful if we can first agree on what it is that is growing. To state formally, measuring material production (Y) can only be done if we reduce it to a single quantity, Q: Y=Q(3) In principle, Qcan be defined in terms of any unit. However, before selecting a particular unit, we must present arguments about why this unit is meaningful. Further- 5One might protest that the reverse may actually be true – that each additional unit of production will cost less. While this may be true in reality, as Harold Lydall notes, “neoclassical theory is built on the ... assumption of absence of economies of scale” (1971 p. 91). 2 MEASURING ECONOMIC GROWTH 5 more, for a unit to be effective, it must be socially agreed upon, and it must not change over time and space. Strangely, economists have chosen a unit – price – that does not uphold this simple principle. 2.1 The Changing Meter Stick Let us begin by looking not at the real world of heterogeneous production, but at an imagined world in which production is homogenous. In this world, only apples are produced, and they are all uniform. In this world, it makes sense to use “apples” as our unit of measurement. If, in the year in question, 300 apples were produced, then: Y= 300 apples (4) Now imagine that our economy begins to produce both 300 apples and 100 oranges (again, all uniform). Now production becomes: Y= 300 apples + 100 oranges (5) This presents a problem: we wish to express Yin terms of a single quantity – but to paraphrase an old adage, you can’t add apples and oranges. We must find a third unit that allows the comparison of “apples” and “oranges”. Again, the unit must make sense. For instance, if we were shipping apples and oranges in a truck, a common unit of mass (kg) would make sense. Alternatively, if we simply wanted to eat them, a unit of energy (calories) would be more appropriate. Since the study of prices is their domain, economists naturally choose monetary value as a common unit of aggregation. This seems reasonable: the price of an orange is much more important to the average person than almost any other metric (mass, energy, etc.). Keeping with this tradition, we now measure output Yin units of dollars. In order to do so, we must know both the quantity of apples and oranges (QAand QO, respectively) and their unit prices (PAand PO). Production now becomes: Y=QAPA+QOPO(6) Using the quantities from above (300 apples and 100 oranges) and adding prices of $3 and $1 for apples and oranges respectively, we get the quantity of production: Y= (300 apples)(3 $/apple) + (100 oranges)(1 $/orange) = $900 + $100 = $1000 (7) Despite the definiteness of our answer, the matter is soon complicated when we realize that our chosen unit (the price of a commodity) changes all the time! For instance, the following year, we might produce the same quantity of apples and oranges, but the price of apples falls drastically to the same price as oranges ( $1). Then, without any physical changes, our measure of output is drastically reduced: Y= (300 apples)(1 $/apple) + (100 oranges)(1 $/orange) = $300 + $100 = $400 (8) 2 MEASURING ECONOMIC GROWTH 6 Which one of these measures of material production is “correct”? Here lies the fundamental problem: both of them are! By choosing price as an appropriate unit for measurement, we immediately remove the possibility of attaining a single measure for the quantity of output because our unit is not socially agreed upon over time. No amount of intellectual gymnastics can get us out of this dilemma. Without an objective way to decide the year in which prices were ‘correct’6, our measure of economic scale is simply not well defined. For those who remain unconvinced by this conceptual argument using imagined numbers, we can apply the same reasoning to an empirical example (Fig. 1). Here we use historical quantity and price data for the production of cars and computers (mostly from the United States). Unlike above, now both prices and the physical configuration of production change. Again, we must choose a ‘base’ year in which prices were ‘correct’, and then fix this price across time. This creates a ‘real’ GDP time-series for our 2 product economy. Different choices of base year drastically change the way we conceive of output growth.7Indeed, the economy simultaneously grows considerably and hardly at all! A further difficulty with real GDP methodology is that commodities change qualitatively over time. Neither the computers nor the cars of 1980 looked anything like those of 2010. In order to combat this problem, statistical agencies attempt to measure these qualitative changes. However, we again encounter a number of fundamental problems. Firstly, we must subdivide a given commodity into relevant attributes. But how do we objectively decide those attributes that are relevant and those that are irrelevant? Furthermore, once we have reduced a commodity to its constituent attributes, how do we decide their relative importance? The most popular method is called hedonic quality adjustment. The Bureau of Labor Statistics summarizes the process as follows: In price index methodology, hedonic quality adjustment has come to mean the practice of decomposing an item into its constituent characteristics, obtaining estimates of the value of the utility derived from each characteristic, and using those value estimates to adjust prices when the quality of a good changes. (Bureau of Labor Statistics 2010)(emphasis added) All quantitative comparisons require a unit of measurement. Here we see that the Bureau of Labor Statistics is attempting to measure the attributes of a commodity in units of utility. This is problematic because utility (a hypothetical psychic flux) cannot be directly measured; rather, it must always be ‘revealed’ through prices (Samuelson 1938). Thus, hedonic measurement becomes circular: distinguishing changes in price from changes in quality requires knowledge of consumer preferences; however, con- 6There is no such objective way to decide the ‘correctness’ of prices (Cochrane 2011). Appeals to the contrary always imply an additional unit used to explain prices. For Marxists, this is a commodity’s sociallynecessary, abstract labor content. For neoclassicists, it reduces to the marginal utility derived from a commodity. In both cases, the argument for a “correct” price rests upon its correlation with a hidden quantity which (conveniently) cannot ever be measured. A more logically sound way to think about prices is that they are always ‘correct’, by definition. 7The US Bureau of Economic Analysis (BEA) is aware of this problem. Its response has been to concede that the choice of base year is subjective. However, rather than conclude that this invalidates real GDP (as I have), the BEA has adopted a new method, called ‘chain-weighting’, that uses a moving average for all base years (Steindel 1995). While this might seem reasonable, it is similar to measuring your height in both meters and feet and then averaging the data to arrive at your ‘true’ height. The result is meaningless. 2 MEASURING ECONOMIC GROWTH 7 50 100 150 200 250 300 1970 1980 1990 2000 2010 2020 "Real" GDP Base Year 2008 "Real" GDP Base Year 1980 1980 = 100 2011 = 129 2011 = 231 Conflicting Measures Historical Data Number of Units 0 100 200 300 400 500 1970 1980 1990 2000 2010 2020 $ 0 5000 10000 15000 20000 $ 0 500 1000 1500 2000 1970 1980 1990 2000 2010 2020 Number of New Computers (1980 = 1) Number of New Cars (1980 = 100) Average Price of New Computer (Right) Average Price of New Car (Left) Unit Quantities Unit Prices Indexed to Historical Production Blair Fix Figure 1: Measuring Production with a Changing Meter Stick Note: This methodology is modeled after critiques of capital aggregation offered by Nitzan & Bichler (2009). Sources: Quantity of cars from Ward’s Automotive Group, “U.S. Car and Truck Sales, 1931-2012” at wardsauto.com/public-data. Quantity of computers from Jeremy Reimer, Total Share: Personal Computer Market Share 1975-2010, jeremyreimer.com. Price of cars from Bureau of Economic Analysis, Table 10.11, Average Price of a New Car, 1970-2011 (using domestic prices). Computer prices are from the Wikipedia entry for personal computer. Note: computer price indices from US Bureau of Labor statistics are unsuitable here because they adjust for changing computer quality (ie: processor speed, memory, etc.). 4 THE INSTITUTIONAL CONTEXT 14 Proprietor Income Costs Income Stream Other Firms Income Stream Other Firms Profit Costs Salary Corporation Self-Employed, Sole Proprietor Self-Employed, Incorporated 50% 50% 50% 30% 20% Decision-Making Power Decision-Making Power Blair Fix Figure 5: Atomistic Institutions 4.1 Atomistic Institutions Atomistic institutions consist of a single, self-employed person (Fig. 5). There are two possible configurations: the sole-proprietorship or the self-employed individual who incorporates his/her business. Modern accounting principles dictate that a soleproprietor’s income be called ‘proprietor income’, and not profit. However, the distinction is in name only – both profit and proprietor income are defined as the total sales less the costs of doing business. If a self-employed individual incorporates, this allows for a conceptual (and legal) separation of income into ‘profit’ and ‘salary’. There are two main benefits to incorporating. Firstly, corporations are limited liability institutions, which allows a legal separation of business and personal assets. In the event of a bankruptcy, only business assets can be seized. The second benefit is that profit is generally taxed at a different rate than a salary. For instance, in 2011, the effective US corporate tax rate was 21%, while the highest tax rate for personal income 4 THE INSTITUTIONAL CONTEXT 15 was 35%.14 Despite these differences, the two forms of business displayed in Figure 5 are, for all intents and purposes, the same. Let us envision a society populated only by these two institutional configurations (similar to the one imagined by Adam Smith (1776). We ask the following question: what is the effect of redistributing income from wages and proprietor income towards profit? There are two possible ways for this to occur. The first is if a sole-proprietor decides to incorporate his/her business. This would eliminate his/her proprietor income from the national accounts, but add wage and profit income in the same amount. The effect would be a change on paper, but no meaningful change in who actually controls this income (the same person in both cases). Alternately, a self-employed person with an incorporated business might decide to allocate more income to profit rather than to salary (if tax rates changed, for instance). Again this has no meaningful effect outside of a re-categorization on paper: in both cases the individual’s total income remains unchanged. For a society populated entirely by atomistic institutions, it is difficult to see how an income redistribution towards profit would change anything but the abstract accounting category used to classify income. 4.2 Flat Institutions We now move on to institutions that include more than one person. We begin with nonhierarchical, or so-called ‘flat’, institutions (Fig. 6). A flat institution is characterized by a complete lack of hierarchy. In its ideal form, this means that each individual has an equal say in all decision-making processes. Our hypothetical, flat institution can either be operated as a non-profit organization (i.e. a cooperative), or as a flat corporation (with ownership divided equally among its members). In the former case, all income in excess of costs is allocated to salaries, while in the latter case, this income is split between profit and salaries. As we did previously, we imagine a society populated only by such flat institutions. Again, we ask: what is the effect of a redistribution of income towards profit? This could occur two ways – either by non-profits deciding to become for-profit, or by forprofits increasing their markup (profit as a portion of total income). In neither case does this change affect the ultimate control of the pre-existing income stream (which is always allocated equally to all individuals). However, the re-categorization of salaries into profit does have the effect of pooling income. For instance, having a group of 5 people control $100 000 in profit is different than having each of those 5 people control $20 000 in salaries. Pooling income allows for the possibility of a larger ‘investment’ than would be possible otherwise. However, the ability to pool income does not require profit. Indeed, the initial income stream is the ultimate source of any pooled income. Thus, if a co-operative wishes to make a large purchase, it may simply divert more of its income stream towards ‘costs’ and less towards salaries. The end result is the same. 14Corporate tax rate is calculated by dividing total before tax profit by total tax collected, using BEA Table 1.12. Income tax rate is from IRS Table 23, U.S. Individual Income Tax: Personal Exemptions. 4 THE INSTITUTIONAL CONTEXT 16 Income Stream Salaries Costs (Non-Labour) Other Firms Non-Profit Institution Income Stream Salaries Costs (Non-Labour) Profit Other Firms Corporation "Flat" Non-Profit Institution "Flat" Corporation 20% 30% 50% 6% 50% 50% 10% Decision-Making Power Decision-Making Power Blair Fix Figure 6: Flat Institutions 4 THE INSTITUTIONAL CONTEXT 17 Leader Salaries Costs (Non-Labour) Income Stream 50% 50% 10% 4% 2% Other Firms Non-Profit Institution Hierarchical Non-Profit Institution Decision-Making Power Salaries Costs (Non-Labour) Profit Income Stream 20% 30% 50% 3% 1.5% Other Firms Corporation Owner Hierarchical Private Corporation Decision-Making Power Blair Fix Figure 7: Hierarchical Institutions 5 CONNECTING HIERARCHY, DISTRIBUTION, AND GROWTH 18 As we did with single-person institutions, we reach the conclusion that when profit only flows to flat institutions, an income redistribution towards profit should have no effect outside of a change on paper. 4.3 Hierarchical Institutions We now move on to hierarchical institutions (Fig. 7). Here we envision the quintessential hierarchy that is marked by a strict top to bottom chain of command, with all decision-making power ultimately residing at the top. We have two possible types of institution – the hierarchical non-profit and the hierarchical corporation. A good example of hierarchical non-profits are state-owned companies like Fannie Mae or PetroChina, while Walmart and General Motors are examples of hierarchical corporations. As before, we are interested in the effect of redistributing income towards profit, but now in a society populated entirely by hierarchical institutions. There are two possible scenarios: either a non-profit organization may become a for-profit (as when a stateowned company is privatized) or a for-profit organization could increase its markup. Unlike our previous examples, here both scenarios imply a significant change in who controls what. A differential increase in profit will serve to concentrate income at the top of the chain of command. The results of our conceptual investigation demonstrate that it is only when coupled with hierarchy that a redistribution towards profit has any meaningful effect on who controls what. In all other institutional settings, introducing/increasing profit has no effect beyond a shift in abstract accounting categories. 5 Connecting Hierarchy, Distribution, and Growth Given the empirical link between redistribution and the growth of energy consumption, and our finding that relative changes in profit are only meaningful if they occur within a hierarchical institution, it seems logical to look for connections between hierarchy, profit, and growth. To do so, we must decide on a metric for hierarchy. Inspired by Nitzan and Bichler’s (2009) concept of ‘breadth’, I propose using the employment share of the largest ncorporations as such a measure (where nis an arbitrary number chosen based on data availability). The logic underpinning this metric is straightforward: large corporations are hierarchical institutions; therefore, the extent to which such corporations dominate total employment should give us an indication of the ‘degree of hierarchy’ of society. 5.1 Hierarchy and Growth In order to connect hierarchy and growth, I continue to use energy per capita as my metric for growth. However, due to the lack of data at the global level, I use primary energy consumption, rather than useful work. My methodology is straightforward: I simply compare corporate employment concentration to energy use per capita and look for correlation. The results of this analysis, undertaken first at the international level (Fig. 8) and then at the national level (Fig. 9), demonstrate a clear connection (across both 5 CONNECTING HIERARCHY, DISTRIBUTION, AND GROWTH 19 10-3 10-2 10-1 100 101 102 Energy Use per Capita (kg oil equivalent) 100 1000 10000 Blair Fix Kuwait Croatia Bangladesh Sri Lanka Netherlands Brazil Russia Mexico Thailand Belgium Portugal Philippines R = 0.62 2 Hong Kong Vietnam Power Regression Top 10 Corporations Percentage of Domestic Labor Force Canada 2008 No Causation Implied Figure 8: Global Corporate Employment Concentration vs. Energy Use Sources: National energy use per capita and total labor force data is from the World Bank (indicator codes EG.USE.PCAP.KG.OE. and SL.TLF.TOTL.IN, respectively). Employment of top 10 corporations (ranked by number of employees) is from COMPUSTAT Global Fundamentals (series EMP). space and time) between corporate employment concentration and energy consumption per capita. From a neoclassical perspective, this finding is puzzling. Indeed, neoclassical growth theory assumes that concentrated power should play no role in the growth process. The evidence, however, suggests just the opposite: growth is consistently associated with a decline in competition and an increase in the control of large corporations. That is to say, growth and the concentration of power appear to be intrinsically related. 5.2 Hierarchy and Profit So far, I have empirically connected relative changes in profit to changes in energy consumption, and I have empirically connected changes in hierarchy to changes in energy consumption. The last piece of the puzzle needed to create a three-way connection between hierarchy, profit and growth, is to link relative changes in profit to changes in hierarchy. In order to do this, I turn to capitalist income, which consists of the sum of profit and interest. An important question to ask is – does the composition of capitalist 5 CONNECTING HIERARCHY, DISTRIBUTION, AND GROWTH 20 12 14 16 18 20 22 24 350 400 450 500 550 1940 1950 1960 1970 1980 1990 2000 2010 2020 R2 = 0.85 United States National Energy Consumption per Labor Hour (Right) Top 200 Corporations Percent of Total Employment (Left) MJ/Hr Blair Fix % of Employment Figure 9: US Corporate Employment Concentration and Energy Use per Capita Sources: Total US employment from BEA Tables 6.5 B-D (Full-Time Equivalent Employees by Industry). Employment of top 200 corporations (ranked by number of employees) from COMPUSTAT (series DATA29). Total energy consumption from EIA Table 1.3 (Primary Energy Consumption by Source). Total labor hours from BEA Tables 6.9 B-D (Hours Worked by Full- Time and Part-Time Employees by Industry). income (the balance between interest and profit) affect hierarchy formation? Political economists have long sought to understand the differences between interest and profit. In Marxist political economy, interest is generally regarded as parasitic,15and profit (while exploitative) is regarded as productive. However, Nitzan and Bichler (2009) challenge this long-held belief. They note that, for the absentee owner, there is very little practical difference between owning debt (and earning interest) versus owning equity (and earning profit). In either case the goal is the same: to transform capital into an income stream. Nitzan and Bichler argue that interest represents the ‘normal’ rate of return, while profit offers the chance to beat this normal rate. While both debt and equity holders should both be considered ‘owners’ of a corporation, there is an important legal difference between the two forms of ownership. Other than when a corporation is in receivership, debt holders have no legal control over business decisions – it is equity holders that have this right. If a particular equity holder owns enough stock (and often owning only a small fraction of outstanding stock is enough) he or she has complete control of business decision-making. If we think of 15Marx referred to interest-bearing capital as ‘usurer’s capital’ (1894, Ch. 36). 5 CONNECTING HIERARCHY, DISTRIBUTION, AND GROWTH 21 -2 -1 0 1 2 3 4 0.5 0.6 0.7 0.8 0.9 1950 1960 1970 1980 1990 2000 2010 2020 % Blair Fix Top 200 Corporations Employment Concentration Growth Rate (10 year moving average, Left) Profit as a Fraction of Capitalist Income (K) (K = Profit + Interest) (10 year moving average, Right) United States Figure 10: Connecting Profit to Corporate Concentration Sources: Total US employment from BEA Tables 6.5 B-D (Full-Time Equivalent Employees by Industry). Employment of top 200 corporations (ranked by number of employees) from COMPUSTAT (series DATA29). Profit and Interest from Bureau of Economic Analysis, Table 1.12, data series: Corporate profits with IVA and CCAdj, Net interest and miscellaneous payments. this in terms of hierarchy, then it is clear that equity holders (with controlling shares) are at the top of the corporate pyramid. Debt holders, on the other hand, are only at the top if a corporation files for bankruptcy (i.e. during periods of crisis). How, then, does the composition of capitalist income relate to hierarchy formation? Figure 10 gives insight into this question. Here we plot the annual rate of change of US corporate employment concentration (smoothed with a 10 year moving average) against the share of profit in capitalist income (also smoothed with a 10 year moving average). The correlation is clear: corporate concentration (i.e. hierarchy) grows more rapidly when capitalist income is dominated by profit, and more slowly when capitalist income is dominated by interest (R2is 0.70 for smoothed data, 0.17 for raw data). Profit, it would seem, is key for hierarchy formation. 6 PUTTING POWER BACK INTO GROWTH THEORY 22 6 Putting Power Back into Growth Theory Having established a three-way link between profit, hierarchy and growth, I now offer my own hypotheses about why this connection exists. While speculative at this point, these hypotheses offer plausible grounds for future inquiry. 6.1 Hierarchy and Group Size The evidence suggests that the growth of energy consumption requires the formation of large, hierarchical organizations that are capable of mobilizing vast groups of people towards a single objective. But why is this the case? Why can’t growth be accomplished by the random interactions of atomistic institutions (as neoclassical theory suggests)? One possible explanation (which I pursue here) is that humans have evolved to function in small, egalitarian groups. Without coercive, centralized power, such small-scale groups will not be able (or willing) to coordinate their actions. In order to coordinate larger groups of people, egalitarian relationships must be abandoned in favor of hierarchical ones. Recent anthropological research supports the hypothesis that egalitarian group size is fundamentally constrained by human brain size. The ability to form social groups is, in large part, a function of genetic inheritance. This becomes obvious when we compare different species: many animals (such as bears) are incapable of forming large groups, while others (such as wolves) do so naturally. All social organisms have evolved mechanisms that maintain the cohesiveness of their groups. In primates, it appears that this has involved the development of large brains. In a remarkable study, anthropologist Robin Dunbar (1992) found that the group size of different primates was highly correlated with the relative size of their neocortex (Fig. 11). His conclusion was that neocortex size places an upper limit on the number of social relations that can be monitored by an individual. That is, since managing social relationships requires computational ability, brain size imposes a limit on group size. From his results on non-human primates, Dunbar (1993) extrapolates to find that human brain size predicts a group-size limit of about 150 (often called ‘Dunbar’s number’). While this number should be considered exploratory, Dunbar notes that Neolithic villages had populations in this order of magnitude. Clearly, however, humans have developed ways of vastly exceeding this social scale: modern cities can surpass Dunbar’s number by five orders of magnitude. If we accept Dunbar’s hypothesis, it follows that any mechanism that allows vast increases in group size must function to limit internal interactions between group members. Neoclassical theory attempts to prove that ‘the market’ is the ultimate organizational mechanism. Unfortunately, ‘the market’ does not act to limit the number of social interactions; instead, it replaces qualitative relationships with quantitative ones (by introducing prices). Therefore, ‘the market’ may act to simplify or ‘standardize’ social relationships, but it does not act to limit their number: any member of a group can still engage in a market exchange with any other member of the group. Unlike the market, Turchin and Gavrilets (2009) note that hierarchical organization allows group size to grow without a corresponding increase in the number of interpersonal relationships. A member of a hierarchy needs to have a relationship only with 6 PUTTING POWER BACK INTO GROWTH THEORY 23 Figure 11: Mean group size vs. relative neocortex size for various primate species Note: Plot icons (squares, cirlcles, etc.) represent different primate genera. Rest of the brain = total brain volume less neocortex). Source: Dunbar (1992). Figure 12: Hierarchical complexity vs. population of six historical empires Note: Hierarchical complexity is counted in terms of the number of distinct administrative levels. Source: Turchin and Gavrilets (2009). REFERENCES 30 Shaikh, Anwar. 2005. “Nonlinear dynamics and pseudo-production functions”. In: Eastern Economic Journal 31.3, pp. 447–466. Smil, V. 1994. Energy in World History. Boulder: Westview Press. Smith, Adam. 1776. An Inquiry into the Nature and Causes of the Wealth of Nations. Edinburgh: A. and C. Black. Solow, Robert M. 1956. “A contribution to the theory of economic growth”. In: The Quarterly Journal of Economics 70.1, pp. 65–94. Steindel, Charles. 1995. “Chain-weighting: The New Approach to Measuring GDP”. In: Current Issues in Economics and Finance, Federal Reserve Bank of New York 1.9. Stern, D.I. 2011. “The role of energy in economic growth”. In: Annals of the New York Academy of Sciences 1219.1, pp. 26–51. Swan, Trevor W. 1956. “Economic growth and capital accumulation”. In: Economic Record 32.2, pp. 334–361. Turchin, Peter. 2010. “Warfare and the evolution of social complexity: a multilevelselection approach”. In: Structure and Dynamics 4.3. Turchin, Peter and Sergey Gavrilets. 2009. “Evolution of complex hierarchical societies”. In: Social Evolution and History 8.2, pp. 167–198. Weber, Max. 1958. “The three types of legitimate rule”. In: Berkeley Publications in Society and Institutions 4.1, pp. 1–11. Wicksteed, Philip H. 1894. An Essay on the Co-ordination of the Laws of Distribution (1932 Edition). London: London School of Economics. Wong, Stanley. 1978. Foundations of Paul Samuelson’s Revealed Preference Theory. New York: Routledge.