PhD_01
Chapter 1: Introduction: Holism
versus reductionism in economic thought
Links to the other chapters:
Title
page, dedication, contents, key words, abstract and acknowledgements
Chapter 1:
Introduction: Holism versus reductionism in economic thought
Chapter
2: the Prisoners' Dilemma
Chapter
3: Arrow's Impossibility Theorem
Chapter
4: The Invisible Hand of God in Adam Smith
Chapter
5: Friedrich Hayek: a Panglossian evolutionary theorist
Chapter
6: Keynes’s methodological standpoint and policy prescription
Chapter 7:
Conclusion; glossary; and bibliography
Chapter 1 Introduction: Holism versus reductionism in economic thought[1]
1.1 Preamble
‘What, if any, is the legitimate rôle of the state in the economy?’ That is a fundamental question – perhaps the fundamental question – for economics. How do (micro level) agent interests and behaviours interact to generate (macro level) social outcomes? Are those outcomes desirable, or should society as a whole, in the form of the state, intervene to modify them? This thesis will investigate these questions and explore the answers that have been given by some characteristic economic thinkers. The thesis thus forms part of an investigation into the views of various writers on the articulation between micro and macro levels in economics, between individual actions and social outcomes, between individual and collective rationality.
The thesis begins, in this chapter, with an introduction to some of the fundamental methodological issues underlying the remainder of the work. A consideration of two problems in twentieth-century political economy – the prisoners’ dilemma and Arrow’s impossibility theorem – will then establish the currency of these themes in contemporary political economy. Attention will then turn to Adam Smith’s ‘invisible hand’, and the group evolutionary theory of Friedrich Hayek. The question addressed is, What is the mechanism by which these writers supposed that individual (micro) rationality translated into collective (macro) rationality? In conclusion, an alternative twentieth-century response to this issue, that of John Maynard Keynes, will be considered.
This structure is dictated by the following considerations. Firstly, the issues of micro and macro, of disjuncture and emergence, of reductionism and holism, and of a providential or indifferent world, are ones which continually re-emerge in political economy. The two issues with which I start the thesis establish that point by reference to two controversies of the second half of the 20th century. Of these, the prisoners’ dilemma turns out to be of fundamental significance for the rest of the thesis. Arrovian impossibility is not directly of the same level of significance, but does play a useful role here, both in illustrating some of the themes of the thesis, and in highlighting just why the prisoners’ dilemma, by contrast, is so important.
The bulk of the thesis, the most important two chapters, consider the invisible hand mechanisms of Smith and Hayek. Given the topic of the thesis, the articulation of micro actions and macro outcomes, it was essential to consider the invisible hand, firstly in its original, and secondly in its modern incarnations. In this way we may see what is enduring and what has changed in the presentation of this theme by providentialist political economists. An important finding of these chapters is that both Hayek and Smith reject the reductionist approach adopted, for example, by modern monetarist and new classical writers in favour of a holistic methodological approach. Finally, the chapter on Keynes is required in order to show what happens when a holistic methodological approach is accompanied by an explicit rejection of the invisible hand: in an indifferent, non-providential world we ourselves are obliged to take responsibility for the unintended consequences of our actions.
The present chapter introduces the thesis by means of a consideration of a recent paper by Mario Bunge (Bunge, 2000) ‘Systemism: the alternative to individualism and holism’. Firstly, I argue, with Bunge, that schools of thought in economics may be characterised according to their stance on a key methodological opposition: that between holism and reductionism. Secondly, I argue that this choice of standpoint has important consequences for policy prescription. Arguing against a simplistic correlation of reductionism and laissez faire, the case is made for two kinds of methodological underpinning for laissez-faire: firstly, reductionist and, secondly, holist plus invisible hand mechanism. On the basis of these methodological preliminaries, the chapter concludes by outlining the subsequent structure of the thesis.
1.2 Reductionism and holism: a response to Mario Bunge
This section and the next constitute a response to Mario Bunge (2000) ‘Systemism: the alternative to individualism and holism’. Bunge’s paper is an extremely interesting article which makes a number of telling points and is evidence of a growing discomfort with the reductionism of the neoclassical school currently hegemonic within the discipline of economics. However, there are two major points which need to be made with respect to the article. Firstly, some translation is required, since the terms ‘holism’ and ‘reductionism’ are used differently in his article and in the present thesis. The second point is that the relation between policy prescription and philosophical standpoint in political economy is both more complex and more fascinating than Bunge’s account would lead one to believe.
The first of these points – that concerning translation between Bunge’s terminology and mine – is considered in this section, which then proceeds to make some initial comments on the significance for political economy of the opposition between reductionism and holism. The next section looks in more detail at the second question, that of the relationship between methodological standpoint and policy prescription in political economy.
Bunge’s thesis is that there are three fundamental research approaches in the social sciences: the two most influential, individualism and holism, being fatally flawed, with only the minority approach of systemism offering a viable way forward. The first two are inadequate – each managing to avoid the other’s error only by committing an opposite error of its own, while the third, systemism, manages to synthesise the other two, accepting the criticism each makes of the other.
For Bunge, ‘holism’ is inadequate ‘because there are no relations without relata’; and ‘individualism’ so ‘because all individuals are interrelated’(147[2]):
“Neither of the two most influential approaches to the study and management of social affairs is completely adequate … individualism is deficient because it underrates or even overlooks the bonds among people, and holism, because it plays down or even enslaves individual action. By contrast, systemism makes room for both agency and structure.” (156-157)
Systemism, apart from being defined negatively with respect to individualism and holism, is characterised by viewing the world from a systems perspective: ‘everything is either a system or a component of a system and every system has peculiar (emergent) properties that its components lack’ (147).
Now, there is nothing in these formulations which is contentious[3], and the points made are valuable ones. However, I think greater clarity can be obtained by stating the matter slightly differently. We are fully justified in endorsing Bunge’s rejection of (what he calls) ‘holism’ as ignoring the fact, that the relations, of which our systems of relations consist, depend fundamentally upon the properties of the relata, the substrate-level entities which are doing the relating. What Bunge refers to as ‘holism’ makes the properties of an entity independent of the properties of its material substrate. This, however, is not the way that the term holism will be employed in this thesis.
We are equally justified in complaining, with Bunge, that what he refers to as ‘individualism’ fails to take into consideration the relationships between agents, the fact that individuals are only nodes in systems of such relationships, and the emergence of properties at the macro, or system level. But, when we do so, we are taking a position on an opposition which has implicitly or explicitly underlain a vast amount of methodological discourse in economics and elsewhere: that between holism and reductionism. In defining ‘individualism’, which I refer to as reductionism, Bunge is implicitly defining a non- or anti-individualist approach (which includes both ‘systemism’ and ‘holism’ in his terms), which I refer to as holism. Essentially, reductionism involves a strategy of interpreting the things we see in the world, and, in particular, economic phenomena, as congeries of substrate-level entities; and holism is the attempt to understand these phenomena as a whole or system, with emergent properties not enjoyed by constituent components. The terms holism and reductionism as used in this thesis may be defined as follows:
Reductionism: the view that an entity at one level can be understood as a congeries, an aggregate of entities at a lower, substrate level, that the properties and behaviour of higher level entities can be understood in terms of the properties and behaviour of its constituent lower level parts, taken in isolation.
Holism: the view that phenomena at one level can be understood as emergent at that level, that a higher level entity can be understood as a product of the interrelationships between its component parts.
Correspondence relating to previous manifestations of this chapter suggests that some discussion is in order at this point. Firstly, the question has been raised (Bunge, personal communication) as to whether ‘reductionism’ and ‘holism’ so defined are ontological or epistemological standpoints. As stated, both definitions are epistemological – ‘an entity … can be understood as …’. A logical further step, adopting a realist standpoint, would imply that a phenomenon can be understood in a particular way precisely because that is the way it is. In that vein, as well as being explicitly epistemological, the definitions would be implicitly ontological. However, that further step is unnecessary for present purposes.
Secondly, the exact definitions of the terms are as controversial as everything else in the debate between the supporters of each point of view. The above definitions are my own, and are as unlikely to please as any others[4]. Correspondents, responding to earlier papers on this theme, have criticised this or that statement about reductionism or holism – but, because their criticism was on the basis of other definitions than those set out above, they failed to engage with the points that I am making[5]. I submit that what is critical here is that holism and reductionism so defined can be shown to characterise two living trends in economic thought and to throw up interesting and enlightening questions about the nature of those trends.
Keynes, for example, in the ‘Preface’ to the French edition of 1939 of the General Theory, clearly advocates a holistic approach:
“I have called my theory a general theory. I mean by this that I am chiefly concerned with the behaviour of the economic system as a whole .... And I argue that important mistakes have been made through extending to the system as a whole conclusions which have been correctly arrived at in respect of a part of it taken in isolation.” (GT: xxxii[6])
Keynes sets out very clearly here what he takes to be the distinguishing feature of the two approaches: that, on the one hand, we can derive correct conclusions from the study of microeconomic phenomena ‘taken in isolation’, but that to extend those conclusions to macroeconomic phenomena leads to error, and, on the other, that the correct approach is (what we would now call) a systems approach, aiming to examine the behaviour of ‘the economic system as a whole’. The fallacy of composition involved in the approach he is criticising here, derives from a difference, between the two levels, in what it is legitimate to take as parametric or given (Keynes, 1973: 293).
Robert Lucas, on the contrary, is a very clear spokesman for the trend in economics which favours a reductionist methodology. The following is taken from the final paragraph of his Models of Business Cycles:
“The most interesting recent developments in macroeconomic theory seem to me describable as the reincorporation of aggregative problems such as inflation and the business cycle within the general framework of ‘microeconomic’ theory. If these developments succeed, the term ‘macro-economic’ will simply disappear from use and the modifier ‘micro’ will become superfluous. We will simply speak, as did Smith, Ricardo, Marshall and Walras, of economic theory. If we are honest, we will have to face the fact that at any given time there will be phenomena that are well-understood from the point of view of the economic theory we have, and other phenomena that are not. We will be tempted, I am sure, to relieve the discomfort induced by discrepancies between theory and facts by saying that the ill-understood facts are the province of some other, different kind of economic theory. Keynesian ‘macroeconomics’ was, I think, a surrender (under great duress) to this temptation. It led to the abandonment … of the use of the only ‘engine for the discovery of truth’ that we have in economics” (Lucas, 1987: 107).
Here we have a clear expression of the desire to reduce macroeconomics to microeconomics, and a characterisation of the Keynesian approach as an illegitimate ‘surrender to temptation’.[7]
Both the passages cited occur in contexts – a preface, and the concluding paragraph of a book – where the authors are standing back from the detail of the theories that they are presenting, and indicating what they regard as the underlying general features of their approaches. What they choose to highlight in both cases is their selection of a holist or reductionist approach. This, I think, establishes, at least a prima facie case, that the issue is worth looking at and potentially useful in understanding the controversy between various schools of thought in the history of economics.
* * *
Some examples can be given to show how this controversy continually emerges in economics:
Dore in an extended review of Binmore (1994), comments thus on the latter’s deployment of evolutionary game theory: ‘with [Binmore’s] new definition of self-interest, methodological individualism, a hallmark of neoclassical economics, is abandoned. In its place is a more holistic and integrated view of society as an integral organism’ (Dore, 1997: 227). According to Lawrence Boland, ‘Demonstrating the dependence of all macroeconomics on microeconomic principles is essential for the fulfillment of the (methodological) individualist requirements of neoclassical economics’ (cited in Nelson, 1984: 576). So Dore and Boland agree that the reductionist standpoint – ‘methodological individualism’ – is an essential component of neoclassical economics. In a similar vein, Schotter, right at the beginning of a book on Free Market Economics, says that in the libertarian, or classical liberal view, ‘society is nothing more than the aggregate of the individuals composing it’ (Schotter, 1985: 2). That is, before anything else, Schotter feels it important to establish the reductionist standpoint of libertarianism.
Kevin Hoover, on the other hand, in a special edition of The Monist on ‘The Metaphysics of Economics’, argues that
“The idea that macroeconomics stands in need of a microfoundational base is a commonplace among economists … [but] ontological reduction of macroeconomics to microeconomics is untenable … while the program of microfoundations may illuminate macroeconomics in various ways, it cannot succeed in its goal of replacing macroeconomics.” (Hoover, 1995)
Feeling it necessary to point out that an example she wants to use for purposes of illustration, although couched in reductionist terms, could be replaced by a more holistic version, Nancy Cartwright remarks that ‘We could attempt to explain macroeconomic regularities by reference to capacities and relations that can only sensibly be attributed to institutions or to the economy as a whole, with no promise of reduction to features of individuals’ (Cartwright, 1995: 278).
Nelson (1984) is an interesting case since at first he seems enthusiastic about the possibility of reducing macro to micro in economics:
“The possibility of reducing at least parts of macroeconomic theory to microeconomic theory is especially exciting …. [I]t seems that the large–scale phenomena dealt with in macroeconomics must be the results of the total effects of the small scale phenomena dealt with in microeconomics. Therefore, one might expect that bridges could be built by merely adding up the microeconomic laws describing the microphenomena to obtain the macroeconomic laws describing macrophenomena …. [I]t does seem that a reduction of macroeconomics to microeconomics would not be plagued with the kind of ontological difficulties which might attend [other reduction] schemes” (Nelson, 1984: 573-74).
His conclusion, however, is pessimistic. There is, he concludes, ‘little prospect of providing anything which deserves to be called microfoundations … it is unlikely that we will see any progress in reducing macroeconomics to microeconomics’ (Nelson, 1984: 593). Cross and Strachan agree that ‘The way forward … is to eschew the reductionist search for finer-grain microfoundations and instead [to] study how complex economic systems can emerge from the interactions between agents’ (Cross and Strachan, 1997: 565).
Finally, Hayek, who spent much of his life criticising the linked errors of macroeconomics, economic statistics and socialism, was amongst the most trenchant of those wishing to deny the legitimacy of macroeconomics: ‘I recognize … microeconomic theory as the only legitimate economic theory’ (KES: 22); ‘I believe it is only microeconomics which enables us to understand the crucial functions of the market process’ (KES: 27); ‘[M]y disagreement with that book [sc Keynes’s General Theory] did not refer so much to any detail of the analysis as the general approach followed in the whole work. The real issue was the validity of ... macro-analysis’. (TBT: 100). As we shall see, however, the citation of these passages by no means settles the question of Hayek’s stance on the methodological issue of reductionism versus holism.
* * *
This section has, firstly, argued that while Bunge’s triad individualism-systemism-holism is a profound one, the opposition between individualist, on the one hand, and non- or anti-individualist approaches, on the other, is of more fundamental interest for the purposes of this thesis, and, secondly, that for the remainder of this thesis the ‘individualist’ approach will be referred to as reductionism, and the non-individualist approach as holism. Further, I have proposed definitions of reduction and holism and tried to show that the opposition between them is a living issue in economics. The next section takes up the issue of the association, which can be seen to emerge here, and which is asserted by Bunge, between methodological premises and policy prescription.
1.3 Policy prescription and social philosophy: reducibility and the invisible hand
A pattern seems to emerge from the examples cited: apparently there is a tendency for libertarians and those on the right of the spectrum of policy views within economics, such as Lucas, to appeal to reductionist methodological premises, while those on the left, those like Keynes adopting a more interventionist stance, are more likely to invoke holistic underpinnings for their theoretical pronouncements. Bunge takes up this point, speaking of the ‘obsolete individualism of … Smith … [and] the neoclassical economists’ (147); ‘the cultural policy of liberalism, which is based on individualism, is one of benign neglect. By contrast, the totalitarian cultural policy, which is based on holism, is one of censorship’ (150-151); ‘all market worshippers espouse individualism’ (151). ‘The radical individualists oppose all social planning in the name of individual liberties … holists swear by top-down planning … they are likely to ignore their aspirations and rights [sc those of the common people]. In either case, the powerless individual, whether forsaken or corralled, has nothing to gain’ (153). By contrast, ‘systemism takes into account social values (ignored by individualism) as well as individual values (ignored by holism)’ (157).
But there is a paradox[8] here. We have just seen that there seems to be an association between a laissez-faire policy prescription and a reductionist methodology. Nevertheless, starting with Adam Smith, a profoundly influential trend, epitomised by writers such as Friedrich Hayek and Armen Alchian, has proposed what has been called an ‘invisible hand theorem’[9],[10] (De Vany, 1996: 427), in which order in human affairs is ‘emergent’, ‘resulting from human action but not design’[11]. This emergence seems flatly to contradict the association between laissez-faire and reductionism just noted. This section will explore that puzzle.
Perhaps the canonical statement of the ‘invisible hand theorem’ is Smith’s statement that
“It is not from the benevolence of the butcher, the brewer, or the baker, that we expect our dinner, but from their regard to their own interest. We address ourselves, not to their humanity but to their self-love, and never talk to them of our own necessities, but of their advantages.” (WN I.ii.2[12])
Even here, in this well-known and apparently simple statement, there is something mysterious about the relation between micro and macro levels. The butcher, brewer and baker do not care about the dinners they provide, so that, in some sense, the desirable social outcome of feeding the members of society is achieved in spite of rather than because of the motives and behaviours of the food providers. The articulation between the motivation and behaviour of agents at the micro level and macro outcomes is not even at first blush a trivial or straightforward question. It seems odd that the trend which is associated with this ‘invisible hand theorem’, the tendency to support a laissez-faire policy prescription, which as we have seen seems itself to be associated with reductionism, should refer to the ‘emergent’ properties of human order, when emergence is precisely what distinguishes the holist from the reductionist approach.
The question therefore arises, whether the invisible hand theorem is consistent with the reductionist or the holist approach. Haworth (1994 Ch 4, Section 1 ‘A logical dilemma’, 32-35) performs an invaluable service by addressing precisely this issue. Haworth’s procedure – part of his philosophical critique of libertarian thinking – is as follows. Firstly, he identifies two ‘theses’, implicit in libertarian thought, which he defines as follows, each illustrated by a statement from a libertarian[13] source. Then the libertarians’ logical dilemma arises from the mutual incompatibility of the two theses:
The reducibility thesis: the fully developed market economy can be understood as the sum or aggregate of its discrete components, the individual bilateral exchanges at the micro level. Sir Keith Joseph: ‘Since inequality arises from the operation of innumerable preferences, it cannot be evil unless those preferences are themselves evil.’[14]
The invisible hand thesis: the market is a ‘paradigmatic exemplar’ of want-satisfaction (ie, unrestricted market forces leave agents better off than any alternative economic environment) because an invisible hand transmutes our self-interested behaviour into socially desirable outcomes. Mandeville: ‘the grand principle that makes us social creatures, the solid basis, the life and support of all trade and employment without exception is evil.’[15]
So reductionism says that evil only comes of evil, while the job of the invisible hand is specifically to transmute evil into good[16]. For Sir Keith the aggregate outcome cannot be evil as long as the preferences it is based on are innocent; for Mandeville, on the contrary, the aggregate outcome cannot be good unless the preferences underpinning it are evil, vicious, selfish. Thus Haworth is able to conclude that the libertarians cannot have it both ways: ‘libertarianism is seriously broken backed in the sense that it must abandon one of its central theses.’ (Haworth, 1994: 34)
My own approach is slightly different from Haworth’s. Firstly, we need to pay more attention to the association, alluded to earlier, between methodological standpoint and policy prescription. There are policy implications of the choice between reductionism and holism. And, indeed, the consequences for policy implied by the approach selected, so far from being a mere scholium, are the tail which wags the methodological dog[17]. If one adopts the systems approach and recognises that the unintended collective outcomes of an unplanned, uncoordinated mass of individual actions may have far from desirable features, then the obvious implication is to see whether there is anything we can do about it. The absence of an invisible hand invites the intervention of the very visible hand of state intervention. The reductionist approach, on the contrary, says that, assuming individuals can be counted on to do the best they can for themselves given the constraints they face, the aggregate outcome of those individual actions will also be the best available: state intervention in the economy is nugatory.
There are two possibilities: we could be living in a world where the reductionists are right or one where the holists are right. Needless to say, I think we reside in the latter. If we lived in the former, the macro level would simply reflect the micro level. There would be nothing for an invisible hand to do. The individual would be directly social, or, what comes to the same thing, there would be no separate category of the social. Individual utility maximisation would directly be social welfare maximisation: the distinction between them would be meaningless. Likewise, macro irrationality would be just a summary of micro irrationality: unemployment would either be a product of irrational behaviour by workers, such as ‘pricing themselves out of jobs’, or it would be the product of a rational desire for leisure[18], and, hence, itself rational. In general, individuals could with confidence be left to get on with it without supervision or intervention. A reductionist world would be a laissez-faire world.
If, on the other hand, we were to inhabit, as in my opinion we do, a holistic world[19], then reductionists would (and do) have a problem. It is fairly obvious that higher level entities are not simply aggregates of their micro components: water does not behave as an aggregate of hydrogen and oxygen; steam, liquid water, and ice do not consist of tiny gaseous, liquid and solid molecules; nor do chairs consist of hard, green, ugly or uncomfortable molecules (Haworth, 1994: 35). As Ken Binmore says, ‘As we all know, you would be very lucky if a small change in your program led only to a small change on your screen. Tiny programming errors typically lead to wild and unpredictable results when the program is run’ (Binmore, 1994: 213). All these properties emerge at higher levels. The problem faced by the reductionist is how to reconcile this fact – of an obvious disjuncture between levels – with the reductionist laissez-faire policy prescription. Libertarians face severe difficulties sustaining a logically consistent reductionism in a holistic world.
The invisible hand is one potential solution to this problem. There are two possibilities. Either one can ignore the disjuncture between levels, and adopt a thoroughgoing reductionist methodology and policy stance – this seems to be line taken by Joseph, Lucas and Friedman[20] – or with Hayek and Adam Smith one can accept that disjuncture, and so adopt a methodological holism, but at the same time postulate a mechanism reconciling that methodological holism with a laissez-faire policy reductionism. Such a mechanism is an invisible hand mechanism. The invisible hand allows us to say, granted that social outcomes are not logically bound to reflect individual behaviour in an aggregative, summary manner, nevertheless a mechanism exists which ensures that in practice they do so. The invisible hand is what allows us to think, and act, in a reductionist way in a holistic world: it underpins reductionism by tacitly conceding holism. Laissez-faire is vindicated, we are inveigled into tying the visible hand behind our back, if we can be persuaded that the invisible hand will do its job instead, and do it better.
What I am suggesting, therefore, is the following: the laissez-faire policy prescription does, indeed, embody a reductionist standpoint. However, there is more than one way of sustaining that standpoint methodologically. One can believe, or at least act as if one believes, that the world truly is reductionist in relevant ways and that supposed macro-level pathology is simply the summation of micro-level behaviour which may or may not be pathological. Laissez-faire is a reductionist policy prescription in the sense that it issues from a reductionist methodological standpoint. Or one can accept that the world is holistic and hence that macro-level pathologies might in principle be emergent at that level, but postulate the existence of an invisible hand mechanism which ensures that the reductionist policy prescription of laissez-faire is nevertheless valid. The latter strategy combines methodological holism with policy reductionism.
* * *
Bunge’s criticisms of the ‘individualism’ (for which, following the terminology I adopt, read reductionism) of the neoclassical economists are well taken[21]. But the simple, one-to-one relationship between this reductionist standpoint and a laissez-faire policy prescription, which Bunge asserts, simply doesn’t exist. Compare the standpoints of Hayek and Friedman. For Friedman, economics is based on the study of ‘a number of independent households - a collection of Robinson Crusoes’ (Friedman, 1962: 13; see also Haworth, 1994: 8). For Hayek, on the contrary, ‘individuals are merely the foci in the network of relationships’ (CRS: 59). So, on the definitions proposed earlier, Hayek subscribes to a very clearly holistic, and Friedman to an equally clearly reductionist methodological standpoint. Yet they still both endorse the same basic framework for policy prescription: laissez-faire. And in bracketing Smith with the neoclassicals in the reductionist camp (148), Bunge is simply in error – as we shall see in Chapter 4.
Writers such as Smith and Hayek are methodologically very distant from the crude reductionism of Joseph, Lucas and Friedman. They tacitly recognise a holistic world by invoking invisible hand mechanisms. For Smith the invisible hand is literally the hand of an omniscient and omnipotent deity desiring nothing other than the maximisation of human welfare. For Hayek, writing in a more secular age, the invisible hand mechanism takes the form of an evolutionary process based on the exploded group selection theory of VC Wynne-Edwards. Hayek’s attempts to distinguish his own stance from that of Keynes, his anxiety to head off a line of thought leading from the holistic or systems thinking premises he shared with Keynes to an interventionist policy prescription, lead him to make the crudely reductionistic statements about macroeconomics and microeconomics which we noted earlier.
The resolution, then, to the puzzle identified at the beginning of this section is as follows. We can understand the contradictory association between an appeal to reductionist methodological underpinnings and an assertion of a holistic invisible hand mechanism on the basis of a dual foundation. The first part is a factual hypothesis about the nature of the world we actually inhabit, namely, the systems view of the world, the hypothesis that the world is a holistic one with all that this implies about the scope and potential for collective intervention in the economy. The second part concerns the laissez-faire policy prescription of the reductionist camp, a policy prescription which both depends on and supports the reductionist methodological stance. Taking the two points together, we can see that some mechanism has to be introduced to mediate between a holist world and a reductionist policy prescription: the invisible hand does the job. The invisible hand, whatever its precise content, whether it comprises the hand of a deity as in Smith, or the result of an evolutionary process as in Hayek, allows us to assert that, in practice, the world can be treated as if it were reductionist. The logical inconsistency which Howarth has correctly identified is an internalisation of the inconsistency between reductionist laissez-faire programme and holistic world. It is illogical because it is attempting to do what is ultimately impossible, namely to reconcile the irreconcilable.
The alternative to both of these approaches is to combine recognition of the holistic nature of the world we live in with acceptance that there is no invisible hand. In this view, rational individual self-seeking behaviour is by no means either the necessary or the sufficient micro substrate for the desirability of social outcomes. Rather, behaviour must be directly social if desirable social outcomes are to be obtained. According to Keynes, for example, egotistical activity uncoordinated by the state may lead to inefficient outcomes. The price system aggregates rational individual actions but the aggregate is an unintended outcome as far as those individuals are concerned. There is no particular reason why unintended outcomes should necessarily be desirable and often they are not. Individuals take responsibility for maximising their own welfare, given what everyone else is doing, but somebody[22] has to take responsibility for organising the aggregate outcome if undesirable aggregate outcomes are to be avoided: ‘there is no design but our own ... the invisible hand is merely our own bleeding feet moving through pain and loss to an uncertain … destination’ (CWXX: 474).
Marx, on the other hand, takes the argument a stage further by arguing, on the contrary, that there is, indeed, a design which is not our own, a design without a designer. Like Hayek, Marx believes this design to be ‘emergent’ at the macro level, but for Marx, unlike Hayek, because it is not our own design, it is alien to us. In the absence of directly social activity, atomistic behaviour spontaneously arranges itself into a self-augmenting parasitic network of social relations which he calls ‘capital’. A society of individual humans thus becomes dominated by an interest alien to that of the individuals comprising it.
* * *
The conclusion of this section, therefore, is that while reductionist approaches may be safely dismissed, embracing holism is no guarantee of getting it right. There is holism and there is holism. Methodological holism combined with the deus ex machina of an invisible hand mechanism can still sustain the inappropriate and unwarranted reductionist policy prescription of laissez-faire. The bulk of this section has concerned the holism-reductionism dyad, at the cost of ignoring the distinction between what Bunge calls ‘systemism’ and what he calls ‘holism’. Bunge drew attention to the important principle that ‘there are no relations without relata’. We both reject the reductionist assumption that the properties of entities are just aggregates of the properties of substrate entities, but we also reject the assumption that the pattern of relationships constituting a system is substrate-neutral and in some sense logically prior to the properties of the substrate entities.
The implication is that writers in the invisible hand tradition, such as Smith and Hayek, adopt precisely this assumption. As far as Smith is concerned, the case is made in Chapter 4, where the invisible hand is interpreted as literally the hand of an omniscient, omnipotent and benign deity. As far as Hayek is concerned, the issue is dealt with in Chapter 5, where I argue that, for Hayek, macro level objects are understood as independent entities in their own right, owing nothing to their material bases in individual behaviour. In both cases the overall macro-level pattern is divorced from the need for a mechanism, such that micro level incentives are consistent with behaviour which sustains that pattern.
The consequence of this analysis is that the association between methodological standpoint and policy prescription asserted in Bunge (2000) can be challenged. There is no simple association between ‘individualism’ and laissez-faire, on the one hand, nor between ‘holism’ and intervention or centralisation, on the other. On the contrary, both ‘individualism’ (that is, in my terminology, reductionism) and ‘holism’ (a holism cut off from material foundations in self-seeking substrate activity) are associated with laissez-faire, while ‘systemism’ (holism which is not so cut off) is associated with a range of policy prescriptions from mild to radical intervention.
In substance, Bunge’s article is a significant contribution to our understanding of the methodology of economics – perhaps best embodied in his bold injunction to ‘see agency through Weber’s microscope, and structure through Marx’s telescope’ (154). The article, however, can be faulted for an over-simplification and misunderstanding of the relation between methodological premises and policy consequences. There do exist reductionist free marketeers of the kind Bunge describes: the pronouncements of Friedman, Lucas and Sir Keith Joseph fit this pattern. But to lump holists such as Smith and Hayek in with them, is both mistaken and allows us to ignore the fundamentally ideological role of invisible hand mechanisms in allowing economists to retain some approximation to efficiency as their default notion of how the capitalist economy actually works.
1.4 The structure of the thesis
From what has been said above, it is clear that a major part of the thesis must be concerned with the problem of the invisible hand. It is mandatory, then, to return to the locus classicus of the invisible hand in the writings of Adam Smith. This will occupy a pivotal chapter, Chapter 4, ‘The Invisible Hand of God in Adam Smith’. This will be followed, in Chapter 5, by a close examination of the position of a leading twentieth century invisible hand theorist, Friedrich Hayek. The thesis concludes with an examination of the consequences of adopting a holistic perspective while rejecting invisible hand mechanisms: Keynes’s proposals for planning and widening state engagement in the economy. Before turning explicitly to the invisible hand itself, however, I examine two episodes in twentieth century political economy where developments within game theory and the theory of social choice were perceived as challenges to ‘the invisible hand theorem’.
Within modern neoclassical economics controversy has been aroused by the discovery of two apparent anomalies or paradoxes. Arrow’s impossibility theorem, for example, appears to show that it is impossible to devise a procedure which can be relied upon to aggregate individual preferences into clear and acceptable collective preferences. And the prisoners’ dilemma seems to contradict the view that individual utility maximisation leads in general to collectively desirable outcomes – not just as a theoretical possibility but as a plausible description of a pervasive phenomenon. Considerable energy has been invested in the attempt to reconcile the prisoners’ dilemma and Arrovian impossibility with the supposition of a ubiquitous and benevolent invisible hand.
One response to such difficulties, however, has been to jettison the idea of collective rationality altogether. Libertarian writers such as IMD Little, argue that only individuals can think or prefer one thing to another, and concepts of collective rationality and social preference are therefore devoid of meaning. The question as to whether a collection of individuals, a society, can have interests and preferences distinct from those of the individuals of which it consists, has parallels with two related questions in psychology, philosophy and computer science, namely, (a) how brains can be conscious when individual neurons are not, and (b) whether artificial intelligence is possible in principle. Both questions relate to the validity of the computational theory of mind.
Chapter 2 concentrates on the challenge to ‘the invisible hand theorem’ posed by the prisoners’ dilemma. Attention will be drawn to a key feature of the problem facing agents when such dilemmas arise: that decision-making is interdependent in content but independent in form. The decision affects others but is taken as if it only affected the individual decision-maker. In Chapter 3 attention is turned to the problem of Arrovian impossibility, and, in particular, to the criticism of Arrow by libertarian thinkers. Taking IMD Little as a principal exponent of this view, a parallel is drawn between his rejection of the notion of a social welfare function and the rejection, by writers such as John Searle, of the possibility of artificial intelligence. Chapters 2 and 3 take the form of an extended review of Barry and Hardin (1982).
What, then, do we learn, with respect to the issues outlined above, from these two chapters? Firstly, they appraise the seriousness of the challenge posed by the prisoners’ dilemma and Arrovian impossibility to the reductionist programme. The conclusion drawn is that both are, in fact, incompatible with the reductionist programme, but that the prisoners’ dilemma is the more profound, drawing attention (a) to both common and conflicting interests of the agents involved, whereas the Arrow theorem is based only on conflicting interests, and (b) to the tension between social content and private form of the decision making process. Secondly, the chapters draw attention to a particularly extreme liberal response to these challenges, and provide an explanation of why that response is wrong.
As just mentioned, in a holistic world the prisoners’ dilemma and Arrovian impossibility pose considerable difficulties for the reductionist programme. In part this is an empirical matter. Do we really encounter prisoners’ dilemmas[23] and non-single-peaked preferences in the world? One possible approach for the reductionists is to say that these do not in fact occur[24]. However, I am not aware of any major attempt to put this argument and therefore will ignore it here. A second possibility is to pick holes in the argument, trying to find an acceptable ‘solution’ to the two ‘paradoxes’. Ken Binmore says well that ‘Personally, I see no paradox at all in the fact that independent choice behaviour by rational agents should sometimes lead to Pareto-inefficient outcomes’ (Binmore, 1994: 103). But why does he need to say this? Because ‘The claim that [that both players defect] is the solution of the Prisoners’ Dilemma seems paradoxical to many authors.’ (ibid) ‘Many’ is in fact an understatement here, and very great efforts have, unsuccessfully, been made to circumvent the two propositions. The point is that there is no absolute or logical paradox involved, but both Arrovian impossibility and the prisoners’ dilemma are paradoxical in the older, weaker sense of ‘contrary to orthodox wisdom’ – the sense in which the Good Samaritan is a paradox. Belief in the possibility of automatic mechanisms which aggregate (a) individual preferences to achieve a consistent notion of what is socially desirable, and (b) individual actions to achieve socially desirable outcomes, is incompatible with these two results.
A third approach, taken by the more extreme proponents of laissez-faire, is to deny collective rationality and social welfare, to deny, that is, that any meaning can be attached to these phrases. According to this view, only individuals can be rational or have preferences. The lesson of Chapters 2 and 3 is that this tactic, which illegitimately privileges a particular level, that of the individual agent, is fundamentally untenable.
Chapter 4 is devoted to an examination of the invisible hand in the work of the writer who invented the term. Adam Smith is revered as the father of modern economics. I shall argue that analysis of his writings, however, reveals him to be a representative of the 18th century tradition of ‘natural theology’, with some medieval elements. Smith is preoccupied with the need to preserve order in society. His scientific methodology emphasises reconciliation with the world we live in rather than investigation of it. He invokes a version of natural law in which the universe is a harmonious machine administered by a benign deity. Nobody is uncared for and, in real happiness, we are all substantially equal. No action is without its appropriate reward – in this life or the next. The social desirability of individual self-seeking activity is ensured by the ‘invisible hand’, that is, by the hand of a god who has moulded us so to behave, that the quantity of happiness in the world is always maximised.
In Chapter 5 attention turns to a prominent twentieth century exponent of the invisible hand. While Adam Smith proposed that individual self-seeking would lead to socially desirable consequences thanks to the intervention of the ‘invisible hand’ of a wise and kindly god, Friedrich Hayek seeks to re-establish the invisible hand in a secular age, replacing the agency of a deity with an evolutionary mechanism. A process akin to natural selection ensures that individual behaviours leading to undesirable social consequences are weeded out. Hayek’s approach hypostatises the old. Hayek is inconsistent in his deployment of this evolutionary mechanism: only its – for him – benign effects are acknowledged, while its undesirable consequences are blamed on a culture of state intervention.
What is the specific contribution which these two chapters make to the overall thesis? I argued, earlier in this introductory chapter, that the notion of an invisible hand is fundamentally ambivalent: at once it sustains reductionism, and the laissez-faire policy prescription implied by it, while tacitly conceding the holist case: social order is emergent. What we want to know is, precisely how this automatic mechanism operates; how the invisible hand is to reconcile individual behaviours in the interest of the greater good of society.
In Smith and Hayek we see two allied but distinct answers to this question. In the writings of Adam Smith we see that the invisible hand is simply the hand of God. Smith’s God has a very simple structure, indeed he is reduced almost to a cipher. Repeatedly and unambiguously we are told that, at least for practical purposes, God has a utility function of one argument: total human happiness[25]. The machine grinds out the summation with perfect accuracy for ever. It is the totality which is always primary for Smith; individuals are assigned very subordinate roles. Deprived of any real freedom or autonomy, they are inveigled, deceived and frightened into doing God’s will, that is, supposedly, maximising human welfare.
Is Smith then in the reductionist or holist camp? I argued above that the role of the invisible hand is to reconcile the reductionist programme with a holistic world. This is certainly the case for Smith. Smith admits a holistic world. His imagery is, in tune with his times, mechanical rather than organic: the world is a great machine rather than an organism. But that is not significant. What is key is that Smith’s deployment of the invisible hand allows him to assume that the maximisation of human happiness takes place automatically – all we need to do is give each other enough space for the invisible hand process to work itself out. Macro level rationality is the spontaneous reflex of micro level rationality.
And what of Hayek? Hayek’s standpoint is structurally very similar to Smith’s – with the modification that a form of evolution is to replace the deity as the mechanism driving the invisible hand. Not only the allocation of resources between competing ends, but also the institutional environment within which that allocation process takes place, are subject to variation and selection. Only those allocations and institutions most conducive to human welfare survive this weeding and sifting process. Once again, the reductionist programme is reconciled with a holistic world. Social order, and other properties of the macro level, are emergent at that level, but they are born perfect and fully-formed. We have no need to intervene at the macro level. Social welfare is again maximised automatically by individual utility maximisation.
What this, at first blush more plausible, account omits is how individual self-seeking behaviour leads to institutions and allocations which are just what society requires. Indeed, just as in Smith individual interests were pre-reconciled by God, so that aggregation was unproblematic, so, too, Hayek admits that this will only work for interests that are already reconciled. He then goes on to build his entire system on the wholly unjustified assumption that, indeed, they are thus reconciled.
The thesis concludes with a presentation of the policy stance of John Maynard Keynes and an examination of its links with his underlying social philosophy. The argument is that Keynes shares the holistic approach of Smith and Hayek, but without their reliance on invisible hand mechanisms. If spontaneous processes cannot be relied upon to generate desirable social outcomes then we have to take responsibility for achieving this ourselves. Individual self-seeking behaviour will lead to socially desirable outcomes only if the institutional framework is right. Setting up the institutional framework in which there are no prisoners’ dilemmas – or, what comes to the same thing, abolishing the existing institutional framework which atomises individuals – implies large-scale collectivisation of the economy.
Send me an e-mail at: a.denis at city.ac.uk.
Back to my research page.
Revised: Saturday, 13 September 2003