An interview with Naked Capitalism's Phil Pilkington on the state of economics (and our Modern Political Economics): Part B

Yanis Varoufakis is a Greek economist who currently heads the Divison of Political Economy at the University of Athens. From 2004 to 2007 he served as an economic advisor to former Greek Prime Minister George Papandreou. Yanis writes a popular blog which can be found here. His treatise on economic theory ‘Modern Political Economics: Making Sense of the Post-2008 World’, co written with Nicholas Theocrakis and Joseph Haveli is available from Amazon.

Interview conducted by Philip Pilkington.

Philip Pilkington: In the book you talk about how humans are extremely hard to model – you go as far as to call them a ‘radical indeterminacy’. Now, some of this comes back to the classical theory of value that we already discussed. That theory of value seeks out the basis of value in the human capacity for labour – essentially putting humans out front and centre when modelling is undertaken. But the neoclassical theory of value – that is the marginal theory of value – also relies on humans to build models, in that it concentrates on their consumption preferences and derives so-called laws from these. Could you talk about this a little?

Yanis Varoufakis: We are coming closer to the heart of the problem. Classical value theory is anthropocentric; as it ought to be. In a world of robots, there will be no value. There will be function, even economies; but no value. Consider a watchmaker, or an artificial intelligence engineer. They have no reason to speak of the ‘value produced’ by some cog doing its thing in a watch, or by some machine that is embedded in a system of machines. They will, I suggest, speak of the cog’s or the machine’s function, properties, energy input or output. But of value? It would be senseless or, at best, metaphorical.

Humans, on the other hand, have a unique capacity of creating value. Starting with John Locke (who argued that property rights emerge when human labour is blended into a piece of land, thus producing property rights over the latter), classical economists (Smith, Ricardo, Malthus, Marx) also assumed that labour breathes ‘life’ into things, turning them from ‘things’ to ‘valuable things’. Then they asked a series of organic or systemic questions: How does this value determine the relative prices of these things? How is it distributed across society? How does it grow? Under what circumstances can it go into reverse (recessions and crises).

When the marginalists arrived on the scene some time in the 1870s (who later turned neoclassical, with the exception of the Austrians who resisted that ‘transformation’), they treated the notion of value with contempt; as a relic of romantic hocus pocus. In its place, they put another notion: utility. All human activity was to be explained not in terms of the value that labour breathes into things but in terms of the satisfaction of some well-defined preference ordering of human consumers. The main and profound difference between them and the classics was their ambition to pose self-contained models of the individual. (This is why they fell in love with Robinson Crusoe and his one-man economy.)

But to have a self-contained theory of the person, they needed mathematically to close their model of Jack without any reference to Jill. Which meant that Jack’s actions (including his labouring) had to be modelled as a constrained optimisation problem; i.e. an attempt by Jack to be Robinson; to choose his actions in order to maximise his ‘utility’ index given his constraints.

Now, whereas Robinson’s constraints involved the scarcity of tools and materials that he salvaged from his ship, Jack’s constraints involve prices that are, at least initially, taken to be exogenous. The end result was a theory of what Jack will do based on the prices he faces (e.g. the price of the bread he buys or of the labour he hires out to an employer). Lastly, the marginalist economist would ‘close’ the model by finding the prices which ensured that that which Jack wanted to sell Jill would want to buy (also known as general equilibrium theory).

From the last paragraph it is evident that value is designed out of this analysis. There is utility and there are prices, which reflect relative scarcity. You say that marginal-neoclassical models of this sort maintain human beings at their heart. Not really. For the model of Jack and Jill that we end up playing with are well and truly dehumanised. Indeed, once we have a snapshot of their preferences, and a list of their constraints, we do not need them anymore. We can, effectively, jettison them. To see this from another angle, Jack and Jill, in this analysis, cannot possibly do anything. There is no process of creation in that model of things, ideas…

Production in marginal-neoclassical theory is a black box story. You describe the inputs and presume a function (the black box) into which the inputs go and from which the outputs emerge. In the absence of time! For, as I have explained already, to ‘close’ these models mathematically, we cannot afford to allow the clock to tick.

In short, marginalism-neoclassicism confine the idea of value to the dustbin but, to have something to say about prices and quantities, they end up with a model in which only cardboard versions of humans ‘live’ a series of instantaneous lives (without any theory of how they proceed from one snapshot to the next; it is what economists euphemistically refer to as ‘comparative statics’).

PP: You’re right, I think we are getting to the heart of the problem. It seems to me that the neoclassicals’ ‘snapshot’ or ‘static’ model of humans does not allow for any creative change. But it is clear that in the real world such change is needed in order for the economy not to fall into stagnation (perhaps the old Soviet economy provided an example of such stagnation). My impression from your work is that if such dynamic change were allowed into the analysis the economists’ models would fall apart. Could you talk about this a bit and maybe indicate what it might mean for thinking economically?

YV: Let me give you an example, one that is close to my heart as it marked my long term engagement with game theory (perhaps neoclassicism’s highest form). Consider a human activity that is full of drama, strategies, cunning, beliefs, threats, pleading, niceties etc: bargaining and negotiations for the purpose of averting explicit failures to exploit mutual benefits (e.g. between trading parties, employers and unions etc.) or out and out conflict (e.g. between neighbours, states, corporations).

Imagine now that we want mathematically to model this bargaining process on the assumption that the bargainers are rational (for otherwise it is impossible to model their behaviour). Now, consider this proposition:

If we could have developed a brilliant theory of conflict, then the possibility of conflict between rational agents withers (i.e. it tends to zero).

Let me now argue that this proposition is inescapable: Let’s call this brilliant theory of conflict T. T, by definition, predicts the negotiating tactics of each rational party. For example, it predicts how long each one of them is prepared to hold out; i.e. to delay agreement, at a cost, in order to achieve a better outcome. But if the parties are rational, and T is unique, then they must work out what T predicts. Thus, to the extent that disagreement and delay is costly to each of the bargainers, they will agree immediately. To see this, think that if we can both predict (even within a range of commonly shared predictive error) how long our conflict/disagreement will last before we settle, and we can also predict (more or less) what that settlement will be, then why not settle now for the same outcome without the costly conflict?

In this sense, if we assume that there exists a unique theory T that produces these predictions, and we also assume that rational people can (just like the best theory) work out what T predicts, there can be no conflict. Have you noticed the paradox? If a uniquely brilliant model of conflict were to exist, then there would be no… conflict (while we should be able to predict the outcome of negotiations between rational parties).

This paradox means that if you set out to write down the equations of a ‘closed’ model of bargaining, then (i) you may well succeed in ‘discovering’ a uniquely ‘rational’ agreement and (ii) you will denounce all instances of conflict as a failure of rationality [since according to the last sentence in the previous paragraph there can be no conflict if there exists a uniquely brilliant theory of bargaining].

The above illustrates nicely the economists’ tragedy: The ambition to create the ultimate model results in a theory of human behaviour which cannot hold unless the humans populating the model operate like inanimate algorithms following slavishly a particular script. The moment you introduce creativity and a capacity to subvert the rules that supposedly govern our behaviour, the whole theoretical edifice collapses. To illustrate this further, take again theory T.

Suppose Jack and Jill are negotiating and both are rational enough to know T. So, Jill expects Jack to expect her to behave according to T. Because she is human (i.e. not an automaton) she has the ‘right’ to ask a ‘subversive’ question that no algorithm can ask: “What will Jack think if I behave in a manner that T has not predicted (something I can do since I know what T predicts)? Is there no significant probability that Jack will panic, thinking that I am irrational (since I have diverged from T), and choose to yield to some of my demands more readily?”

There is, I wish to argue, nothing irrational about this question. Which means that it is possible that Jill will rationally diverge from the bargaining behaviour that theory T prescribes as the uniquely rational behaviour of Jill. But this is another, more ruthless, contradiction:

Theory T supposedly prescribes Jill’s uniquely rational bargaining behaviour while, at the same time, Jill can rationally contradict theory T.

In summary, as you say, introducing dynamic change and genuinely human capacities for rational subversion into the analysis causes the economists’ models to fall apart.

(*) You may wish to note that this paradox was the starting point of my 1991 book Rational Conflict, Oxford Blackwell

PP: I can’t help but note that the more I look at these arguments the more I detect a strong sense that they are shot through with problems that we generally refer to as ‘sexual politics’ (and I mean that in the colloquial sense of the negotiations that daily take place between the sexes rather than anything to do with feminism) – but I think to go in that direction would be a whole different, if perhaps interesting, discussion.

Personally, I’ve long thought that social science – and in this economics is only one example – should simply banish determinism when it is applied to the individual (‘situational determinism’, I believe it was called by Spiro Latsis). I’ve also long thought that this would mean the liquidation of microeconomics as a discipline. Would you agree with these sentiments? If not, what would you consider worth salvaging from the wreckage?

YV: Yes, I do. The problem with economics is that, at best, it can offer an interesting theory of what an economy populated by algorithms will look like. A type of Matrix economy (here I am referring to the Wachowski brothers flick) in which machines produce other machines, and the ‘stuff’ that keeps them going, on the basis of complex but given instructions, or algorithms. In such an economy two things are impossible: (a) value creation and (b) rational behaviour that is premises on the subversive question “And what if I do not behave according to the rules that ought to govern my behaviour?” Such economies can be modelled mathematically. Human economies, in which value is produced because the human imagination is irrepressible and unbounded (and thus capable of subversive thoughts and ideas that bamboozle even the smartest algorithmic computer), cannot be similarly modelled. In the end, it is as simple as that.

Which brings me to your final question: What can be salvaged from the theoretical wreck that is economics? My answer is: The process of discovering the limits of analytical reason. By studying critically all models, we end up none the wiser about quantitative outcomes but much, much smarter about the complexities of really existing capitalism. Our exploration of economics may take us, in the end, right back to the point we started: Not having a clue about when the next crisis will hit, what sectors will dominate, whether the stock exchange will pick up soon or not. But, while we shall not have a determinate model of prices and quantities, we shall be much more appreciative of capitalism’s motivated irrationality, its penchant for surprising even the powers that be, its capacity to create incredible wealth and untold suffering by means of precisely the same process. Or, as T.S. Eliot once wrote:

We shall not cease from exploration

And at the end of all our exploring

We will return to where we started

And know the place for the first time

– Little Gidding, one of T. S. Eliot’s Four Quartets

PP: Right, well that certainly takes care of a certain type of theoretical economics, but my understanding is that this sort of economics is based mainly on microeconomics. It seeks to explain behaviour at the level of the individual or the firm. The contradictions are so replete I have to agree with you: the only point in studying it is to understand how limited it is.

But what about macroeconomics? This wasn’t really dealt with in the book, although it was used in an applied manner in the non-theoretical part of the book. What do you think of the research potential for macroeconomics and its ability to solve problems? I’m alluding specifically to those schools that claim that the Keynesian revolution in economics was aborted by shifting the focus to the microeconomics that we’ve been discussing. Might this not be a fruitful path of study?

YV: It is perfectly true that, sometime in the 1970s, macroeconomics was replaced by a microeconomic model of some Robinson Crusoe-like ‘representative agent’ whose microeconomic behaviour was assumed to mimic that of a complex capitalist economy! In effect, the distinction between microeconomics and macroeconomics was eradicated, with the former occupying the driver seat. At this point it is worth remembering that macroeconomics was first ‘invented’ by John Maynard Keynes in a bid to understand the Great Depression. In essence, macroeconomics was a child of the midwar Crisis, one not too dissimilar to the one our generation ‘entered’ courtesy of the Crash of 2008. What was Keynes’ main contribution? How did he carve macroeconomics out of the preceding micro-based marginalist economic analysis?

Keynes’ central idea was that two markets behave differently to the rest: the market for labour and the market for money (or capital). Whereas in the vegetables market a price fall will eliminate an excess supply of cabbages, when the economy slumps a fall in the wage, i.e. in the price of labour, will not reduce the excess supply of labour (i.e. unemployment). And, similarly, a fall in the price of money, i.e. in the interest rate, will not reduce the excess supply of savings (i.e. it will not stimulate investment).

The failure of these two markets (labour and money) to operate along the lines of microeconomic thinking (i.e. a reduction in ‘price’ leading to an increase in ‘sales’) meant that the ‘laws’ of a macroeconomy are different to the laws that apply to firms and households. And what laws are these? The main idea is that macroeconomies have a tendency to stumble and fall; to produce crises (nothing new here, as Marx was the first to make the point). However, Keynes added that, after the Fall, macroeconomies may stay ‘fallen’ for a long time, unable to pick themselves up, dust themselves down and grow again. Once caught in an equilibrium of negative expectations, capitalism is perfectly capable of staying depressed. At that point, it needs a boost. And since interest rate reductions will not do the trick (and are anyhow impossible once interest rates are lingering near zero), the only thing that we can do is have the government take over the role of, at least temporarily, investing.

This is a simple, logical insight for which one has no need of models or determinism. Indeed, one way in which we can sum up Keynes’ major insight is by means of the statement: We are damned if we know! The point here is that, indeed, we have no way of predicting when crises will hit, or even to know why they occurred (not even ex post). In a Socratic twist (“I know one thing; that I know nothing”), Keynes’ wisdom entailed a major concession to indeterminacy followed by a recommendation (government investment) of how simply and efficiently to deal with the consequences of the inevitable, yet utterly unpredictable, crash.

The problem with Keynes was the… Keynesians. Especially those who, beginning with Paul Samuelson, decided that the thing to do is turn the great man’s insights into… models. These models turned out to be extremely naïve and to give the false impression to two generations of young economists that the truth about crises, slumps and fiscal policy is to be found within the confines of these models (which, paradoxically, tried to turn Keynes’ great insight, that we are damned if we know, into a… geometrical model). So, when the collapse of the first post-war phase came in 1971, giving rise to an historical discontinuity (e.g. stagflation), it was ever so easy to show that the Keynesians’ models were wrong. At that point, the baby (Keynes) was thrown out with the bathwater (the Keynesians’ models). Macroeconomics thus, effectively, ceased to exist – returning back to its pre-1929 ways. As a result, the current crop of mainstream economists are just as ill equipped to deal with this Crisis as President Hoover’s lot were in 1929.

To conclude, let us by all means reinstate Keynes’ macroeconomics to its rightful place. But let us, at the same time, ditch the Keynesians’ naïve belief in modelling the weird, wonderful and ultimately un-modellable workings of the labour and the money markets. Then and only then will we be granted a glimpse of really existing capitalism(s).

PP: This seems to all come back to the interpretation of Keynes and the ISLM model – which, I think, is what you mean by the dominant Keynesian school in the post-war era. But, as I’m sure you know, there are schools of thought that completely rejected the ISLM approach. Certainly the post-Keynesians, like Minsky, rejected such an approach. This school seems to have focused almost entirely on monetary operations and how money-flows enter the economy. This seems to me a fantastically interesting and down-to-earth approach. Do you have anything to say about it?

YV: Not much, for the simple reason that I agree with you. The IS-LM model symbolises the bastardisation of Keynes. Whereas Keynes’ greatest insight was about the special nature of labour and money markets, which makes these markets recalcitrant and unable to auto-correct via drops in price (when there is an excess supply of labour or savings; i.e. a Crisis), Samuelsonian ‘Keynesianism’ tried to squeeze the great man’s insights into a neat piece of geometry. But of course, if it could be so squeezed, then the ‘problem’ with these two markets could also be dealt with by means of some technical solution (since the geometry implied that restoring equilibrium and full employment was only a matter of adjusting certain parameters). But Keynes’ significant point was that things were not that simple. That ‘animal spirits’ are fickle and indeterminate, and no amount of ‘tweaking’ of interest rates and government expenditure could reliably restore full employment.

The small band of Keynesians who refused to bastardise Keynes, authors like Leijonhuvfud and Minsky, continued to say sensible things at the time when IS-LM Samuelsonianism was bandied about as… Keynesianism. In this sense, economists of the Samuelson ilk set up a straw man version of Keynesianism which, unsurprisingly, was blown over by the first gust of the 1970s stagflationary winds. Alas, at that point the Samuelsonian ‘Keynesians’, instead of admitting that Leijonhuvfud and Minsky had got it right, chose to defect to neoclassical modeling instead. They became the equivalent of Reagan-Democrats: Neoliberal-Samuelsonians.

So, to wrap up, it seems to me that once one’s mindset identifies economics with modeling, the urge to keep modeling overcomes any philosophical, ideological or analytical commitment to an economics that retains any link to really existing capitalism.

Topics: Economic fundamentals, Free markets and their discontents, Guest Post, The dismal science