On Economic Methodology: An Interview with Sheila Dow

Sheila Dow is Emeritus Professor of Economics at the University of Stirling, Scotland, and Convener of the Scottish Centre for Economic Methodology. Her latest book ‘Foundations for New Economic Thinking’ is available from Amazon.com

Interview conducted by Philip Pilkington

Philip Pilkington: Your book is all about the importance of methodology in economics. Up until now methodology has largely been side-lined. I recall, for example, that Paul Samuelson made a comment to the tune of “Those who cannot do science do methodology”. That strikes me as being broadly the mainstream attitude to any discussions of methodology in economics. In the face of this why do you find methodology to be so important?

Sheila Dow: Methodology is about the way in which we do economics, what we think is a good or bad argument and how we choose between theories. It’s not just about whether to use this technique or that technique. Among other things, the methodology we use assumes a particular way of understanding how the economy works and what kind of knowledge is possible.

I think it is always important for economists to understand the basis of what they do, and to be able to justify their own choice of methodology. But it is particularly important now, when the financial and economic crisis is challenging much of what was in the mainstream of economic theory. Those who dismiss thinking about methodology are assuming that there is one best way of doing economics on which everyone agrees. But there are in fact different ways of understanding the economy and therefore different methodologies.

If any economist wants to be persuasive about the reasons in favour of her own approach to economics she needs some understanding of how others approach the subject. The first step for many economists therefore is to accept that there is more than one way of doing economics; this itself requires some methodological awareness.

PP: In the book you suggest that having an awareness of methodological issues allows us to distinguish between different schools of thought. Perhaps you could reflect on this a bit?

SD: I see each school of thought as being identified by its own methodology. Of course there are grey areas between schools of thought and individual economists who don’t fit neatly into any one school of thought. Nevertheless it seems to me that the notion of a school of thought is useful shorthand for categorising different ways in which economics is done.

Rather than making communication worse, I see such categorisations as improving communications. If someone calls themselves a Post Keynesian, for example, we have a good idea of how she understands the economy and what she regards as good arguments. Disputes between schools of thought really need to focus on this level. Members of different schools of thought arguing about theory won’t get far unless they recognise that they understand the economy differently, use concepts differently, use different styles of argument and so on.

We can also reduce talking at cross purposes if we recognise which bodies of theory are methodologically similar, and thus belong to the same school of thought, and those which are methodologically different. Thus I would categorise all forms of mainstream economics (rational expectations theory, New Keynesian theory etc.) as one school of thought with a shared basic methodological approach, while New Keynesian economics and Post Keynesian economics, for example, are methodologically very different.

PP: In the book you attribute great importance to issues of uncertainty and the dichotomy between rationality and irrationality. Could you talk about this and in particular how it impacts upon questions of methodology?

SD: Yes, this question illustrates why it is important to consider the methodological foundations of economic theory, not least because words like ‘uncertainty’ and ‘rationality’ mean different things in different schools of thought. Mainstream economists generally take uncertainty to mean risk, as in quantifiable risk. So they don’t take account of unquantifiable risk, which is what non-mainstream economists mean by uncertainty.

Because non-mainstream theorists understand the economy as an open, evolving system, they think that uncertainty is really important. Where risk can’t be quantified, expectations can’t be captured by a probabilistic statistical forecast, and so we need other guides to our behaviour, like conventional judgements, expert judgement or assuming that the future will be like the past. Uncertainty then is an indicator of how much confidence we have in any forecasts and in fact how far we are prepared to commit to forecasts at all. According to this view, economists themselves face uncertainty in their understanding of the economy, so their methodology is designed to deal with this.

The mainstream definition of rationality is based on the idea of individuals maximising their utility on the basis of full information (within probability distributions, sometimes with some particular information concealed). Anything else is defined as irrationality. But for non-mainstream economists, where uncertainty is the norm, this type of rationality isn’t feasible. We use reason along with evidence and the kind of conventions noted above in order to make judgements about the future. But sometimes uncertainty is so high that it is difficult to form any judgement. Emotion then plays an important part, for example in encouraging action in spite of uncertainty. This behaviour may be unreasonable if it flies in the face of evidence, but is not automatically classed as irrational.

PP: In the book you discuss what you refer to as a “Babylonian method” for doing economics. In my understanding you associate this method with Keynes and the Post-Keynesians and contrast it with the “dualistic” approach generally followed by the mainstream. Could you outline the differences between these methodologies and explain why you think the Babylonian method superior?

SD: The Babylonian mode of thought is a way of thinking. It is based on an understanding of the social system as being open and evolving, so that any knowledge is inevitably partial (and subject to uncertainty). It picks up on Feynman’s account of Babylonian mathematics, which allowed for different starting points for different chains of reasoning, depending on the problem at hand. It is a pluralist approach in that it allows for a range of methodological approaches, focusing on different aspects of a complex system. It also implies that each methodology should be pluralist, in the sense of employing a range of methods. Babylonian thought is not atomistic, i.e. it doesn’t require that all arguments be built up from a common starting point of isolated individuals. Rather it is organicist, allowing for and addressing complex interactions and evolutions. Nor is it dualist. There can be degrees of uncertainty and interactions between reason and emotion, for example, and variables can be endogenous to one chain of reasoning while exogenous to another chain, depending on the problem at hand.

The mainstream mode of thought is based on a closed-system approach, requiring all arguments to be built up on microfoundations expressed in terms of atomistic individuals. The resulting methodology is monist: there is one best methodology, which is formal deductive mathematical reasoning. It is a dualist mode of thought, since a closed system only allows for certainty or ignorance and rationality or irrationality, for example, and defines variables as always either endogenous or exogenous.

I prefer the methodology implied by the Babylonian approach since it fits the way I understand the economy. In other words I put realism ahead of the internal elegance of the mainstream approach. The Babylonian mode of thought itself describes how I think, but this is not something I consciously choose. Mainstream economists evidently feel comfortable thinking in a different way. I’m not qualified to say why this is so, but I can point out the damaging consequences, e.g. of uncertainty denial.

PP: In the book you also discuss the use and abuse of mathematical formalism in economics. Perhaps you could talk a bit about mathematical formalism and how it is misused in economics?

SD: By mathematical formalism I mean the insistence on theory being expressed in formal mathematics. In mainstream economics this takes the particular form of deductive mathematics, i.e. deriving everything deductively from a common set of principles about rational optimising individuals. Behaviour is shown as being based on rational calculation with respect to information which is normally known with certainty. This allows definitive results to be derived using deductive mathematics. At the same time, when mainstream economists consider any behaviour which does not fit this pattern, mathematical formalism limits how it can be treated, if at all.

All models are closed systems, but this methodological approach means that the economy itself is treated as a closed system. So, as a methodological position, it is coherent only if you can accept that the economy is indeed a closed system. This is not my own belief, but I suspect that many mainstream economists would also find it hard to support.

But it is not a simple matter of everyone choosing one methodology or the other, because the mathematical formalist approach includes the view that it is settled that this is the best way of doing economics. This view lies behind a lot of the mainstream refusal to discuss methodologies other than mathematical formalism, or indeed methodology in general. So, not only is mathematical formalism contestable in terms of limiting how far economics can apply to an open-system economy, but it is also contestable in claiming to rule out other methodologies.

This is not at all an argument against the use of mathematics – it’s an argument that economics should not be exclusively mathematical. An open-system approach (based on the Babylonian mode of thought) can include mathematical argument as a partial contribution to an overall theory along with other types of argument. Mathematical formalism can be appealing because it involves a strict application of rules and yields definite conclusions. Alternative methodologies may be less appealing because the conclusions are less definitive and can vary from context to context. Nevertheless they are, in my view, more rigorous in terms of taking account of the nature of the subject matter.

PP: It seems to me that the use of mathematical formalism limits an economist’s ability to take into account institutions and institutional structures. Perhaps you can say something about this – especially given its importance after 2008.

SD: Yes, I agree. There is some scope for analysing institutions in a formal mathematical way as a partial analysis of a particular context. It is the deductivist structure of mainstream mathematical formalism and its exclusivity which are the problem. Mainstream theory can only include institutions as constraints or as the outcome of rational optimising behaviour by atomistic agents.

The financial crisis has exposed these limitations. Mainstream money-macro theory did not include banks, for example. Because mainstream methodology doesn’t provide room for uncertainty – or the role of institutions in helping us cope with uncertainty – it was difficult to address the crisis. It was particularly difficult to address the breakdown of markets, as when the interbank market froze. Further, a calculative account of choice ignores the importance of confidence and trust for the financial system (including central banks) and provides little guidance as to how to restore them when they are lost.

A non-mathematical non-mainstream analysis of the evolution of banking systems and of behaviour under uncertainty in contrast provided an explanation for the crisis and guidance as to how to address it.

PP: You mention in the book that postmodernism has had some influence on economics. Could you briefly describe how much influence it has had and on what areas?

SD: Postmodernism had its most obvious impact on economics through its critique of modernism, an approach which had had important influences on mainstream methodology. In particular postmodernism challenged the idea that it was possible to establish general theories, tested against independent facts. Unfortunately the critique was understood by many people in dualistic terms to imply that no knowledge was possible. Much of my work has been designed to explore the middle ground, of knowledge which is uncertain but still provides a basis for action.

But postmodernism had a less obvious but more pervasive influence within mainstream economics. At one level it encouraged the withdrawal from policy activism. At another level it encouraged a subjectivist approach to knowledge which allowed mainstream economists to continue to avoid addressing uncertainty. It was assumed that individuals subjectively estimate probabilities, even if they cannot do so objectively because of the open nature of the real world. Of course this still does not address uncertainty (which refers to varying degrees of confidence in estimates or a complete unwillingness to commit to making an estimate).

PP: Finally, let’s talk about alternatives. In the book you favour a “structured pluralistic” methodology. Could you explain what this means and why you think it is the type of methodology economics should pursue?

SD: The term ‘structured pluralism’ is intended to signal a difference from a pure form of pluralism; this pure form is the ‘anything goes’ position. At the level of methodological pluralism, structured pluralism means that we can roughly categorise a finite range of approaches (or schools of thought). These schools are communities, with shared understandings of the real world and of how best to build knowledge about it. As a socio-epistemic activity, therefore, economics has to be structured within groupings, and there is a logistical limit to how many schools can function effectively within the discipline. This limited range of schools helps us be methodological pluralists in the sense of allowing us to recognise and learn about other schools of thought than our own.

At the level of methodology itself, each school of thought uses a range of methods suited to its view of the world and its view of knowledge, so in this sense a pluralist methodology is also structured.