Skip to content

Interview on innovation with Peter Swann

↓ Jump to responses

Download the newsletter ›

Peter Swann was formerly Professor (and now Emeritus Professor) of Industrial Economics at Nottingham University Business School. Most of his research has been about the economics of innovation. Here he answers some questions from Stuart Birks

Swann1. Why do you think innovation is an important topic for economic analysis?

It’s a good question. If you had asked me this when studying the economics of innovation for my PhD, I would probably have said that I don’t really know if it is an important topic or not, but I am very interested in it. In retrospect, it was just a happy accident that my research came to focus on the economics of innovation. I didn’t appreciate at the time how important this field would be.

A few years later (say, mid- to late-1980s) however, I would have given a completely different answer. By then, it was clear to me that this really was an exceptionally important topic which had been surprisingly neglected by economists in the twentieth century. Solow’s important study from the 1950s had established that a large proportion of economic growth in the US could be attributed to technical change rather than to growth in the use of capital and labour (Solow, 1957). Economic histories by David Landes (2003), and others, had shown just how important innovation was in the process of industrialization. And Chris Freeman’s pioneering empirical work had shown us just how important innovation was in some of the industries with the highest growth rates (Freeman, 1982).

By then, also, I had started to look back to some of the classic works in economics (Adam Smith, Karl Marx, J.S. Mill and others) and found that innovation was prominent in all of these. And, of course, Joseph Schumpeter (1942) had put innovation at the centre of his analysis of capitalism. So at that stage, the question was not so much, “why do I think innovation is an important topic for economic analysis?”, but rather, “why did so many economists in the twentieth century think that innovation was not an important topic for economic analysis?” Of course by the late 1980s, the economics of innovation had taken off in earnest as an exciting new field within industrial economics.

Indeed, I remember attending a seminar towards the end of my PhD studies when a prominent theorist from one of the world’s greatest universities gave a theoretical paper on the economics of innovation. He prefaced his talk with the observation that if you were to look at the leading economics literature of that time, you could be forgiven for thinking that economics and innovation had almost nothing to do with each other. I remember this remark clearly because it really was a rather shocking observation, but it was indeed an accurate statement about the economics literature of that time.

Now, however, my answer would be quite different again. Over the last 15-20 years or so, I became preoccupied with a more cautious assessment of innovation. Have we overestimated the benign effects of innovation? Are the benefits of innovation starting to show diminishing returns? Are the returns to the innovator much greater than the value to society as a whole? And, can innovation do damage in some cases?

Joseph Schumpeter (1942) had already given us the answer. Any student of the economics of innovation will have come across his famous description of innovation as, “a perennial gale of creative destruction”. Yes, innovation is destructive as well as creative, though with luck the creative effects are greater than the destructive. But as I started to examine various case studies of the destructive side of innovation, I saw that these can be substantial (Swann, 2014). Schumpeter would not have been surprised in the least, but when I have shared the message of these case studies with many people working on the economics of innovation, they react with bewilderment and even a little frustration. Their response seems to imply that I am being irresponsible to cast doubt on something that is so important for economic growth.

I don’t think that is irresponsible at all. Indeed, I think it is essential that economists don’t get carried away with their own rhetoric. So my answer today to your question would be this: it is important to analyse and understand the economics of innovation so that we can form a clear and realistic account of the good (and bad) things that it can do.

2. How would you view the mainstream economics approach to innovation?

I really need to break this question into two parts. (2.1) How would you describe the mainstream approach to the economics of innovation? And (2.2) what do you think of the mainstream approach to the economics of innovation? Let us start with (2.1).

Twenty years ago, I think it would have been fair to say that much of mainstream economics treated innovation as something that just enhances productivity. The main research question was therefore: to what extent does innovation enhance productivity? As for research methods, mainstream applied research on the economics of innovation was dominated by econometric studies – as was most mainstream applied research in economics.

We should remember, however, that the mainstream economics of innovation (like any other discipline) is always evolving. We see signs that the mainstream will, after some time, adopt ideas from heterodox economics and even from other social sciences. This diffusion of ideas is helped by several ‘ambassadors’ who belong to one community but also associate with the other.

Today, I would describe the mainstream approach in a different way. The theoretical perspective is broader than it was twenty years ago. Nevertheless, mainstream economists still have a more constrained and limited theoretical perspective on what innovation does to the economy than you will find amongst heterodox economists. Related to this, mainstream economists still seem to consider a narrower range of research questions about innovation than heterodox economists. Thanks to the gradual diffusion of ideas from heterodoxy into the mainstream, the range of questions considered by the mainstream does grow over time. But so too does the range of questions considered by heterodox economists. I think it is fair to say that in this respect, at least, the mainstream is often quite a few years behind the heterodox community.

Some readers may ask: can you actually give some evidence for your assertions about the comparative range of research questions in the mainstream and in the heterodoxy? Yes I can. The Community Innovation Survey (CIS) is a survey about innovation carried out in all the member states of the EU. Traditionally, the UK CIS has included a large number of questions, and for this reason, response rates were probably lower than they could have been for a shorter survey. Some of the government economists and statisticians who designed the survey asked researchers for their views about the range of questions to be covered. The responses are instructive. Mainstream econometricians typically wanted CIS to be more ‘econometrics-friendly’, with just a few questions on the key issues of concern to the mainstream, a higher response rate and a richer set of panel data. Heterodox economists, in comparison, were mostly happy with the large number of questions because this made it possible for them to consider a wide variety of questions.

Now I can turn to question (2.2): what do I think of the mainstream approach to innovation. Let me start with theoretical frameworks and research questions. The short answer is that I have at times found these rather limiting and frustrating. This was especially true when most mainstream economists seemed only to be interested in the effects of innovations on productivity. That is a natural question to ask about process innovations, perhaps, but I was really more interested in product innovation. Up to a point, you can fit product innovations into this framework if you are prepared to stretch what we mean by increased productivity. But I think this limited perspective was unhelpful because it focused only on one of the things that innovation can do, and ignored many other things that innovation can do. While the mainstream has now moved on from an exclusive focus on the productivity effects of innovation, I still find that the lag between heterodox work and the mainstream is frustrating. I find that I want to discuss questions that the mainstream simply does not consider.

What, finally, do I think of the research methods used by the mainstream? This is something I have already discussed at length (Swann, 2006) so I won’t repeat myself. But let me correct one misunderstanding about that book. Some have accused me of being very anti-econometrics, but that is quite wrong. As that book makes clear, I am not anti-econometrics, but I am anti-monopoly. Econometrics holds a massive ‘market share’ in mainstream applied work, a near monopoly, and most other applied techniques have been completely marginalised. That is what I object to! I am convinced that econometrics has a place amongst the portfolio of research techniques we use, but it should be a much more modest position than it holds at present.

3. How might a cross-disciplinary approach improve understanding?

I have two things to say about this. The first is an obvious point, perhaps, but is very often overlooked.

While Adam Smith extolled the virtues of the division of labour in the Wealth of Nations, he also talked about its vices (see the discussion on division of labour in Book V Chapter I). The main virtue is the way the division of labour enhances productivity, while the main vice is that people whose work involves endless repetition of similar tasks develop a very limited perspective on the world.

If we are to reap the benefits of the division of labour, somebody – at least – must recombine the fruits of each labourer’s work. In a manufacturing process, that recombination is ensured by the assembly of a physical product for sale. But in the research world, there is no assurance that the recombination will take place. Indeed, I can only think of a few scholars who indulge in this recombination activity. Career advancement primarily depends on excellence within your narrow discipline, and there is very little incentive to work at recombination.

Without that recombination, economists work in ‘splendid’ isolation. Not everyone thinks that is a good idea. Hayek, for example, believed that, “the economist who is only an economist is likely to become a nuisance if not a positive danger.” (Hayek, 1967)

The second way in which a cross-disciplinary approach can improve understanding is really an extrapolation of what we learned above about the gains from encouraging a heterodox research approach. As I said before, the heterodox community has a less restrictive framework in which to think about the economics of innovation and can therefore ask a broader range of research questions. A truly cross-disciplinary approach can take this process even further still.

I would identify four areas in particular where I have found cross-disciplinary work especially valuable to me. These are:

a) The sociology of consumption
b) Engineering economics
c) The geography of innovation
d) The psychology of creativity

When you learn about the sociology of consumption – for example, Bourdieu’s work on distinction (Bourdieu, 1984) – you start to realise that the mainstream economic theory of consumption and demand is a very special case. It is not wrong – because there are cases where people do genuinely behave in the same sort of way as that theory suggests. But you start to realise that there is a great deal more that influences consumption. In the context of the economics of innovation, the sociology of consumption opened my eyes to a whole host of other factors that could influence the likely market success of a product innovation.

By engineering economics I mean the sorts of economic principles governing technological change that emerge from scientific and engineering analysis. A leading example of this is the work around Moore’s Law (concerning the rate of growth of computing power per semiconductor chip). This scientific and engineering analysis allows us to make some very precise predictions about the costs of producing computer chips, and how that falls over time. Compared to what any other method can offer, engineering economics is something of a revelation – though it can only be used in a limited range of contexts.

When I did some work on geographical patters of innovation twenty or more years ago, economic geography was starting to emerge as a hybrid discipline in its own right. It has now developed in to a healthy hybrid discipline. I can’t say whether the dialogue between economists and geographers is yet as well developed as it should be, but the momentum is there.

Finally, I have found it fascinating to compare and contrast the psychology of creativity with the economics of innovation. In some ways, the two complement each other, but in other ways they are clearly in tension. That makes for a truly fascinating dynamic which has not, as I see it, been properly explored yet.

I’m sure that readers could cite many other examples.

4. How would you suggest that we frame the issue of innovation? And what is your approach to the analysis of innovation?

I think that the ideal ‘frame’ for the topic of innovation should be as loose-fitting as possible. A tight framework may contort our data to conform with a bad theory, because it leaves no opportunity for the data to ‘rebel’. So here I am at one with Sherlock Holmes – the great (fictional) detective: “It is a capital mistake to theorize before you have all the evidence. It biases the judgment.” (Doyle, 1887)

Turning to your second question, I would describe my approach to the economics of innovation in three key steps.

a) First, we need to move on beyond Friedman’s test – as set out in his Methodology of Positive Economics.

Friedman (1953) argued that the test of any theory or model should be the accuracy of its predictions and not the realism of its assumptions. Like so many of the principles that have guided mainstream economics, this test is not wrong as such, but it is based on a strong assumption. It assumes, implicitly if not explicitly, that our interest in the model is limited to the specific data used to test it and the specific context that gave rise to those data. That may be true for some econometricians, but it is not true for all economists – especially those who want to use the model for policy purposes.

When you use models for policy purposes, it is common to ask: what happens if we make radical changes to some of the variables in a way that has not happened before. A model based on false assumptions may predict a particular set of data well. It may even perform well outside that sample with another set of data from a similar context. But it cannot work well over an unrestricted domain. Sooner or later, a model based on false assumptions will make poor predictions.

My concern about Friedman’s test, as it stands, is that it is used by econometricians to neutralise criticism from heterodox economists who use other research methods that contradict econometric results. As I have argued at length elsewhere (Swann, 2006), this means that economists make too many Type II errors – that is, they fail to reject false theories – and these errors can be very costly. I would prefer to replace the Friedman test with a tougher test. If a model predicts well and is based on realistic assumptions, then it is a good model. Otherwise, it is a bad model.

b) Second, we need to loosen the grip of mainstream theory and mainstream assumptions on our view of innovation.

I sometimes think that trying to understand all aspects of the economics of innovation within the constraints of the neoclassical framework is a bit like trying to play tennis in a straight-jacket. I think this perfectly sums up the frustrations I have felt in trying to analyse innovation within an over-restrictive framework.

As I look back over some of the key advances in the economics of innovation, it seems to me that each of these loosens one of the assumptions of the existing mainstream approach. So, for example, an important step was the work of several authors to loosen the framework so that it could properly incorporate product innovations. Nelson and Winter (1982) released the economics of innovation from the restrictive static framework of the mainstream and launched the field of evolutionary economics. Von Hippel (1988) developed the idea that user knowledge was one of the most important sources of information for innovation, and then later (Hippel, 2005) considered the idea that users could actually play an active role in driving the direction of innovation. And so on.

My recent work (Swann, 2014) has taken a few more steps to loosen the grip of mainstream thinking. I questioned the idea that innovation was generally socially benign and studied in detail the destructive side of innovation. And I dropped the assumption that innovation is primarily a business activity and started to look at ‘common innovation’ – innovation by ordinary people, families, clubs and local communities to enhance their well-being.

c) Third, we need to use a wide variety of applied research techniques.

In earlier work (Swann, 2006) I described at length the case for using a wide variety of applied research techniques. I can summarise the arguments very simply as follows. All these different techniques have their strengths and all have their weaknesses. The weaknesses of one technique can often be addressed by using another technique. A research study that uses many techniques should therefore be able to address many weaknesses. There is no universal solvent or universal technique that is the best to answer all economic questions.

 

REFERENCES

Bourdieu, P. (1984) Distinction: A Social Critique of the Judgment of Taste, Routledge Kegan Paul

Doyle, A. Conan (1887) A Study in Scarlet, http://www.gutenberg.org/

Freeman, C. (1982) The Economics of Industrial Innovation, 2nd edition, Frances Pinter

Friedman, M. (1953) ‘The Methodology of Positive Economics’, in Essays in Positive Economics, University of Chicago Press

Hayek, F.A. (1967) ‘The Dilemma of Specialization’, Chapter 8, Studies in Philosophy, Politics and Economics, Routledge Kegan Paul

Hippel, E.A. von (1988) The Sources of Innovation, Oxford University Press

Hippel, E.A. von (2005) Democratizing Innovation, MIT Press

Landes, D. (2003) The Unbound Prometheus: Technological Change and Industrial Development in Western Europe from 1750 to the Present, 2nd edition, Cambridge University Press

Nelson, R.R. and S.G. Winter (1982) An Evolutionary Theory of Economic Change, The Belknap Press of Harvard University

Schumpeter, J.A. (1942) Capitalism, Socialism and Democracy, Harper

Solow, R.M. (1957) ‘Technical Change and the Aggregate Production Function’, The Review of Economics and Statistics, 39, 312-320

Swann, G.M.P. (2006) Putting Econometrics in its Place, Edward Elgar Publishing

Swann, G.M.P. (2014) Common Innovation: How we create the wealth of nations, Edward Elgar Publishing

 

From: pp.10-13 of World Economics Association Newsletter 5(1), February 2015
http://www.worldeconomicsassociation.org/files/Issue5-1.pdf

Download WEA newsletter Volume 5, Issue No. 1, February 2015 ›

1 response

  • David Dahmen says:

    Swann has given an important and well thought out interview. I would have preferred to have seen a more ample list of references including authors such as Brian Arthur and Joel Mokyr who have brought the systems dynamics perspective of the Santa Fe Institute to the area of economic studies.
    In particular the above authors emphasize how asset pricing does not generally reflect equilibrium theories which has been a major criticism with respect to central bank policy. Innovation has to do also with dynamic disequilibrium when East Asian governments act rapidly for geopolitical motives in ways that have economic consequences leave Western slow moving democracies flat footed.

Respond to this article

You may use these HTML tags and attributes: <a href="" title=""> <abbr title=""> <acronym title=""> <b> <blockquote cite=""> <cite> <code> <del datetime=""> <em> <i> <q cite=""> <strike> <strong>

Please note that your email address will not be published.