Exploring the epistemological growth of macroeconomic thought.

SECTION I. Introduction¹

This paper explores the nature and construction of macroeconomic thought in order to understand the growth of knowledge within macroeconomics. Employing theories of epistemological growth as the basis for assessing macroeconomic thought, this paper seeks to answer two key questions:

1 What is the nature of epistemological growth within macroeconomic thought?

2 Does the nature of macroeconomic thought allow us to know more about the economy today than we did before?

Sections II and III will deal with the first question, and I will respond to the second question in Sections IV and V by offering a modified Kuhnian perspective to the analysis of the epistemology of economics.

My initial hypothesis is that epistemological growth within macroeconomic thought can be described as evolutionary. In the case of macroeconomics – a dynamic and interactive field – knowledge evolves not only through selection and survival of theories, but also in accordance with events and actions induced by economic agents. As Karl Popper had concluded, resultant theories can only be provisionally true, regardless of the extent of successful empirical testing.

However, from a Kuhnian perspective, the evolutionary nature of macroeconomic knowledge seems to imply that we cannot know more about the economy today than we could before. The result is that macroeconomic knowledge does not seem to grow in terms of giving us an increasingly coherent body of knowledge that we may apply to current economic phenomena.² Instead, what we may apply to current phenomena is merely the part of knowledge that is coherent with the current economic paradigm. Theories from preceding paradigms may be applied only to fail in the face of incommensurability across paradigms. My aim is to show that a modified Kuhnian framework (which relaxes some of the assumptions in Kuhn’s model) may be a possible explanation for the epistemic evolution of macroeconomic knowledge.

SECTION II. Ways of Considering Evolution in Economics

Herein we will take economics to be a social science, although it is possible that other frameworks are employed in analyzing economic knowledge. Social science may not qualify as a science per se; often the distinction lies in the nature of the phenomena studied. Social sciences tend to deal with human phenomena; the dynamism of interactions between the observer and the observed adds a dimension of complexity that draws the discipline away from being a ‘proper science’. Nevertheless, epistemological frameworks established for the study of the natural sciences can be relevant in our present analysis. The study of economics is often categorized under the positivist school; it is likened to science not by practice, but by its objective – to generalize a wide range of phenomena into suitably simplified models. (Perlman 1986)

In the next section I will examine the field of economic knowledge through the two key frameworks found in scientific epistemology – Popperian falsificationism and Kuhnian paradigmatism. Popperian falsificationism is the notion that scientific knowledge is obtained via the process of falsification. Theories may never be proven necessarily true by the principle of induction that scientific methodology appeals to, but can be proven necessarily false.³ By rejecting falsified theories, Popper argued that we might arrive at a closer understanding of truth(s). Thus, Popper appealed to falsification as a means of contributing ‘negatively’ to knowledge. In doing so, theories can only ever be contingently true or necessarily false, i.e. no theories are necessarily true with falsificationism. This implies that we can never expect any theory to be applicable without abandon. (Popper 1959)

Kuhnian paradigmatism is the notion that scientific knowledge consists of several ‘paradigms’. A paradigm defines a set of scientific ideas that are logically coherent in explaining currently observed scientific phenomena, while remaining incommensurable with other paradigms. Scientific history is defined by periods of normal science that are punctuated by revolutions in which, following the discoveries of anomalies, a preceding paradigm is overthrown and a new paradigm is born. Science progresses within its paradigm as theories are used consistently to explain observed phenomena, but one is rendered incapable of drawing comparisons of truth and progress across paradigms due to incommensurability. (Kuhn 1962)

Both sets of ideas suggest that some form of epistemological evolution is taking place in the field of knowledge studied, albeit in very different ways. Popperian ideas imply that knowledge evolves progressively, i.e. the future resembles the past in some ways, and advances by altering only the parts that have been falsified or shown to be empirically deficient.⁴ Kuhnian ideas, on the other hand, seem to imply that knowledge evolves somewhat randomly, i.e. entire sets of ideas are overthrown with the change of paradigms, and even when the future resembles the past, the entire ideological set has been altered such that any resemblance is only superficial.

SECTION III. The Nature of Macroeconomic Knowledge

Having established the groundwork of our analysis, it is necessary to relate it more explicitly to the field of macroeconomic knowledge, in particular to the nature of macroeconomic knowledge. Understanding the nature of macroeconomic knowledge will help to explain why the discipline may be more amenable to a particular epistemological framework. Hence, I will attempt to extract the nature of macroeconomic knowledge from a brief historical overview before applying it to the frameworks set out above.

It seems to me that we can conceive of macroeconomic knowledge (herein taken from the Classical to New-Keynesian traditions) as being characterized by three broad chronological categories. (Backhouse 1996) The Classical tradition and some of its precedents may be considered to be the first category. They were then followed by the Keynesians, along with business cycle theorists as a second category. The recent rise of the New Classicalists and New Keynesians alongside certain postmodern schools may be considered as a third category. Macroeconomic knowledge under each category has been characterized by different traits. One way of comparing the three is to consider the implication of economic knowledge in each: advocacy of the invisible hand in the first category, prescription for government intervention in the second, and a move towards psychological and behavioral insights in the third. We may then apply the epistemological frameworks of Popperian falsificationism and Kuhnian paradigmatism against this trichotomy.

Through the lens of a Popperian, macroeconomic knowledge seems to have evolved by means of falsifications. For example, the concept of aggregate supply as being fixed in both the short and long runs according to Classical analysis was ‘falsified’ by the empirical observation of price and wage stickiness in the real world economy.⁵ This then led to the postulation of the extreme Keynesian approach of a completely price or wage dependent aggregate supply (illustrated by the horizontal aggregate supply function). In the third period, however, the Keynesian and Classical approaches were fused under the Neo-Classical synthesis (which accounted for some degree of imperfection in goods and labor markets) to arrive at an upward-sloping aggregate supply function, whose result also cohered with similar theories, such as that of the Lucas supply function. We may also consider other instances of falsifications in macroeconomic knowledge, such as the rejection of the Phillips curve or the augmentation of the theory of money from Classical to Keynesian periods. However, it is not entirely true that knowledge in the examples shown has been falsified. While they have been proven false, such theories are still being employed in explanations of new economic phenomena. Thus, we may wish to seek an alternative explanation using the Kuhnian framework.

From the perspective of a Kuhnian, we may think of macroeconomic knowledge as having evolved through a series of changing economic paradigms. Paradigms are coherent idea sets marked by interparadigm incommensurability, and arise out of revolutions. The trichotomy defined can thus be considered three paradigms of macroeconomics. In the first instance we have a paradigm that rests on the assumption of perfect markets, and in the second we have one that rests on the assumption of imperfect markets. In these two instances it is evident that there is incommensurability across paradigms – the insights one achieves under perfect markets certainly cannot be spoken of in the same manner as those achieved under imperfect markets. For example, the concepts of a Natural Rate of Unemployment (NRU) and a Non-Accelerating Inflationary Rate of Unemployment (NAIRU) clearly illustrate incommensurability.⁶ Another example of incommensurability can be found in the notion of money neutrality (Classical) and money non-neutrality (Keynesian). However, it would be too much of a hasty generalization to assert that there are no grounds for inter-paradigm communication – for it is a known fact that differing paradigms contain similarities (e.g. price and wage are concepts that are used in the same way throughout paradigms, real and nominal distinctions notwithstanding). Thus it seems that some degree of redesign is required in order to explain macroeconomic epistemology through frameworks used in scientific epistemology.

SECTION IV. A Modified Kuhnian Approach to Economic Epistemology

As we have seen earlier, neither the Popperian nor Kuhnian frameworks seem individually sufficient for an account of macroeconomic epistemology. Kuhnian ideas seem to find greater success than Popperian ideas but are nevertheless only capable of offering an impoverished account. In order to give a better account of the evolution of macroeconomic knowledge, I propose a modified Kuhnian framework for the purposes of our analysis.

The modified Kuhnian framework changes the definition of a ‘paradigm’ such that a paradigm is a logically coherent set of ideas that contains several subsets, each of which can be partially commensurable with preceding and/or succeeding paradigms. This implies that macroeconomic knowledge is still defined by periods of normal and abnormal ‘science’, punctuated by revolutions, which signify paradigmatic shifts. Paradigms are, however, only partially incommensurable with each other, and may not always be incommensurable on the same grounds. As long as any degree of incommensurability exists between two paradigms of thought, they can be considered distinct paradigms. Since paradigms are not entirely incommensurable but only partially so, some degree of communication across paradigms can be afforded. It is this allowance for communication that drives the revolution of macroeconomic knowledge – it acts as a door from the old paradigm to the new.

Yet this does not mean that there is any degree of linear truth progression within macroeconomic knowledge as a whole, for the partial incommensurability means that one cannot consistently hold the views of two paradigms in their entirety at the same time – only discrete parts of each. That is, (for example) I may not consistently agree that money is both neutral and non-neutral, however it may be possible for me to hold that aggregate supply is upward-sloping in the short run, but vertical in the long run. By allowing for partial incommensurability, we would be able to avoid the problems that arise when we attempt to apply complete incommensurability across paradigms to macroeconomic knowledge, which is a key failure for direct Kuhnian paradigmatism in the context of macroeconomic epistemology.

SECTION V. Inferences and Implications

In sum, I have argued that developments in macroeconomic knowledge may be better understood in terms of Kuhnian paradigms, albeit with some caveats. What this implies is that there is in fact no ‘growth’ in the field of macroeconomic knowledge insofar as it refers to some progression towards truth and certainty across paradigms and the ability to prefer a newer paradigm to an older one on the basis of it being ‘more true’.⁷ I do not dispute that the body of knowledge grows trivially (in terms of an accumulation of possible theories), but have argued that the active and applicable parts of current debate are not directly comparable to those that have been debunked or proven to be lacking in empirical content.

Yet, there is no reason why this conclusion ought to dissuade us from continuing our pursuit of new theories in macroeconomics. Although we have seen hitherto how the truth of theories does not seem to extend beyond their constitutive paradigms, the objective of macroeconomic knowledge differs from that of, say, scientific knowledge. The way it does so, however, does not jeopardize the value of pursuing new theories in the field. The primary objective of macroeconomic knowledge (or indeed, economic knowledge as a whole) appears to be the modeling of economic phenomena so as to prescribe possible policy approaches towards improving or stabilizing the economy. New theories will aspire to explain changing economic conditions that old theories may no longer be able to account for. As such, we may think of the evolution of paradigms as representing ‘growth’ in a different sense. As paradigms change, economic knowledge ‘grows’ by reinventing itself to retain practical relevance and value – factors that any policy maker will concede to be more important than a commitment to absolute truths in macroeconomic knowledge.

Yuh Yiing Loh is a second-year Philosophy, Politics and Economics undergraduate at the University of Oxford.

Endnotes

1. My sincerest gratitude goes to Jarrod Lee and Aaron Chee for their time and effort in providing invaluable feedback and ideas on drafts of this paper.
2. Herein I refer to ‘growth’ not in the classical sense (e.g. material growth in knowledge propositions regardless of relevance and application) but in an epistemological sense (e.g. growth by adding knowledge propositions to a superstructural epistemic foundation, or addition of new propositions to an existing web of coherent knowledge propositions.).
3. The Problem of Induction, as introduced by Hume (1748):
   1. Premise I: There are no observations of the causal relation between two events, i.e. there can only be an observation of their (repeated) conjunct appearance.
   2. Premise II: Induction, by making a finite number of observation statements, is not a valid inference in deductive logic.
   3. Conclusion: Scientific theories cannot be verified by observations because causal relations can only be inductively inferred.
4. In other words, it relies on the assumption that nature is uniform.
5. Note that the Classical analysis was not entirely falsified (according to the Popperian definition) following the empirical observation of price and wage stickiness, thus the use of apostrophes. The Classical analysis was acknowledged to be deficient by its assumption of price and wage flexibility, but nevertheless continued to be a useful model for approximating market behaviour in ideal cases. Whereas if we stick to a strickt Popperian definition of falsification, the continued application of the Classical analysis would have been unacceptable – a unique feature of economic knowledge that will be explored further in Section IV.
6. 1. Natural Rate of Unemployment: The Classical economists argue that the market clearing condition applies to the labour market, and assert that in the long run, the economy will tend towards the natural rate of unemployment (NRU).
   2. Non-Accelerating Inflation Rate of Unemployment: The Keynesians, on the other hand, argue that wages are sticky downwards and hence the market clearing condition does not hold; in the long run the economy tends to a non-accelerating inflation rate of unemployment (NAIRU – the level of unemployment required to stop rising inflation) as a result of a bargaining equilibirum achieved by workers and firms.
   3. These two measures of unemployment are incommensurable on the grounds that NRU relies on the market clearing assumption whilst the NAIRU measure denies the very same assumption. Since we cannot possibly affirm and deny the same assumption simultaneously, incommensurability arises between these two theories.
7. Truth can be considered in two contexts: on one hand, truth can be objective, i.e. there is one truth and new paradigms become ‘more true’ by progressing towards this objective reality; on the other hand, truth can be subjective, i.e. it explains current economic phenomena accurately and thus one theory is ‘more true’ than another by explaining current economic phenomena accurately (though it might not be as accurate in explaining past or future phenomena). In this particular instance, I am referring to the former version of truth in economic knowledge.

Bibliography

Backhouse, R. E. 1996. Interpreting Macroeconomics: Explorations in the History of Macroeconomic Thought. New York: Routledge.

Dow, S. C. 1996. The Methodology of Macroeconomic Thought: A Conceptual Analysis of Schools of Thought in Economics. Cheltenham: Edward Elgar.

Kuhn, T. 1962. The Structure of Scientific Revolutions. Chicago: University of Chicago Press.

Perlman, Mark. 1986. Perceptions of Our Discipline: Three Magisterial Treatments of the Evolution of Economic Thought. Journal of the History of Economic Thought, Volume 7, Issue 02. 9–28

Popper, Karl. 1959. The Logic of Scientific Discovery. London: Hutchinson.