This article first appeared in the Newsletter of the Royal Economic Society in 2013. It is republished here with thanks.
Imperial College London
Since the crisis of 2008, there has been much discussion of the future of economics, and especially the future of its teaching. This has generated many interesting contributions from many perspectives.[i]
Several themes have emerged that apparently command wide support in many quarters, including prospective employers of economics graduates. Among these is a group of inter-related topics that could be seen as addressing the need to pay more attention to reality:
- better connection with the real world, including familiarity with economic history;
- ability to handle data, and knowledge of data sources;
- ability to synthesise evidence, and more broadly an emphasis on the application of economics.
Drawing these themes together, they can be expressed as the importance of a solid grounding in evidence. This includes some understanding of the way that it is obtained, as well as possession of substantive knowledge. It also encompasses evidence both descriptively on the main features of the economy and how they change over time, and mechanistically on what brings about those features and changes, i.e. how the economic system works.
Underpinning this approach is the fundamental principle that everything that is taught should be defensible in factual terms. Not that it is necessarily “true”, as that would greatly limit the scope of the discussion, but rather that there is evidence for and (probably also) against. The corollary is, things that are known to be untrue should not be presented as fact. It may seem odd to emphasise this, but it is motivated by the observation that some elements in the standard curriculum, and in textbooks, are incompatible with the evidence. Not merely oversimplified; false. I return to this issue below.
Plurality of types of evidence
In the past twenty years, economics has seen a huge growth in empirical research. Much of this is on practical issues, such as the influence of class size on educational attainment, and of educational investment on future earnings. These are important policy questions. In teaching, however, one needs to prioritise. The types of evidence that are most relevant are (a) those relating to the broad features of the economy, both descriptively and mechanistically; and (b) evidential support for (or against) the major economic theories, e.g. each of the components of the General Theory.
Evidence can be of many kinds, and different types of course have different requirements. Business schools need an emphasis distinct from that of economics courses (and may currently be better in this respect). For economics, data handling has featured prominently in the discussions; in addition, it is important for students to have an appreciation of where data come from, and how this may affect their quality – the strengths and weaknesses of key datasets. In substantive analyses attention should be directed more at what is driving the data, causally, than on quantification of estimates.
There are important sources of evidence other than statistical data, for example surveys, qualitative as well as quantitative.[ii] Evidence need not consist only of generalisations; specific events and case studies can also be instructive.[iii] So can descriptive studies, including on how particular sub-systems of the economy work.[iv] Secure knowledge comes from bringing different approaches together, and if necessary addressing any inconsistencies.
Behavioural economics is now part of the mainstream. This signifies a shift in emphasis towards how things actually happen, away from a focus on how they might happen in an ideal world. If one of the frequent criticisms of the curriculum is that it needs to move closer to reality, this is a good opportunity for teaching (and modelling) to be founded on real behaviour patterns, and on acknowledging that causation in economics involves multiple causation not determinism. Behavioural postulates based on imaginary axioms may be able to generate precise predictions, but this is of little value if these are wrong – accuracy should take precedence over precision.
Economic history is substantively important. It is the indispensable record of how economies actually behave, the particular structures they have, and how they change. In addition, economic historians pay a great deal of attention to the above-mentioned issues such as data quality, and the discipline encompasses specific events and narrative as well as quantitative methods and consideration of general historical processes. Importantly, the comparative method is very strong, and provides insight – although not necessarily decisive conclusions – into the process of long-term change.[v] It is very important not to be centred on the experience of only one country, e.g. the UK or the US (or even both), but to have a broader global perspective. This gives a sense not only of what happened in a particular economy at a particular time, but also of what did not happen.
If it’s untrue, don’t teach it
The reader could be forgiven for responding that all this is straightforward commonsense. However, there is a problem. The RES Steering Group propose many elements that should be added to the economics curriculum, but are silent on what should be removed. Moreover, its recommendations begin with asserting the strength of mainstream economics, albeit hedged with stressing the need for humility and for honesty about its limitations. The unfortunate truth is that “mainstream” economics still contains elements that are contradicted by the evidence, but which are presented as if they were true. I will focus on just one example, the shape of the firm’s average cost curve.[vi]
It has been known for over 60 years that in manufacturing at least, the U-shaped average cost curve is rare – estimates range from 5 to 11 percent of firms. Nevertheless, it is ubiquitous in textbooks, and presented as a general truth. Unfortunately, the historical discussion of this issue became diverted into whether or not this finding “invalidated marginalist theory”, which is odd given that the theory is tautologously true. We ought to stop teaching the U shape as the typical relationship between costs and scale, for the simple reason that it is false.[vii]
Falsity is not the same as simplification. For specific modelling purposes, a particular assumption may be useful, even though it is known to be a gross over-simplification. The classic case is rationality: it is clear that economic behaviour does not always accord with the tenets of classical rationality, but there may well be situations in which the rationality assumption is both useful and harmless. In such cases, it would be explicit that its inclusion is justified pragmatically, and should not be taken to imply that it is a realistic description of the actually-existing psychological process.
In such a situation, the importance of the distinction between the model and the reality it sets out to represent becomes crucial. We need to encourage an attitude in our future economists that they are as aware of the many features and forces of real situations as they are of the simplified, and hopefully elegant, model. Thus, if a macro model omits the financial sector, implicitly assuming that it plays only a passive role, then this omission would be visible. Students need to have a broad and multi-faceted understanding of the real situation, so that they can see which elements have been selected for the model, and which have been omitted. Again it is a question of being able to see what is not there.
A good model for how to do this is biology. Whereas in economics we tend to rely on rather sparse “stylised facts”, biology has progressed by building up from multiple descriptions (including of what is obvious to the observer), moving on to empirical generalisations, and then to attempts at explaining established patterns. This process leads to empirically-based theories that can subsequently be subjected to testing/refinement/rejection.
Currently, prestige is attached to the more abstract parts of the division of labour within the economics profession – and the model of general equilibrium is indeed a beautiful creation. For the curriculum, however, a better balance is needed. In particular, the mutual dependence of theoretical categories and evidence should be emphasised. The discourse should focus less on competing theories and models in the abstract, and more on the relation between each of these theories/models and the evidence. Students should be judged as much on knowledge of the real world as on knowledge of models.
[i] I will not attempt a list, but restrict myself to mentioning the recent contributions of Diane Coyle and David Simpson in this publication – for the former, see http://www.res.org.uk/view/article7Apr13Features.html
[ii] An example of a survey that is useful for quantitative analysis is the Michigan Survey of Consumers [http://www.sca.isr.umich.edu/]. A more qualitative example is Blinder AS, Canetti E, Lebow D and Rudd J (1998) Asking About Prices: a New Approach to Understanding Price Stickiness, New York: Russell Sage Foundation. The methodology can be criticised, but in the current state of knowledge it is still able to generate important information.
[iii] For example, a highly instructive (if idiosyncratic) collection of single-country studies can be found in Rodrik D (ed.) (2003) In search of prosperity, Princeton: Princeton University Press.
[iv] See for example, Ryan-Collins J, Greenham T, Werner R and Jackson A (2012) Where does money come from? London: the new economics foundation.
[v] The reason for the wealth of some countries and the poverty of others has been a particularly rich topic. Important contributions include Landes D (1998) The wealth and poverty of nations, WW Norton & Company Inc; Pomeranz K (2000) The great divergence, Princeton: Princeton University Press. The diversity of views in this literature shows that some controversies are not easily settled. A different topic, the growth of the state in the past 200 years, has been forensically analysed in a comparative perspective by Lindert P (2004) Growing public, Cambridge University Press.
[vi] I am abstracting from distinctions such as that between short- and long-term, and that between the firm’s perceptions and reality, in order to emphasise the main point. Note that this is part of micro – indicating the inadequacy of the widespread view that the only problems are in macro and finance.
[vii] Eiteman, WJ and Guthrie, GE (1952) ‘The shape of the average cost curve’, American Economic Revue 42, 832-38; see also Blinder et al., op cit.