This article first appeared in the Newsletter of the Royal Economic Society in 2018. It is republished here with thanks.
Imperial College London
Prompted by Ben Chu’s observations about the state of economics in our July issue, Michael Joffe, Imperial College London, explains what he thinks is really wrong with economics.
Ben Chu’s report of the RES Conference in the last Newsletter1 (no. 182, July 2018) begins and ends by referring to recent accusations against the profession. He contrasts them with the actual content of the conference, finding the accusations to be misplaced. And yet, there is some disquiet about the current state of economics. This invites the question, is there a problem? If so, what precisely is it?
When theory and evidence don’t match
It is now widely recognised that some parts of traditional ‘textbook’ theory do not correspond well with the evidence. There are five possible ways of relating to this situation, which are not mutually exclusive:
(i) True belief mode: maintaining the theory is considered more important than the awkward facts that contradict it. This involves being in denial and clinging to discredited aspects of traditional standard theory, even when they have been shown to misrepresent the real world.
(ii) Reactive mode. The converse is the view that demonstrating the falsity of (aspects of) standard theory is in itself a contribution to knowledge. In its pure form, it is mere criticism and produces no new ideas. Typically, it involves the belief that all mainstream economists are dedicated to true belief mode and repeats the same accusations that have been around for many decades. This should not be confused with the legitimate critiquing of existing theories that is part of the process of developing new theory.
(iii) Distancing mode: the explicit rejection of traditional simplifications and dogmas. For example, Diane Coyle emphasises that nowadays ‘Economics does not require that people be rational, calculating automatons … it assumes there is a fog of uncertainty … Economists know that few markets are perfectly competitive, … and disequilibrium is the norm.’2 It focuses attention on what aspects of standard theory are not regarded as true. The positive case is that economists are mainly concentrating on producing good empirical work, e.g. evidence that is policy relevant.
(iv) Incremental mode maintains that traditional theory is a good starting point, that can be modified to fit the empirical data. It provides a benchmark from which one can study departures, e.g. ‘altruism towards our children’ as a departure from ‘the infamous homo economicus theory’, ‘irrational behaviour when drinking’, market imperfections, etc.3
(v) Selective replacement mode — evidence, ideally of diverse types, is used as the basis for generating theory. More on this below.
The first two positions, the extremes, clearly do not move the debate forward. In fact, they obstruct progress. When reactive mode gets the target wrong, it generates discussion around the wrong issues. The same old arguments are traded back and forth: perfect competition is assumed vs no, we study market imperfections; assumptions are unrealistic vs. we relax those later; humans are not rational vs we study departures from rationality; etc. I will focus on the other three.
Before proceeding further, it is necessary to say what I mean here by ‘theory’. I follow the usage in the natural sciences: an account of the causes that bring about the observed phenomena. It is not necessarily in the form of a model. Development of theory requires the ability to integrate disparate sources and types of evidence, and to produce a causal explanation. In economics this will typically involve multiple causes, and heterogeneity between different economies and at different times. The result is an empirically based causal theory.4 Ideally, for each causal relationship one can characterise both the mechanism and the phenomenon that it produces. Models can then be built, embedded in the wider descriptive theoretical account, by selecting some of the identified causes.
Responding to the mismatch
Distancing mode. Economists who put forward this view are often doing good empirical work, which may have important policy implications. Examples include the impact of immigration on training for native citizens, the causes of the decline in the share of national income going to labour in rich countries, and the effectiveness of international aid.5 Some past examples have had an important impact on policy, notably the evidence on the limited employment consequences of minimum wage legislation. This view is a legitimate defence against the false criticism of being over-dependent on unrealistic theorising and abstract mathematics. But by maintaining that their work is independent of standard theory, there is a danger of ignoring the need for new and better theory. We could be in danger of neglecting such ‘big questions’ as the existence of large-scale unemployment in some economies, the apparent ability of modern (post-industrial revolution) economies to grow indefinitely at previously unprecedented rates, and the now-well-recognised tendencies towards growing inequality in some countries, as well as periodic crises.
Incremental mode, in taking standard theory as the starting point, accords it special priority status. This can narrow the focus and lead to the neglect of other causes. An obvious example is the way that DSGE modelling was used before the financial crisis, with an unrecognised assumption that the financial sector would continue to operate perfectly. In such a case, the traditional accusation of over-simple models is valid. A solution is to embed one’s model in a previously-constructed empirically-based causal theory, which allows one to see what is being omitted.
This is not the only, or perhaps even the main, problem. And it is certainly not confined to macro. A more general implication of incremental mode is that it necessarily implies two processes: one that accords with theory, and one that modifies it. An example is the concept of ‘bias’ in behavioural economics, if taken to imply that standard theory corresponds to one causal process, and in addition one or more other processes are operating. This works well when the theoretical starting point is causally correct. But if there is no actual causal mechanism corresponding to standard theory, then neither of the suggested causal processes can be considered to exist — a case of double error.6 This may be far more widespread than just behavioural economics.
In some cases, standard theory implies the wrong causal direction. In such a case, any suggestion of an incremental modifying force will be meaningless. Much of the academic literature concerning the copious poor-to-rich country capital flows (especially from China to the US) that began around 2000 is rooted in the idea that all capital should flow from rich to poor countries, due to their differential rates of return — the Lucas puzzle or ‘paradox’.7 However, the correct causal direction is readily discerned from the available historical/institutional and statistical evidence: rather than the existing capital stock being the primary cause, it is clear that the phenomenon results from the underlying economic forces (in China’s case, highly profitable manufacturing) that generated the capital flows. This is recognised by the participants and by well-informed economic journalists.8
Selective replacement mode implies that theory is derived from evidence, rather than the more time-hallowed tradition of starting with an a priori model. This is routinely done in the natural sciences such as biology,9 but has so far only been patchily applied when generating economic theory. There have been numerous objections to this idea, some based on old ideas from the philosophy of physics, but they do not stand up to scrutiny.10
Crucial here is that when a satisfactory causal account is achieved, this leads to a replacement of the pre-existing theory. All too often in economics, one finds that old theoretical notions continue to be presented alongside the evidence that demonstrates their inadequacy. They do not disappear. Thus, the evidence on China-to-US capital flows has not led to the abandonment of the erroneous starting point of the Lucas puzzle. And to take another example, the source of money in the modern economy (e.g. Britain) is now well understood,11 but one still commonly encounters the notions that banks act simply as intermediaries, or that they ‘multiply up’ central bank money to create new loans and deposits.
Ultimately, the feasibility of selective replacement mode — and evidence-based economics12 more generally — can only be conclusively demonstrated by doing it in practice. It is achievable, especially now with the abundance of good evidence, improved analytic methods and prevalence of empirical research. It requires going beyond evidence collection on individual topics, and consolidating often-disparate types of evidence into a coherent theoretical structure. Its implications then need to be subjected to continuing hypothesis testing. It is likely that such a process is already occurring in some subdisciplines; if so, it would be helpful if this were made more widely known for economists in other subdisciplines and other interested people.
It the accusation is that economics is over-theoretical, abstract and mathematical, there may now also be the opposite danger: neglect of theory in the pursuit of evidence that is practically useful. Perhaps more prevalent is the insistence that traditional theory provides a good starting point, even when the evidence demonstrates that it actually makes the task of causal explanation more difficult.
2. Coyle D, ‘In defence of the economists’, Prospect Magazine. https://bit.ly/2JRuxvW
3. ‘Dismal ignorance of the “dismal science” – a response to Larry Elliot’, Prospect Magazine. https://bit.ly/2Osty7j
4. For a natural science example, consider the germ theory of disease. https://www.cogentoa.com/article/10.1080/23322039.2017.1280983.pdf
6. Joffe M, ‘Mechanism in behavioural economics’. Journal of Economic Methodology, 2019, in press.
7. Lucas R E Jnr. 1990. ‘Why doesn’t capital flow from rich to poor countries?’ American Economic Review 80 (2): 92-96. http://www.econ.nyu.edu/user/debraj/Courses/Readings/LucasParadox.pdf. In this paper, Lucas explains how zero flow is possible, but not copious poor-to-rich country flow.
9. https://www.cogentoa.com/article /10.1080/23322039.2017.1280983.pdf
10. Joffe M, ‘Economics based on evidence’. Paper presented at the AHE 20th Anniversary Conference, Leicester, 2018.
11. Ryan-Collins J et al. 2011 Where does money come from?, New Economics Foundation; McLeay M et al. 2014 ‘Money creation in the modern economy’, Bank of England Quarterly Bulletin. Q1: 14-27; http://www.res.org.uk/view/art5Apr17Features.html