The economist as outsider

We economists present ourselves as outsiders, looking down objectively on human societies, but that isn’t how society sees us. This claim to impartial outsider status weakens economics. Like the protagonist of Albert Camus’ great novel, L’Etranger (The Outsider), we are discovering that it isn’t possible to be disengaged from society. Economics has to reconnect—and what’s more, it isn’t hard to see how it can do so.

When I say ‘we’, I do not just mean ‘I’, but a large number of academic and professional economists who, not surprisingly, have been re-evaluating the subject since 2008. In this article, I draw on many conversations with other economists over the past year or so.

Young people studying economics certainly think the character of the subject needs to change; there are student-driven reform movements at several universities. There is tremendous public interest in the economy too—a real passion to understand what’s going on in the world. Although students are eager for reform, the strongest impetus for these efforts comes from employers. Many of those I’ve spoken to in the past year or two have complained about the narrowness of economists. They can hire graduates who are technically very able, but who are wholly unable to apply what they’ve learned to the real world, have no practical data skills, are unable to communicate with non-specialists, are unaware of context or recent economic history, and have not been taught the new, policy-relevant areas of economics, particularly behavioural economics.

Wanting to see change in economics is, therefore, not a fringe agenda; nor is it just a question of changing the curriculum or the academic research agenda. Most of us were trained in pre-crisis days. Unless we really want to remain the slaves of defunct economists, the kind of impact assessment work undertaken widely in public policy and consultancies needs to change too.

Madness in our methodology

My first point is hardly novel, but it needs saying because economists have continued to shy away from the implications. It is that economics often cares more about logical rigour than reality, despite the increasingly desperate attempts by reality to get economists to pay attention.

Economists put models at the heart of their methodology. A model is an attempt to make sense of the world by including only relevant detail. A good model is a powerful tool for analysis and prediction.

One well-known example is the map of the London Underground. It’s sometimes a flawed guide—for example taking tourists down two deep escalators to travel 260m on a train, and then up in a lift, if they want to travel from Leicester Square to Covent Garden. Still, it’s an accurate enough representation of London’s geography for its purpose and an invaluable guide for passengers. Its combination of reasonably accurate analysis, parsimonious representation and sheer elegance makes it a model for economic models.

However, many economic models fall short of the Tube map standard. Most often, their failure is one of inaccuracy by over-simplification. Economists value logic, parsimony and elegance—and then believe the model. It’s a kind of cargo cult mentality, in which a useful tool becomes a venerated shibboleth. The linguistic philosopher, Alfred Korzybski, warned about this kind of error: ‘The map is not the territory’.[1] The aim in modelling has to be finding a happy medium: between navigating around London only via the Tube; and making the opposite error of piling on descriptive detail without any analytical abstraction in a kind of Borgesian extreme of regarding the entire territory as the only possible map.[2]

Take game theory, rightly considered to be one of the jewels in the crown of the discipline. It models formally how people behave in strategic situations, where what one person chooses to do will depend on what the others do. Of course, our assumption is that the players choose rationally: they will calculate what is in their best interests, given their assumption that everyone else will do the same. Game theory has been applied in practical contexts ranging from business strategy to spectrum auctions, with apparently excellent results.

Ariel Rubinstein is a game theorist who has experimented with specific games, or strategic contests, with his students, audiences at his lectures, and respondents to his website surveys, collecting more than 13,000 responses. His firm conclusion is that people do not behave in life as they are assumed to do in game theoretic models. Relatively few end up at the Nash equilibrium outcome predicted by theory. More opt for ‘naive’ solutions that take no account of how others might respond, and more still show signs of thinking strategically but getting the calculation wrong. What’s more, Rubinstein points out that if you are rational and can do the sums then playing the game on the assumption that others are too will leave you worse off than assuming that they will act naively or capriciously.[3]

Of course, many economists now acknowledge the importance of incorporating into economic models assumptions about decision-making that bear a closer relation to how people make decisions. This is the behavioural economics revolution. The results so far seem impressive, and so behavioural economics models have moved quickly from the research lab and seminar room to the corridors of power, and implementation in policies. Policymakers and regulators rightly have great interest in learning more from these models, and applying the lessons.

Cognitive science in general is obviously relevant to understanding how individuals make economic choices, and serious economists want to absorb and use this knowledge.[4] I think many economists are now pretty comfortable that psychological studies give regularities that can be used in economic models. For example, risk aversion or kinks in preference functions can easily be accommodated in a formal model, as can loss aversion, which field experiments are suggesting is an extremely powerful force in people’s decisions.

Another encouraging sign that economists and economic policymakers are embracing evidence is the enthusiasm for randomised control trials (RCTs) and field experiments. Often linked with behavioural models, these methods started in the development context but are quickly migrating to other areas of policy. The idea is that trials and experiments, if properly constructed, with participants randomly assigned to the treatment or control groups, will provide robust evidence of ‘what works’. For anybody keen to assess the impact of a policy or intervention, the combination of psychological realism and rigorous method looks unassailable. The approach allows realism without stooping to the ‘ad hoc’.

Unfortunately, interpreting economic evidence is not simple. There are limits to what can ever be known.[5] In economics we are typically trying to test hypotheses about a small number of variables in a complex world of millions of variables, with a huge amount of feedback and simultaneity, and using a relatively small amount of data, of dubious quality. It is difficult in the extreme to establish causality rather than correlation. Neither RCTs nor more realistic assumptions in economic models make a difference to the sheer difficulty of the empirical challenge. Nate Silver writes in his recent bestseller The Signal and The Noise:[6]

The government produces data in literally 45,000 economic indicators each year. Private data providers track as many as four million statistics. The temptation that some economists succumb to is to put all this data into a blender and claim that the resulting gruel is haute cuisine…If you have a statistical model that seeks to explain eleven outputs but has to choose from among four million inputs to do so, many of the relationships it identifies are going to be spurious.

Economists need to rely less on models alone. This is tough. We’ve all been trained to love the analytical muscle of modelling. We’ve extended the range of economic models by embracing behavioural economics and new empirical methods. Nevertheless, we need to supplement the analysis far more with narrative approaches, from both economic history and other social sciences such as anthropology and sociology. A combination of a model, a field experiment, and a narrative account of the cultural context and history relevant to the experiment could add up to a genuinely powerful approach to economics.

In the immediate aftermath of the financial crisis, there were many calls, from economists, for economists to demonstrate greater humility. David Colander presented to the ASSA meetings in 2011 a suggestion for a code of ethics with the title ‘Creating Humble Economists’. He wrote:[7]

Back in 1927, Lionel Robbins argued that, ‘What precision economists can claim at this stage is largely a sham precision. In the present state of knowledge, the man who can claim for economic science much exactitude is a quack.’ Despite the advances economic science has made, that remains true today. Yet, all too often economists allow lay people and policy makers to believe that our policy suggestions have far more scientific foundation than a neutral objective observer would give them.

Policy in Wonderland

Moving on from methodological issues to the practicalities of drawing up policies or assessing policy impact, in practice economists tend to ignore the consequences of economics being a social science, involving sentient beings who, all too often, change their behaviour in response to policy changes, or even policy assessment. In principle, economists know this, but they typically ignore the implications. We wrongly imagine we can stand outside the context we’re evaluating.

The habit of taking this stance also often disguises a transition from objective or positive assessments to highly subjective or normative conclusions. Milton Friedman, in his essay on the distinction between positive and normative economies, was adamant that positive conclusions can and should be the aim:[8]

[D]ifferences about economic policy among disinterested citizens derive predominantly from different predictions about the economic consequences of taking action – differences that in principle can be eliminated by the progress of positive economics – rather than from fundamental differences in basic values…a consensus on ‘correct’ economic policy depends much less on the progress of normative economics proper than, on the progress of a positive economics yielding conclusions that are, and deserve to be, widely accepted.

In fact, we conflate ‘is’ and ‘ought’, a conflation that is being highlighted by behavioural economics but which occurs frequently in economic assessments.

There are several levels to this. At the most basic level, economists have often forgotten to include any behavioural reaction in their assessments, or omitted it because it makes the models intractable. Unfortunately, reality has little interest in tractability. It’s like the game of croquet Alice finds herself playing, when the mallet turns out to be a flamingo and the ball a hedgehog. Alice is the policy economist, assuming the subjects of her intervention will respond in a fixed way. They do not. They are affected by each other and the path their decisions take over time is impossible to predict. Gregory Bateson said that in social science, the game is to discover the rules of the game.[9] It’s confusingly self-referential, and economics is intolerant of this characteristic.

Over the years, there have been countless examples of policy interventions that do not take account of the fact that people will react. Although this naivety is becoming somewhat less common, it does make impact assessments much harder, and there is a wilderness of inherited unintended consequences.

One example was in the UK Competition Commission’s 2003 inquiry into the market for extended warranties on electrical goods. This market, which was found to be uncompetitive, had come into existence in place of people using their insurance to cover breakdowns. In 1997, the government increased the insurance premium tax on cover for domestic electrical goods to 17.5% to match the rate of VAT at the time. It was meant to level the playing field. However, retailers can recover much of their VAT paid, so a bright management consultant sold the biggest retailer the idea of moving from insurance cover to the new extended warranty contracts. A £16bn a year market, from which very little consumer welfare derived, was born because an official assessing the tax increase had not thought that retailers might change their offer.[10] The terrain of public policy is littered with examples of what’s become known as the Peltzman Effect, or risk compensation, after the observation that safety regulations can increase risky behaviour in the context of roads and traffic. Although economists are familiar with the concept of moral hazard in finance, the wider applicability of this point is not often recognised.

The tension between observing and inevitably participating as well is common to all the social sciences, and causes discomfort to all conventional social science. It is widely ignored, and I think economists are even less likely than other social scientists to pay any attention to this tension. The economist’s outsider perspective is therefore problematic at the level of reality. It is also problematic at the level of morality, where economics suffers from several inconsistencies.

Our aim in using economics in the policy arena is to serve the public interest. We have in mind a concept of social welfare. Surely it is right to aim for the perspective of the impartial observer, to try to make an objective assessment of the welfare effects of a policy intervention? This is a constant theme of liberal theories of justice. At the same time, we know that economists and policy advisers are only human and we will respond to incentives and maximise our own utility.

Lessons from cognitive science have obvious useful applications, for example in evaluating financial regulations or competition remedies. Economics predicted that a bigger choice of providers of directory enquiries would increase competition, but in fact the deregulation increased concentration in the market.[11] Economics predicts that consumers will use APRs to compare the cost of loans, but if that were so none of us would borrow on credit cards, never mind on payday loans. Behavioural economics will prove effective in policies ranging from competition remedies, to financial and consumer regulation, to social policy. The idea of ‘choice architecture’ to ‘nudge’ people towards decisions that are better for them—on their own criteria—inevitably turns economists into paternalists.

So what does this do to our outsider status? I’ve described the confusions or ambiguities about our perspective as economists—are we inside the model or outside? Impartial observer or self-interested agent? Does the paternalistic economist know best, or is the consumer king?

The ambiguities matter because economists present themselves as technocratic experts in the domain of public policy, the toilers after truth discovering ‘what works’. New techniques like RCTs encourage this self-perception, the economist as objective scientist in a white lab coat, experimenting on society.

It might not be possible to sustain this self-image. Not only behavioural economics but the whole, very successful, project of empirical microeconomics will put economics firmly into the political arena, the arena of making normative choices.

Conclusions

I began by saying that economists make the same mistake as Camus’ Outsider: it is not possible to stand apart, disengaged from the society we’re studying. It probably never was, but that outsider perspective has become unsustainable.

In short, political economy is back. It is certainly back in a no-growth economy with no rising tide to make distributional shifts acceptable. And it is back because the genuine advances in economics in the past couple of decades, in empirical microeconomic research, will bring more and more examples of conflicts between ‘what works’—assessed according to the efficiency criterion of the ‘objective’ economist—and what people believe or want, even if it is not rational, or not even reasonable.

Mr Spock’s greeting, ‘live long and prosper’, is an apt motto for economists, but like the half-Vulcan, half-human Mr Spock himself, we ultimately can’t stand outside a particular moral and political perspective. We economists are right at the heart of the societies we study.

Diane Coyle

Contact: Fod Barnes