What IS wrong with economics

Last article I wrote about what people misunderstand when criticising economics. This time around I’m going to join them and give my two cents regarding what is wrong or at least worrying about economics. I will talk about two issues: the mathematisation of the discipline and economics becoming increasingly divorced from applications. On the second count there is the related issue of the absence of much discussion of values or politics or psychology in economics education. What these things boil down to is economics’ desire to be considered a science. I will argue that this is a misguided attitude, at least to an extent.

Ask any economics student about trends in the discipline and they will note how it is becoming increasingly mathematical. For example, in the case of some current generation Emeritus professors, they can sometimes only perform ordinary least squares regressions, and maybe probit analysis. They employ research assistants to do their statistics for them. My generation of economists is expected to have taken six or more statistics courses and be fluent in a variety of techniques as well as their underlying mathematical principles. This expansion in the statistical arm of economics is understandable and I think welcome. Statistics have become much more accessible with the advent of modern computer technology and so it is understandable that their presence in the discipline has expanded, and empirics are much more convincing than theory in many cases. In almost all cases, theory is much convincing when backed up by data.

One concern though is the rise of papers doing statistical analyses to confirm things we already know extremely well. For example, a slew of papers is published every year (albeit rarely in top journals) discussing the use of new instruments to replace instruments that are already extremely good and explain things we have already confirmed. Often these new instruments aren’t even as good as the established one. This phenomenon can be partly explained by the ‘publish or perish’ culture that dominates modern academia, which encourages people to write statistical papers because they are often quick, cheap and easy. But part of it is also the result of a generation of economists who aren’t very good at thinking about problems, formulating research questions and applying innovative techniques because they have simply be drilled on foundational mathematics. Including more and more statistical and mathematical material in curriculums is squeezing out theoretical and applied courses. Core micro and macroeconomics courses of course remain, but students don’t have many electives left over to study things like labour economics, welfare economics, public economics, energy economics, case studies in economic policy etc. As a consequence they are somewhat ill informed about the various economics problems that exist in the world.

Alongside the expansion in required statistical courses has been a dramatic expansion in the pure mathematical skills required to undertake graduate study in economics. To return to the comparative analysis done earlier, recently retired professors have noted to me that Hamiltonian equations where just starting to become popular when they completed their doctorates. My generation is expected to be fluent in Hamiltonian techniques by the end of undergraduate. Systems of fourth order differential equations and advanced set theory are not uncommon amongst graduate dissertations. In fact, such things are increasingly the norm. Other colleagues tell me that most American graduate programs are now recruiting almost exclusively students with undergraduate mathematics majors alongside their economics degrees and turning away high performing candidates with combined humanities-economics degrees.

The most important consequence of this increasing emphasis on mathematics is that economics is becoming increasingly irrelevant to policymakers, and sometimes even reality. Mathematical models need to be highly stylised. To achieve this they need to make a lot of assumptions, some of which aren’t entirely fair. Often they need to look at an enormous period of time (sometimes even an infinite time horizon). They also need to have a neat internal logic. For example, many contemporary models converge to steady states and balanced growth paths in ways that don’t seem entirely natural. The more these factors are present the more modern economic models come to explain trends in ways that aren’t very useful. For example, modern growth theory, barring some rare exceptions, can only explain growth across the entire world economy, or at most in some frontier economies like the United States. Its key variables are the existing level of capital in an economy and the rate new ideas are generated, both of which are determined exogenously. This gives no help to developing economies who adopt technology from other countries rather than generating it themselves, or who might have specific development goals through which they approach their growth targets.

In the old days economists started with theory, then made a mathematical model and then cross-checked it with statistics. In most cases the endogenous variables—the ones that could be controlled—were very accessible policy instruments. Perhaps the most obvious example is Keynes’ attitude to monetary and fiscal policy in downturns. While Keynes was indeed a mathematician, his work on these topics is largely devoid of mathematical explication. Instead, he uses word based arguments to expound his thesis. It was up to Hicks and host of other subsequent researchers to develop Keynes’ theories into a model and then to check whether the model fit. In the meantime, the depression was resolved, crucially because Keynes’ model relies on the interest rate and the money supply, both of which can be controlled by the central bank.

Nowadays things are different. Models often do not directly address policy problems and in many cases the key variables in models are completely inaccessible to policymakers. For example, the gravity model of trade states that the volume of trade between two countries is principally determined by their proximity and the size of their economies. Not only is this conclusion blatantly obvious but it also has no policy implications because a nation cannot move itself closer to another, and policymakers are typically interested in using trade to increase the size of their economies, not the other way around. Crucial trade policy questions, like the impact of reduced tariffs and other protections on domestic industry or the relative difficulty of convincing a polity to open their borders, are not included the model. Yet it is the gravity model that dominates trade economics courses rather than the study of the political economy of trade. This is despite the economic history of trade being much more a political history than an economic one.

The political economy of trade provides a neat window into just how disinterested in reality economics is becoming. Economic theory is unanimous in its endorsement of the efficiency gains provided by trade. However, it is largely silent on the equity outcomes of trade liberalisation, and also passes over the complications posed by political factors. As a result, huge numbers of economists advocated the liberalisation of agricultural sectors in Africa in the later part of the 20^th century with disastrous results. Economists had not calculated that political complications would dominate economic rationalism in negotiations between the United States, Europe and African nations in relation to trade barriers. The result was a situation where African nations were open to American and European products but American and European markets remained closed to African goods. When Dani Rodrik, an Economics Professor at Harvard, began pointing out the record of trade liberalisation in Africa to his colleagues his was asked ‘not to give ammunition to the Barbarians’. The Barbarians in this case are people who oppose liberalisation of American industries. They are indeed foolish, but they are also politically powerful. Assuming politics does not make a trade theory work any better in practice.

Another example is provided by the recent Global Financial Crisis, which brought modern ‘self-correcting market’ theories into the spotlight. Some economists had argued against the need for government intervention because it would inhibit the market’s ability to correct itself back to a stable equilibrium. Over long time horizons, it does indeed seem that markets self correct. If you look at a line plot of growth rates over the last five hundred years you will see a steady upward trend. However, these economic models say nothing of who bears the burden of crashes (the poor), how many lives are eased through periods of economic turmoil by government assistance and intervention in markets. Anyone with a modicum of humanities training will understand that while smoothing by government might increase the economic cost of a downturn compared to just letting the market correct, it decreases the human cost of downturns considerably.

I’ve inadvertently started discussing my second point—the absence of politics and values from mainstream economics education. Students are taught models and ideas regarding what causes market health and market failure, what government is good at and what it isn’t, and what affects factors, variables and parameters like the money supply, the savings rate, the marginal propensity to consume, labour inputs etc. At no point is there any discussion of what we might like the economy to do other than grow (or the costs of growth, for that matter). There is no discussion of what humanity needs from its economy and its economists, what economic issues should take priority (growth or employment, for example), whether market outcomes are just or whether social planners are better able to make ethically sound decisions.

Some may argue that neglecting values perspectives is sensible because you don’t want your economists to be politicised during their education. But I am in no way suggesting that answers be presented to these questions, merely that they be discussed. Moreover, if you do not enlighten economists regarding the values dimensions of their work they will come to inadvertently enshrine economic ideals as values instead. Economists who have not been taught the social and human dimensions of their discipline will come to enshrine things like growth, innovation, efficiency and productivity without understanding the consequences of prioritising those things. For example, a colleague of mine recently argued that the pervasive inequality of the United States was justified because it drives innovation, and life without such innovation is boring. This despite the fact that innovation is extremely difficult to measure while inequality is not, that ideas don’t have feelings while humans suffer in a very real way, that innovation benefits only those who can afford to adopt it, that boredom is a nebulous notion that has not entered in a meaningful way into any studies of happiness, that a good game of soccer and a nice meal is arguably just as entertaining as the latest iPhone upgrade, and that ‘progress’ is extremely difficult to even define—was the theory of Marxism-Leninism an innovation? Was the collectivisation of farms in China ‘progress’? What about the coal fired power plants that are now destroying our biosphere, or the internationalisation of Quinoa that is destroying Peruvian communities? What about the financial ‘innovation’ of securitised debt obligations? If economists are not taught from the perspective of real world issues and human needs but from the perspective of abstract mathematical models they will inevitably produce ideas that serve mathematical rather than human ends.  

So if all these changes in economics are bad, why are they happening? They are happening because economics wants desperately to be considered a science and science needs to be controlled, repeatable and testable. Policy directed research is never any of these things.

The humanities are regularly dismissed by hard scientists as ‘wishy-washy’ because they are often divorced from any possibility of empirical validation. Economics has historically been somewhat better because of its emphasis on empirically validating all its conclusions. There is an entire, compulsory branch of economics called econometrics that attempts to adapt statistical techniques to applications in economics (and sociology and politics etc). This has made economics substantially more solid, generally speaking, than say gender studies.

Yet despite incorporating empirics on a fundamental level, economics has still come under criticism from natural scientists. This is because of insurmountable problems in econometrics that mean it can never approach the purity of the controlled experiments conducted by scientists. The data that goes into econometric models is not controlled. In a laboratory setting you can fix each variable—soil acidity, sunlight, water levels, nutrient levels, the genetic makeup of a plant under study etc—and then adjust only one variable. In this way you can see controlled effects. Outside of the emerging field of experimental economics (and even there) economics cannot do this because its data is drawn from real life and therefore comes with baggage. For example, if you are studying the impact of race on university entry scores you may identify various characteristics—gender, IQ, athletic ability, parental income etc—that you can observe, measure and then plug into a model. But many other factors will always be present that you cannot account for because everybody is different, every social context is different, every age is different etc. The data economists use is therefore always contaminated in a sense.

Instead of occupying an enviable niche between the more abstract social science disciplines economics has instead tried to get ever closer to the pure sciences, shedding more and more of its connection to reality along the way. In the abstract world of mathematics and computational models one can make very clean statements and run tests that are controlled and repeatable. But where the complexity of reality is present, especially where politics is present, things instantly get muddy. For these reasons, economics has moved away from policy, away from real world applications and away from empirical and applied economics towards theoretical economics.

For a variety of reasons, purer economics is welcome. The research in the theoretical branches of the discipline is very important, interesting and informative and may well result in very interesting and important applications down the track. However, some part of the discipline must break away and fill the vacuum that is emerging. Society needs highly skilled humanities economists, not just maths-economists (or science-economists in the case of behavioural researchers and those engaged in tracking evolutionary aspects of economic behaviour). Yet our graduate programs increasingly reject these students (yes, I have a vested interest in the arguments I am presenting here) in favour of those highly trained in mathematics. Maths is important, but couldn’t there be two separate streams of graduate economics, one that recruits mathematical economists and teaches them computational methods and economic applications for mathematics, and another that recruits humanities trained economists and teaches them how to engage with social science problems using mathematical techniques? Public policy schools are starting to fulfil this role, but many of them do not offer economics PhD programs independently of their university’s economics faculties and so recruitment is still driven by math skills and core courses are still dominated by mathematical dimensions. This does not bode well for public policy where economic issues are increasingly coming to dominate.

So what are the problems with economics? It is not the rational actor fallacy, or that all economists are bastards. Rather, it is that the discipline is overly committed to becoming akin to the pure sciences and, in the process, neglecting those of its aspects that allow it to make substantial contributions to civic life and public policy.