A Science Too Human? Economics*
Bernard Guerrien (The Sorbonne, Paris)
(This article was commissioned for La Science en 2004 of the L’Encyclopedia Universalis.)
Economic science is far from being exact: the divisions between economists are notorious and their predictions are subject to disputes and revisions. Nor has there been any major discovery in economics in 2003 or in 2002 or in preceding years. One could even ask if there has ever been any; the annual awarding since 1968 by the Swedish Royal Academy of Sciences of a prize for economic science (commonly called the “Nobel prize for economics” although from this it does not follow that it is one) fails to convince.
Nevertheless, economic relations exist; they even constitute an important part of human activities, and a scientific mind can only try to understand them. Generations of economists, of whom the most famous have often had a solid scientific education, have tried; thus one cannot ignore their thought, and eventually the influence that they may have had on the evolution of societies. So doing an update on economic knowledge and on the theories of economists is consistent with a scientific approach – even if, in the end, the results are slight or subject to caution. Knowing that we don’t know, or that we know little, is also part of scientific knowledge.
Economic science and science
The term “economic science” is usually used to designate a collection of economic theories. By “science” one generally means a body of knowledge or set of theories about which there is a broad consensus: they are considered to be true on the whole because they have been verified – or at least non falsified – by experience or observation. But in economics it is not uncommon to see different theories coexist for a long time, although they concern the same phenomena and give rise to divergent, maybe opposite, predictions. One can offer two reasons for this, which explain why the situation in economics is radically different from sciences in the strict sense.
Economic theory and experimentation
Theories, whatever they may be, are in the beginning the fruit of imagination, of beliefs, or even sometimes of the opinions of those who formulate them. In order to sort them out, so as to retain only one concerning a given phenomenon, the ideal method is that of controlled experiment, where the studied phenomenon is isolated, by conserving only what is taken into account by the theory, apart from certain perturbations considered as trivial. But in economics such experiments are not possible. As John Stuart Mill noted in 1843:
The instances requisite for the prosecution of a directly experimental inquiry into the formation of character would be a number of human beings to bring up and educate from infancy to mature age ; and to perform any one of these experiments with scientific propriety, it would be necessary to know and record every sensation or impression derived by the young pupil long before it could speak. It is not only impossible to do this completely, but even to do so much of it as should constitute a tolerable approximation. One apparently trivial circumstance which eluded our vigilance might let in a train of impressions and associations sufficient to vitiate the experiment as an authentic exhibition of the effects flowing from given causes. (A system of Logic, Book VI. chap 5. point 3)
Knowledge and laws in economics
Certainly astronomers, for example, do not carry out experiments. However they use results reached by sciences that do undertake them and, above all, give an essential place to observation. The regularity of physical phenomena, their repetition, their universal character (in time and space, at least on a certain scale), enable astronomy to explain a great number of phenomena and, even, to make highly accurate predictions.
The situation is very different in economics, where it is impossible to find situations which in the main are differentiated only by the action of one or a few well identified factors – the first step towards the establishment of causal relationships and therefore of laws.
It is why it is not possible to find in economics laws taking the form of precise relations, always verified, between two or more variables, other things remaining unchanged -- this last condition being almost never verified, even approximately. Economists however create confusion by using the word “law” where it does not apply. This is the case, for example, of the “law of supply and demand”, according to which the price of a good whose supply exceeds its demand tends to decrease – or to increase in the contrary case. As soon as one tries to give a more precise content to this so called “law”, one perceives that it is very vague: who makes the price vary? And how? Is this price unique? Can’t it happen that the people who demand the product organise themselves and refuse to pay a higher price? Can they not purchase other goods instead? In fact, the use of the verb “tend” fits what economists can at the very most hope to achieve: to detect some tendencies in the phenomena studied.
Tendencies rather than laws
The word “tendency” suggests a direction, but not a certain result. The tendency is in itself the manifestation of a law, but this one does not appear clearly because of the existence of disruptive non negligible elements, that can be called “counter tendencies”, and for which it is not possible to isolate their effects. Thus, in place of talking of a “law” of the equalization of profit rates (widely evoked by David Ricardo, John Stuart Mill and even Karl Marx), we will say there is a “tendency”, because this equalization can take time and also resources in collecting information and in comparing various profit rates and the risks with which they are associated.
A more controversial case is that of the tendency for the profit rate to decrease (set out by Karl Marx). The idea is simple: if one thinks that all value comes from labour, and that with time, the accumulated labour (under the form of machines, equipment, offices, etc.), or “dead labour”, increases in comparison to the living labour; then the profit rate must diminish. But here there is only a “tendency”, which can be blocked by an increase of the profit (the share of the living labour appropriated by the capitalists), or by a diminution of the value of the “accumulated” labour (obsolescent or unused equipment). The problem faced by the theoretician, if the diminution of the profit rate is not evident, is to know whether it is due to the existence of counter tendencies or due to the erroneous nature of the theory, i.e., the tendency to diminution does not exist. Because it is not possible to have a controlled experiment to settle the matter, both points of view can continue to coexist indefinitely. ( . . . )
Economic theory and self-realisation
There is another aspect of economics that distinguishes it fundamentally from the natural sciences; its theories can transform the world that it studies. This is what we call, not altogether correctly, self-realisation.
The discourse of economists, their predictions and their speculations often turn out to be erroneous. Even so economics influences the people at which it is aimed, and whose actions shape economic life and constitute its substance. People base their economic decisions on more than just “objective” factors such as tastes, available technology and the distribution of resources. They also base their decisions on their beliefs at the time of deciding, for example, beliefs regarding the “business climate” or future prospects. There is also the fact that the government, the big firms and the participants in stock exchanges act according to economics theories – which often take the form of mathematical models – whose form has thus an effect, more or less important, on economic reality, even if this influence is not that assumed by these models. It is this action of the subjective, of beliefs, on the real world that is called, somewhat incorrectly, self-realisation – which concerns the very special case where what has been predicted is realised as a direct consequence of the prediction itself.
Where then is the “reality”, the “real” world that the science intends to analyse and understand independently of the opinions and beliefs of the scientist? Two typical examples show why the answer to this question is not obvious. Assume that one has observed that an upturn in the stock market accompanied by a decrease in interest rates and an increase in household expenditure has been followed, say 4 out of 5 times, by a revival of the economy. If these conditions (stock market upturn, low interest rates, increase of spending) are observed at a given moment, and if the idea according to which they should entail a revival is widely held, then those who share it will, by their actions, make it happen. The revival is thus as much a consequence of shared beliefs, concerning a causal relationship, as it is a consequence of the causal relationship itself (assuming that it really exists).
Another example concerns the stock market, where the beliefs of investors play a central role. Take the price of options, i.e., the premium that someone has to pay at a certain time to have the right to buy a security or commodity at a future date at a given price (fixed in advance). This premium will depend in particular on the expectations held regarding future fluctuations of stock market prices, on their “volatility”. It is thus that Fisher Black and Myron Scholes (Nobel Prize winners) proposed a formula to calculate the price of options – assuming among other things that stock exchange prices follow a law of the type “random walk”. If all actors in stock markets adopt this formula, attributing the same value to these parameters, then this formula will, very precisely, give the observed prices of the options. But is one to say that Black and Scholes’s model perfectly explains reality, as if it were independent of the model? No, of course not. One can, at the most, note that there is consensus among the actors on the price of options – all agreeing on the price given by the formula of Black and Scholes, which plays the role of a convention. Here the conjunction of the subjective and the objective is at a maximum.
The beliefs of members of a society and the theories and models of the economists – resulting from their beliefs – are facts, data, which can play an important role in the economic life. Even if they are difficult to figure out and define, a scientific approach in economics must take them into account – even if this has the consequence of rendering vain or impossible purely mathematical formulations (something the profession has difficulty accepting).
Economy and Mathematics: a serious drift
The prominence that economics books and journals give to mathematics, sometimes very complex, is impressive. Together with physicists, economists are probably those who use advanced mathematics the most. There would appear to be a paradox here: mathematics being synonymous with rigour and precision, how is it that they can play such a role in a discipline where vagueness reigns?. The answer probably lies in the roots of this vagueness: the economic and social world being particularly difficult to grasp schematically, to reduce to simple laws, the temptation is great to flee it and to take refuge in fictitious worlds, in models having little to do with what we can observe (especially concerning forms of social organisation), but which lend themselves to endless mathematical developments.
It is symptomatic that among the journals of reputable disciplines, that it is those of economics that have, by far, the highest proportion of purely theoretical articles, with lots of mathematics but without any concrete data (this also happens in theoretical physics, but much less). Some economists – including famous ones who have sometimes built their reputation on their mathematical expertise – lament this state of affairs. These include the Nobel Prize winners Wassily Leontief, John Hicks, Paul Samuelson, Robert Solow and Joseph Stiglitz. Nevertheless, the recruitment and selection processes for economics teachers and researchers continue to privilege those who demonstrate (particularly in their publications) their knowledge of mathematics, thereby perpetuating the situation or even making it worse. This approach that economics uses to give itself the image of having a scientific character can, however, have the opposite effect, by providing evidence that economists are charlatans and pedants, who try to impress others with their formulas, while the predictions that follow from them leave, at the very least, much to be desired.
Economy and Ideology
The desire to prove that economic science could be different from other human and social sciences, because it can be put into a mathematical form, also leads to aberrations. It was with this desire that the currently dominant theory of the formation of prices was originally proposed in the 1870’s by Leon Walras, who above all sought to determine prices that would, according to him, be “fair”, that is to say, such that the rights of each person would be respected. In order to do this, Walras conceived a form of social organisation where prices are “called out” by an entity exterior to the traders, and where there is “tatonnement” (without exchanges) until the “faire prices” are reached, (these prices happening to be those which equalize the global demands and supplies). The mathematic form which was progressively adopted to represent this system happens to describe a very centralized economy, where the person who calls out the prices, and makes them vary, plays an essential role, especially by organising the exchanges (they can only occur through him). However, this very special form of organisation is necessary for the demonstration of what is considered to be the principal result of economic theory: there exists a system of prices which equalizes, on the basis of these prices, the supplies and demands. The demonstration of this “existence theorem” – for which the Nobel Prize was awarded to those who first made it, Kenneth Arrow and Gerard Debreu – is undoubtedly a piece of technical wizardry. But it also is the source of a great confusion, because it is systematically presented as proving “mathematically” that a market devoid of hindrances – “perfect” – always leads to a desired situation (where the choices of participants are compatible, and thus realizable). This is absurd, because the demonstration assumes a very centralised form of organisation, the opposite of the idea one generally has of market systems. Only a central planner can possibly be interested by this “theorem”. Even so it is generally what formal economic theory relies upon and pretends relates to markets.
Another example of this kind of absurdity are the “representative agent” models, very fashionable since the 1990’s (one here thinks of another Nobel Prize winner, Robert Lucas). These models suppose that the observed evolution, “macroeconomic”, of some of the basic variables of an economy (such as the GNP, consumption and investment, and employment and price levels) can be assimilated to the choice of a unique individual (obviously imaginary), who is both a consumer and producer, and who decides to divide his available time (present and future) between labour and leisure, and to divide what he produces between consumption and investment. Various mathematical techniques are then used – among others, the optimisation of non linear programs – to determine the share which enables this individual to maximise his (present and future) satisfaction. The result obtained is then compared to that of the economy on the whole (at it appears in statistical series, concerning employment, production, etc.), and theoreticians then try to give the parameters, which characterize the “representative agent”, values which reproduce at best the observed evolutions.
Only ideology – here, the belief in the all-powerfulness of mathematics joined to the virtues of the “market” – can explain why people, otherwise very reasonable, can dedicate their time and energy to these types of models.
Do we need economists and economic theories?
For a critical mind, the situation in economics is like this. On the one hand, it consists of an important accumulation of facts, data and statistical treatments, more or less elaborate, which try to bring out relations or tendencies by relying on relatively simple theories – but between which it is hardly possible to discriminate, because the elements which are not included in each theory are numerous and often non negligible. On the other hand, it consists of endless speculations, which use mathematics like Moliere’s doctors used Latin, trying to make us believe in the scientific character of the discourse, when on the contrary, it is the scientific approach which is sacrificed.
Many economists, however, both undertake sensible studies, on specific points, using a certain number of simple ideas, and yet participate in the speculations of the “grand theory”, when it has (almost) nothing in common with what they do when they carry out their empirical studies. The “simple ideas” that are the basis of these studies are generally old ideas, the fruit of observation and of the experiences of our societies. Thus, there is now a fashion for the “asymmetry of information” (a theme which led Joseph Stiglitz to his Nobel Prize in 2001). This term refers to the fact that in many transactions the parties involved do not have the same information regarding the object of the transactions. A typical example is that of the relation between insurer and insured, or between a banker and a borrower. Insurers and bankers have always known of the problem and tried to avoid it, but without talking about “asymmetry of information” or trying to put it in a mathematical form. Stiglitz, however, has earned his stripes (and the Nobel Prize) by “demonstrating” that the existence of asymmetries of information profoundly modifies the behaviours and the allocation of resources – a thing we have known for long time. But he also has lead concrete studies, based on observations and available data, where he shows the importance of the asymmetry of information to many important questions of economic policy. To do so, he called upon a few simple ideas, accessible to all, far from the mathematical formulas of his academic publications. The conclusions he comes to, and the policies he recommends, are however far from being approved unanimously, as testifies the controversy – at the end of the 1990’s – between the IMF and the World Bank (where Stiglitz was at the time the chief economist) on the way to tackle the crises which then affected certain developing or “in transition” countries. It is clear that mathematics are not the element which will enable the settling of the controversy, and that behind it lie very different visions of the world and different arguments concerning especially the consequences of the intervention of the State.
It is obviously unsatisfactory not be able to settle such matters. But knowing what are the arguments advanced, and on the basis of what observations and from what data, is part of scientific knowledge. Given the importance of things economic in the life of our societies, this knowledge is necessary, even though it is inevitably limited.
* Translated by Emmanuelle Benicourt and Edward Fullbrook.