A Philosophical Framework for Rethinking Theoretical Economics and Philosophy of Economics
Here Gustavo Marqués discusses the main tenets of his book with the above title, published by the World Economics Association.
- Bookish Economics
In his paper Credible Worlds: the status of theoretical models in economics Robert Sugden called attention to the difference between “real” and “model” worlds. In fact, Sugden conceives the model world as a “parallel” world, which triggers questions about whether a connection among them exists and in that case how should it be characterized. Other authors have considered the possibility that economic models are autonomous from both theories and the facts of real market economies, which opens the door for claiming that they could be described as completely imaginary worlds. Is it possible that a good part of economics consists of theoretical developments that have no connection to actual market economies? The history of some branches of science shows that this possibility cannot be readily dismissed. Consider the history of anatomy during the XIV century in Europe. It was taught at universities, and the classes consisted of two different simultaneous activities. From his desk the teacher delivered a speech about the subject of the day. He was the academic authority in his classes. The content of the speech was taken from classical texts, some of them written during the III century before Christ (particularly by Galeno). Given the impossibility of dissecting cadavers (this practice was impeded by a legal disposition that prevailed during a good part of the Middle Ages) the teacher had not had contact with the organs of real human bodies and literally repeated what Galeno wrote. The problem is that Galeno had not had access to human bodies either, and in some cases he attributed to humans what he observed in pigs and monkeys. Meanwhile at the center of the room a butcher and an auxiliary of the teacher were dissecting a cadaver to illustrate to the audience the main points of the lecture. But to their surprise what they saw was sometimes very different from what was taught. They were inhabitants of the real world but the teacher was living inside what may be called a bookish world.
In this book I argue that a good part of mainstream (and maybe some non-mainstream) economics is in a situation that is similar to that of middle age anatomy: it is trapped within the “model world”, solving theoretical problems suggested by previous texts. Even if finding solutions to bookish oriented problems may be found exciting by the members of the community, they may be irrelevant for approaching the main economic problems of market economies.
- Bookish Economics inspired conventional theoretical practice
From the very beginning, the construction of economic theory as a scientific discipline deliberately assumed a sharp break with the main features that characterize real economies. Mill, Senior, Menger, Walras and many other classical economists have stated explicitly that the results of economic science are dissociated from their applications. Lucas has been even more explicit about this issue and recognized that denying uncertainty is a condition for economics to be scientific. As a result, most of mainstream theoretical economic practice involves making up imaginary worlds inhabited by truly rational agents, where regularities governed by calculable risk can be found, and has focused on the examination of the properties of these representations. Such models are built on the following set of ontological and epistemological assumptions:
1) Specific economies “contain” laws, mechanisms or some kind of regularities arising from ex – ante rational agents´ decisions. They are invariant (stable) features of the economic processes, lying below the surface of economic phenomena.
2) One of the main roles of theoretical practice is to discover these invariants.
3) This invariant knowledge also applies to the future and is obtainable ex-ante (mainly by models).
To these ontological and epistemological assumptions conventional economics usually adds a practical one:
4) Without having the invariant knowledge mentioned in (2) and (3) it is not legitimate to implement economic policies (and whatever intended policy lacking such a foundation is doomed to failure).
From this perspective it is natural to stress that the role of social or economic theory is to investigate the unwanted (and unexpected) consequences of agents’ decisions. And it is also clear why those theorists need not consider incorporating uncertainty and common sense and expert knowledge into their approach. Besides, lobbying activity has no place when agents are able to take ex-ante rational decisions.
- Mainstream Philosophy of Economics (MPE): a history of submission and failure
Standard epistemology and philosophy of economics, which assumes a naturalized view of science, has capitulated to the standard way of doing economics. This orientation is usually referred to as “recovering the practice” of economic theory, which means to describe as legitimate any contribution of standard Bookish economics. However, MPE has failed in its many attempts to solve the so called external validity problem. So, it couldn’t justify the explanatory capability of Bookish Economics and its practical value for the implementation of economic policies. It leaves Maki’s question without a clear and well founded response: “Fact or fiction? Is economics a respectable and useful reality-oriented discipline or just an intellectual game that economists play in their sandbox filled with imaginary toy models?”
Focusing its attention on Bookish Economics and taking an a-critical stance regarding standard models’ results, mainstream philosophy of economics seems to be in the same weak position as the “Teachers” of Anatomy who were overcome by the rough practice of the butcher!
- Some suggestions for reorienting Economics and Philosophy of economics
If it is assumed that (a) both agents and theorists are aware they are facing an uncertain context, and (b) they hold epistemic and ontological beliefs consistent with this state of affairs, the proper way to approach economic phenomena should be very different from those that guide current modeling practice. Particularly, instead of mechanisms or economic regularities that keep running independently of agents’ expectations, the decisive role of lobbyists within open-ended and uncertain processes based on expectations should be incorporated into the analysis. The following assumptions could be the philosophical core of a new conceptual framework for economics:
1) There are economic processes based on expectations and characterized by radical uncertainty. Agents involved in such processes act in two different ways (as decision-makers or as lobbyists).
2) Ex-ante knowledge of invariant sequences of events is generally not possible (because there are few if any sequences of this kind); more importantly, such knowledge is unnecessary as support and justification for the implementation of economic policies.
3) The role of theoretical practice is to identify the many feasible “branches” of a “tree of plausible outcomes” as well as the restrictions that each sequence of events faces.
4) It is not known (and it is not possible to know) ex-ante what “branches” of the tree (what sequences of feasible alternative events) will prevail. Science cannot help us with this.
5) Other types of knowledge (common and practical knowledge as well as practical skills) are crucial for shaping those processes. It is a sort of know-how knowledge, closer to management and administration than to scientific economics.
6) Although – as was shown in point (3) – theoretical practice has an important role to play in shaping processes, what is crucial in this endeavor is another practice, which we denote as lobbying (interventional) practice, which is performed by a wide range of economic players (mostly different kind of interest groups who are able to operate on the relevant context and agents’ expectations).
instead of mechanisms or economic regularities that keep running independently of agents’ expectations, the decisive role of lobbyists within open-ended and uncertain processes based on expectations should be incorporated into the analysis.
The whole concept of theoretical practice should be rethought if economic processes consistent with the above assumptions are the target. Chapter Eight argues that there are some mechanisms of economic transmission at macro level which represent feasible (or credible) sequences of economic events. I insist on the concepts of “feasibility” and “credibility”. Feasible sequences can happen (they are attainable in our world), although it all depends on the interventions of many different lobbyists along the process. Correctly interpreted – as open ended sequences, not as mechanisms – feasible models could be useful. Consequently, when I speak of an alternative theoretical practice I am not demanding the invention of a new way of doing economic theory (a demand that would be rather foggy). Part of the required theory is already available (I mean, the feasible one): it offers sequences of stages which in principle could be reached and provides points of intervention for governmental administration and the several interested lobbyers.
Philosophy of economics could also be reoriented. To go beyond mainstream philosophy of economics the usual ontological and epistemological assumptions of conventional modeling practice should be critically examined and philosophical support to the above mentioned assumptions should be provided. It means to incorporate into the agenda an analysis of the ontological features of economic processes, like uncertainty, and to call attention to the decisive role of the practice of lobbing, Such problems as what kind of rationality, learning and useful theoretical practice can be achieved under uncertainty should be addressed.
- Philosophical Problems posed by assuming open ended, uncertain and intervenable processes
Assuming open ended processes subject to interventions and pervaded by radical uncertainty means that some problems should be rethought anew and put into the philosophers’ agenda. These are some of them:
There are two basic notions of individual rationality. First, there is a subjectivist and coherentist notion which conceives “rationality” as consistency among agents’ beliefs, preferences and actions. This is the one involved in Expected Utility Theory. Second, rationality is seen as a sort of calculation driving decisions that precedes the realization of their future consequences. I call it well-grounded ex-ante rationality. “Well grounded” because it is supposed to be founded on a calculation that takes into account the best available evidence (where “best” is that which is more determinant for the expected results). “Ex-ante” because the rational character of the decision can be recognized before its results are manifest. This is the kind of rationality that Keynes had in mind.
Following Keynes’ remarks some scholars maintain that under radical uncertainty there is room for rational action in this second (stronger) sense. They offer three different arguments in support of this claim: (a) agents can employ induction in order to form “rational beliefs” about future events; (b) agents can perform some kind of rational economic calculation; (c) agents are able to decide rationally following a number of “techniques” or conventions. These three strategies find some support in Keynes’ work. The first is developed in his A Treatise on probability, the second is described in Chapter 11 of his General Theory, and the latest is also suggested in his General Theory but it is made fully explicit in his article “The general theory of employment” published in 1937. In Chapter Five I contest the validity of these arguments. I claim that agents cannot behave rationally under uncertainty assuming rationality is of the well-grounded ex-ante type.
Could agents facing uncertainty behave rationally if rationality is understood in a subjectivist and coherentist sense? The answer to this question is vital for my distinction between “complexity” and “uncertainty”. I concede that rational behavior of this type is possible in a complex world, but I deny that possibility if uncertainty prevails. However, it has been claimed that even in these circumstances agents might nonetheless be rational. Contrary to this view, I claim that if some reasonable ontological and epistemic beliefs are attributed to economic agents, to be coherent they cannot behave rationally in this sense.
To get this result it must be realized that all agents’ beliefs should be taken into account (not only agents’ subjective probabilities regarding expected results, but their epistemic and ontological beliefs too). Radical uncertainty means that they do not have vital information about those future events which are going to influence the results of their present days’ decisions, and they know that. Besides, they also know that the influential impact of these causal events is more relevant the more they are close to the time in which the consequences of their present decisions will be revealed. This is so because the nearer in time is a causal event from its consequence, the less is the opportunity for intervening if a corrective action were needed.
Agents’ ontological and epistemic beliefs should be considered as a key part of their uncertainty, which means that they know that the most relevant causal factors affecting the results of their actual decisions cannot be grasped when their decisions are taken. These philosophical convictions show them that no ex-ante rational decision is available under true uncertainty.
What sort of behavior, if any, would be “rational” then? My answer is that provided economics cannot surrender rational terminology, being rational is to intervene a posteriori of the original decisions for the purpose of validating them. Agents who believe in the uncertain nature of the world are driven to become lobbyists.
What about theorists’ behavior? If theorists believe, like agents do, that the scenario is uncertain, their theoretical practice would have to be consistent with this belief, as long as their models take into account the special nature of their intended target. If the purpose of economic theory were to render account of the behavior of subjects that know (or think they know) they suffer uncertainty, their theoretical practice should assume that agents behave consistently both with respect to their set of preferences and expectations (which is the usual requirement) and regarding their epistemological and ontological beliefs. Agents should be represented as behaving consistently with respect to all their beliefs. Theoretical models in which agents are modeled as lobbyists should be seen not as mechanisms but as feasible (credible) sequences.
Testing and learning
The use of such essential concepts in economics, like testing and learning, are not conceptually problematic in conditions in which certainty or risk is assumed. They are not problematic given complexity (in my sense) either. But conceiving open ended economic processes in which uncertainty reigns, where consequently there are “no laws”, nor “invariants” or “mechanisms” to discover, both experimental practice and learning require particular justification. Certainly, we can gather precise information, restricted in space and time. The problem is that taking uncertainty seriously puts in question the relevance these data have for future situations. (See Chapter Seven)
Pluralism and Realism
A richer taxonomy of economic models, which leaves room for distinguishing among Bookish models and Open ended feasible sequences, may be of concern for the current discussions concerning pluralism and realism. Models that at first sight seem to be mutually inconsistent, may belong to different (incommensurable) categories: one may pertain to the world of intellectual exercises and the other refer to possible courses of real world markets. As long as they belong to different categories incompatibility is out of the point, because their purpose or intended target is different. Both models may be retained without paying the price of endorsing inconsistence as a virtue (or, at least, pretending not to see a sin in it).
On the other hand, those economists who are realists can readily pay no attention to the whole Bookish Economics. But, I suggest, they can benefit from the set of plausible sequences approaching an open ended, uncertain and intervenable world, all of them realizable. Even if many of these sequences seem not to be compatible accounts of economic laws or markets mechanisms when they are understood as mechanisms, they are all realizable and compatible views as long as they are seen as future possible paths to perhaps different ends. Feasible economic sequences are not incompatible descriptions or accounts of the same target, even if they are rival sequences and may inspire different policies. Realist and pluralists’ may consistently have a positive appraisal of all those views.
- Interventionist view of Economics
The Hayekean and Popperian programs for social sciences considered that their main task was to find out those not deliberately pursued (and often not wished) consequences of our actual decisions. This is the main program of Bookish Economics, which presupposes that there are laws or mechanisms acting on the back of individuals’ consciousness and commanding economic phenomena in an independent and inexorable way. To back such approach is the main task of mainstream philosophy of economics. To this project I oppose a different one in which given open-ended and intervenable economic processes based on expectations lobbyers strive (often successfully) to attain wished results. In fact, this is the whole point for the existence of lobby!
Against the view that sees Economics as just conceiving how reality works I promote an active view of Economics. However, I hold no commitment with voluntarism. Even if a lobbyist tries to impose a particular feasible sequence to get some definite results, he may be unable to attain it simply because he does not manage to gather enough strength to impose on the set of institutions and bring about the required measures.
Therefore the sequence of events that ultimately prevails does not express some sort of pre-established order. Rather it has been socially constituted, and it might not have happened at all. The temptation to naturalize the prevailing social relations is huge and history is full of examples of this class. It has been shown that presumed natural laws, like the incapacity of women to assume government responsibilities, were just social constructions. Economic feasible sequences have the same status and can be successfully implemented and managed.
A conventional economic view asserts that all the implemented policies designed to increase the income of the lower-wage people via fuelling aggregate demand will be a failure. Conventional economists say so because they assume some kind of law or mechanism which determines that fuelling the aggregate demand will result in inflation (not in more income and employment). Supposedly they know this result ex–ante and by scientific means. Other heterodox economists, on the contrary, think that the way to reach an increase in employment starts with an increase in aggregate demand. Supposedly they have discovered an alternative mechanism opposed to the conventional one! I invite economists and philosophers of economics to pay less attention to the search of automatic mechanisms able to take us (necessarily) to a success or a failure are available, and consider that political decisions and practical measures will be crucial in the resulting events. The final outcome is open (and there is some room for a sound administration on the part of the government).
Our approach emphasizes the deliberate aspect of the interventionist behavior and the possibility that such strategy may be successful. That is to say, we find it feasible to enforce some desired results by altering the relevant context or influencing the present decisions of other agents. If we are right, it is better to perceive social and economic events less as commanded by laws or spontaneous mechanisms than like open ended and intervenable processes constructed by the joint actions of the many mutually conflicting relevant agents.
The shift proposed means that a scientific approach like the one assumed by the usual way of modeling can be of little use either for analyzing market operations or as a guide for implementing efficient policies. The nature of economic processes suggests that it would be far more useful to apply a political analysis able to reveal their open-ended nature and the diversity of economic interests at play than a supposedly scientific analysis which aspires to find invariant laws, mechanisms or regularities.
From: pp.2-5 of WEA Commentaries 7(6), December 2017