The WHY of Crazy Models

{bit.ly/WEAwhy} I was professionally trained as an economist, and learned how to build models with the best. As described in detail in a previous post on “The Education of An Economist“, it was only by accident that, a long time after graduate school, I learned of glaring conflicts between the theory I had been taught, and the historical evidence about effects of free trade and trade barriers. Further exploration along this direction dramatically widened the chasm between the economic theories I had learnt, and the historical and empirical evidence all around me. This led me to a set of puzzles which I have been struggling with for the past two decades. [1] Why is that economists are not aware of the conflict between economic theories and empirical evidence? [2] Why is it that economists do not care, when such conflicts are pointed out to them? In “Trouble With Macro“, Romer expresses these same two points as follows:  “The trouble is not so much that macroeconomists say things that are inconsistent with the facts. The real trouble is that other economists do not care that the macroeconomists do not care about the facts. An indifferent tolerance of obvious error is even more corrosive to science than committed advocacy of error.”

Once we move from the easy-to-establish fact that economists use crazy models, to the much more difficult meta-question of WHY economists use crazy models, one apparently obvious answer suggests itself: it is because economists are crazy. A referee once accused me of thinking that “economists are blinkered idiots“. Actually, from close association with the tribe, I know that some of the best and brightest human beings are economists.  This intensifies the puzzle: how can some of the smartest people believe the stupidest theories?  To pick some of the most flagrantly stupid theories:

1. Economists believe in rational expectations, even after the Global Financial Crisis took the entire profession by surprise. This requires not believing in uncertainty, despite overwhelming evidence to the contrary — see  Foundations of Probability 1
2. Economists believe in utility maximization, despite a huge amount of empirical evidence against it (see  Behavioral vs Neoclassical Economics).
3. Economists continue to believe in, and use, DSGE models, considered crazy by Solow.  Even sympathizer and supporter Blanchard acknowledged that these models  make assumptions profoundly at odds with what we know about consumers and firms. (see Quotes Critical of Economics).

How can someone in possession of his senses declare that it is midnight, when a bright sun is shining overhead? A powerful hypothesis is an ideological bias, as many have argued. However, from personal experience, most economists I know are not ideologically wedded to capitalism. It is not ideological commitments or class interests which drive the lunacy embedded at the heart of economic theory.

Instead of specializing to economics and economists, I began to think about the broader question: Why do people come to believe in theories (true or false)? Pondering this question led me to a startling realization. I had been conditioned to believe in positivism: We acquire knowledge by observing empirical evidence and formulating theories in the light of observations. The fact is that the body of human knowledge is the work of millions of scholars over course of centuries. No one man can hope to acquire more than a tiny fragment of this knowledge. Even the brightest mathematical genius, deprived of the heritage of human knowledge, and left to his own devices would be unable to progress much beyond third grade mathematics with a lifetime of work. So we have no choice but to take at face value, and accept without question, the vast portion of what we learn. This insight was expressed clearly by Kuhn in his study of the Structure of Scientific Revolutions. Scientists are trained to dogmatically believe in a paradigm, without questioning the fundamental methodologies which they learn like an apprentice. Similarly, economists learn how to do economics by examples, without any discussion of the methodological frameworks within which they are creating their theories.  The implication is that the vast proportion of knowledge that we have is inherited and unexamined knowledge; it has to be this way, because our lifetimes are too short to enable us to examine and verify the enormous structure of existing human knowledge. If massive errors have been made in our intellectual heritage, we will generally accept these without question.

Taking it to a personal level, I was taught a theory of knowledge, and a methodology of acquiring knowledge, without any explicit discussed of either of these topics. The textbooks that I studied made the assertion that what is contained in this book is knowledge (without explict statements to this effect). Similarly, the methods used to construct this body of knowledge — mostly theorem-proof, but with some others thrown in eclectically, as needed — were what I learned about methodology. I was not offered any choices about theories of knowledge via  discussion of epistemology, and I was not offered a choice about alternative methodologies. I started to question these methods only after I came to the realization that this body of “knowledge” was deeply flawed, and the “methodology” I used to arrive at “truth” often led to stark falsehoods. The problems with “Western knowledge” as it developed over the past few centuries are too numerous to list, but perhaps the root of all the problems is the (mis)conception of “objective” knowledge.

Objective knowledge is knowledge which has been detached from the knower – the subject who is in possession of this knowledge. Realizing that this cannot be done would lead to a revolution in the theory of knowledge (for an extended discussion, see “The Illusion of Objective Knowledge“). To take small steps, consider the difference between the “empirical” and the “actual”, as introduced by Roy Bhaskar in his philosophy of science known as Critical Realism. The empirical is the sense data that we perceive, and the actual is what is really out there in external reality. It is obvious that there are actually “trees” out there in external reality, but at the same time, our only access to these trees is via our five senses — we do not have any direct access to external reality, only to our perceptions. Because of differences in perspective and time, no two observers of the same tree will ever actually have exactly the same sensory experiences — the ’empirical evidence’ for the tree, and a description of its appearance, will vary radically from observer to observer and also vary radically with time and position for a single observer. However the “actual” tree is  much more stable as an entity – although it too grows over time.

Just as there is a difference between what we observe, and the actual object in external reality, so there is a chasm between knowledge that I acquire which is subjective knowledge belonging to the subject (me), and objective knowledge – a disembodied entity, the real, indisputable, and objective truth, invariant across time, space, and independent of the observer. Three central illusions of Western epistemology are:

1. There exists of body of OBJECTIVE knowledge – universal truths independent of observers. Also, we CAN aspire to get this knowledge.
2. We can filter out our subjective imperfections to distil and separate the perfect objective truths (facts) from our imperfect subjective analyses (opinions).
3. The “scientific method” is the (only) methodology which can be used to reach objective truths, and this is the only type of knowledge worthy of the name.

Clarity is achieved by considering the polar opposite positions. For a long and detailed explanation, see Hilary Putnam: Collapse of the FACT/VALUE Distinction and other essays.  At least as an idealization, we admit the existence of objective fact, and also of purely subjective opinions. But the vast majority of human knowledge consists of a mixture of the subjective and the objective in a way that the two are “inextricably entangled”. Another way to say this is to say that human knowledge consists mainly of our life experiences, which are quintessentially non-scientific. This is because our life experiences arise from our personal interactions with external reality – this interaction mixes our subjectivity (opinions, emotions, identity)  with the objective (facts, social realities, history, geography).

How does this help to solve the puzzle of how intelligent people can come to believe in nonsensical theories? Considerations discussed above provide one important piece of the puzzle — other pieces will be discussed later, separately. Modern Western epistemology teaches us that the only knowledge worthy of the name is objective and universal knowledge. But actually, nearly all of the knowledge we have is our personal subjective life experiences. “Methodology” then becomes the name of the process whereby our personal life experiences can be converted into universal truths, so that it counts as knowledge. Timothy Mitchell (2002, Rule of Experts) writes that: “The possibility of social science is based upon taking certain historical experiences of the West as the template for a universal knowledge. Economics offers a particularly clear illustration of this.” The process by which economists proceed is to take some “axioms” of human behavior as self-evident universal truths (even though they are actually false). Then they proceed to build models of the economy on the assumption that all human beings follow these rules of behavior. It is widely recognized that the process of modeling involves populating the economy with mechanical robots guided by mathematical rules, and computing the outcomes. The fact that this process leads to models disastrously in conflict with reality is ignored. This aspect, of how it became acceptable and fashionable to ignore reality, requires further explanation.

We started with a sequence of six posts on the nature of economic models, meant to clarify why economic models ignore reality. These six posts are: Mistaken Methodologies of Science 1, Models and Realities 2 , Thinking about Thinking 3, Errors of Empiricism 4, Three Types of Models 5, and Unrealistic Mental Models 6. This sequence is to be continued with further detailed explanation of the nature of economic and econometric models. To understand the difference between observational models as used in econometrics, and real structural models, we need to introduce and explain causality and how it affects analysis, even though econometricians ignore it. This is done in a sequence of 5 posts on Simpson’s Paradox.

Postscript: Next few posts in this sequence are: The Knowledge of Childless Philosophers 8, Beyond Kant 9, and Mistaken Methodology of Econometrics 10. A later post also addresses this same theme more succinctly: Why Do Economists Persist in Using False Theories?