Wednesday, August 21, 2019

Econometrics and the problem of unjustified assumptions — Lars P. Syll


This is important but may be too wonkish for those who are not intimately familiar with econometrics. So let me try to simplify it and universalize it.

The basic idea in logical reasoning is that an argument is sound if and only if the premises are true and the logical form is valid.  Then the conclusion follows as necessarily true.

This is the basis of scientific reasoning.

In modeling, a set of assumptions, both substantive and procedural, is stipulated, that is, assumed to be true. In a well-founded model all the assumptions that make substantive claims are known to be true empirically on the basis of evidence. This is called semantic truth. The logical truth of logical form is formal proof. This is called syntactical truth. Only the former contains substance. The latter is purely procedural.

A key methodological assumption of the scientific method is naturalism. Being "scientific" signifies being based on observation, rather than say, intuition or "common sense," that is, self-evidence. No self-evident first principles — that's doing philosophy, not science. Not that such speculation is not useful. It's just not science and should not be conflated with science. There is often a tendency to do so.

This presents two major difficulties with scientific modeling versus philosophical speculation. The first is the empirical warrant of the starting points, the stipulations that are assumed to be true and serve as the premises of the argument. The second is knowing that all relevant information is included in the assumptions. This is called identification.

Paraphrasing Richard Feynman, we do science in order to avoid fooling ourselves and we are the easiest ones to fool (owing to confirmation bias, for example). This requires following scientific method scrupulously when substantial claims are made.

Keynes pointed out to Roy Harrod that econometrics did not conform to this strict procedure and that owing to the nature of the subject matter, economics was "moral science," which at the time signified what we would now call "philosophy." The social sciences and much of psychology fall into this category. They are basically speculative exercises that employ some formal methods that may be scientific, or not. 

Accounting is a formal method that is proto-scientific in the sense that double entry it is made up of tautologies. But the entries can be checked for substance against journals and inventories. It is a method to prevent fooling ourselves on one hand, and to prevent cheating on the other.

When accounting tautologies (identities) are interpreted causally, then causal explanation demands empirical corroboration through data, e.g., measurable changes in stocks and flows.

Lars P. Syll’s Blog
Econometrics and the problem of unjustified assumptions
Lars P. Syll | Professor, Malmo University

See also

Bond Economics
Comments On "Business Cycle Anatomy"
Brian Romanchuk

1 comment:

AXEC / E.K-H said...

Economics: No method to the madness
Comment on Lars Syll/Tom Hickey on ‘Econometrics and the problem of unjustified assumptions’

Neither Lars Syll nor Tom Hickey is known for having contributed anything of substance to the scientific progress of economics yet they cannot stop waffling about methodology.

Lars Syll tells us: “Econometrics is basically a deductive method. Given the assumptions, it delivers deductive inferences. The problem, of course, is that we almost never know when the assumptions are right. Conclusions can only be as certain as their premises ― and that also applies to econometrics.”

Wake up, Lars Syll, this is a truism for about 2400 years: “When the premises are certain, true, and primary, and the conclusion formally follows from them, this is demonstration, and produces scientific knowledge of a thing.” (Aristotle)

J. S. Mill was well aware of the pivotal question of methodology: “What are the propositions which may reasonably be received without proof? That there must be some such propositions all are agreed, since there cannot be an infinite series of proof, a chain suspended from nothing. But to determine what these propositions are, is the opus magnum of the more recondite mental philosophy.”

Now, what are the foundational propositions of mainstream economics? The verbalized Walrasian axiom set reads: “HC1 economic agents have preferences over outcomes; HC2 agents individually optimize subject to constraints; HC3 agent choice is manifest in interrelated markets; HC4 agents have full relevant knowledge; HC5 observable outcomes are coordinated, and must be discussed with reference to equilibrium states.” (Weintraub)

Because these foundational propositions are shock-full of NONENTITIES all mainstream models that are based upon them are methodological garbage and therefore econometric testing cannot possibly work. This does NOT mean that econometrics has to be thrown on the muck heap. Lars Syll’s conclusion “Econometrics doesn’t establish the truth value of facts. Never has. Never will.” is plain methodological nonsense. What is called for is a Paradigm Shift, i.e. the replacement of the Walrasian axiom set.#1

Keynes understood that microfoundations are false but he messed up macrofoundations. Historical fact is that Keynes was too stupid for the elementary math that underlies macroeconomics. To this day, Keynesians, Post-Keynesians, and MMTers use the sectoral balances equation (I−S)+(G−T)+(X−M)=0 which is provably false because it lacks the balance of the business sector, i.e. macroeconomic profit.

Macroeconomic profit is determined by macroeconomic accounting which is nothing but elementary math. The philosopher Tom Hickey maintains: “Accounting is a formal method that is proto-scientific in the sense that double entry it is made up of tautologies. But the entries can be checked for substance against journals and inventories. … When accounting tautologies (identities) are interpreted causally, then causal explanation demands empirical corroboration through data, e.g., measurable changes in stocks and flows.”

The point is that economists are too stupid to get macroeconomic accounting right and to determine macroeconomic profit: “His [Keynes’] Collected Writings show that he wrestled to solve the Profit Puzzle up till the semi-final versions of his GT but in the end he gave up and discarded the draft chapter dealing with it.” (Tómasson et al.)#2, #3

To this day, economists do not get the pivotal economic magnitude profit right. Econometric testing cannot work when applied to misspecified models by folks who fail already at the level of the elementary mathematics of macroeconomic accounting.

There is no point in expecting from Lars Syll/Tom Hickey or any other methodological loser to ever produce the true macrofoundations.#4

Egmont Kakarot-Handtke

References
https://axecorg.blogspot.com/2019/08/economics-no-method-to-madness.html