Here is the draft introduction and conclusion to my chapter on economics as history rather than physics.
Economists love to dress up as physicists. We like to put theories into the form of equations. However, there are important differences between physics and economics, and these differences are particularly pronounced in the case of macreconomics…
1. Politicians and journalists want answers to questions such as, “How many jobs will (or did) a certain stimulus proposal create?” However, it is not possible to give reliable answers to such questions. Economists who purport to do so are misrepresenting the state of knowledge that actually exists.
2. Economists would like to know which theories are ruled out by the data and which theories are supported by the data. However, our ability to make statements along these lines is quite limited.
3. I believe that the study of macroeconomic events is going to have to be comparable to the study of revolutions, wars, or other historical events. There will be many plausible causal factors per event.
4. It will not necessarily be the case that the best explanations for macroeconomic events will be a single “model” that uses the same causal factors for every event. Instead, each important macroeconomic event may have important idiosyncratic elements involved.
5. Many very different explanations for an event will be consistent with the data.
6. Neither the use nor non-use of equations will ensure clarity or logical consistency. Confusion may be embedded in verbal descriptions of macroeconomic theories. Confusion also may be embedded in equations.
7. Neither verbal descriptions nor equations express verifiable relationships. Macroeconomic hypotheses will contain assumptions that will be highly contestable.
I won’t repeat this comment again, but this post tends to confirm some of my forebodings concerning directions your book might take. The body of the chapter may contain incisive criticisms of econometrics, but these general conclusions seem, to the extent they are true, rather trivial.
Saying in effect “It’s complicated” and “Don’t be so sure” is making things too easy for yourself. Saying that economics is less certain and precise than physics is not really addressing a serious position, and neither is observing that economists must depend on contestable statements. So what? Since when do you object to contestable statements? You make them all the time and think that you’re right and those who disagree are wrong. Nor can one avoid doing this.
And I don’t see that this kind of thinking leads anywhere much. Not only do I believe that you have macroeconomic opinions (as recently as yesterday’s very interesting post (invoking “data”) on the monetary dial), but those in positions of power necessarily have such opinions, if only by implication. “Take no action concerning X” is also an economic policy. People will, and should, try to predict what will have the best consequences, making do with whatever bases for judgment (data, Intuitive plausibility, etc.) they have at their disposal. It can be useful to show that some of these bases are more reliable than others, but not to say that our conclusions will be uncertain or “unverifiable”. We have to do something or other, so we need guidance as to the least bad choice.
Well, it’s just one chapter. I hope in the rest of the book you basically ignore all this and tell us what we should believe about macro and why.
Frank,
Maybe what it means is that there is a lot of fraud going on with macroeconomic “science.”
“Take no action concerning X” is also an economic policy.
Indeed.
But if you can demonstrate, first, that any action you take is essentially unfalsifiable against your goals and, second, that taking action biases the actions of others in ways that are more destabilizing than taking no action, then that is pretty good evidence that the preferred economic policy should be “Take no action concerning X for all values of X”.
Yesterday morning NPR had a piece on the currency problems of, in particular, Brazil and India because of the impression that the Fed will take its foot off the accelerator. You can argue that, well, that would have happened anyway at some earlier time if the Fed didn’t exist or was operating on some predetermined market monetarist rule. But it’s hard to deny that behaviors around the world have been biased by behaviors of the Fed, and that usually independent and hedgeable random variables have become interdependent. That’s how bubbles form and risk predictions fail and financial disasters happen.
What did Darwin do?
He produced science. Not history. Not physics. Science.
How did he do that?
He identifies a problem raising pattern no one had conceived before, and he identified a causal mechanism which could account for that problem. (See Ernst Mayr’s One Long Argument).
Economics does the same thing. See Hayek’s casting of the problem raising patterns and the causal mechanisms.
The “history or physics” false choice goes back to the Germans, who failed to understand Smith and Menger.
Historicism is as much a mistake as scientism/positivism.
But what to do with macroeconomics?
What to do with macroeconomics?
We start simple.
Learning in the context of changing conditions and relative prices explains order.
What explains systematic patterns of order failure within that order of the kind we call ‘the business cycle’?
A start is to look for systematic failures or kinks in the causal mechanism of learning — ways in which the learning process can go systematically off track.
Well, doesn’t this happen?
Don’t we has all sorts of evidence indicating systematic yo-yo-ing in leverage, shadow money, liquidity, the debt service ratio, credit-to-GDP ratios, asset prices, and sectors like construction and transportation? (See a host of BIS research papers, including this one: http://www.bis.org/publ/work421.pdf )
We have a problem-raising systematic pattern — a systematic disturbance in the ordered pattern of plan coordination via learning and prices.
That is what a scientific problem looks like.
And we have the elements for a causal mechanism to explain this systematic pattern — the yo-yo made possible by money, credit, liquidity change, asset value change, and endogenous money and asset grow and destruction in the banking and financial, all tied up to the real economy in a systematically connected pattern across time thru interest rates and production pathways along with their inputs to production, including labor.