“Markets are not efficient because we assume they are; they become efficient through a discovery process in which profit and loss guide innovation, reduce inefficiency and generate wealth. This process is what makes markets a better alternative to the state. Implicitly, these critics assume the state can correct market imperfections – ignoring that it suffers from its own limits of knowledge and benevolence. If they judged the state by the same standard they apply to markets, the picture would change: both have flaws, but only one has a built-in mechanism for improvement.”
– Mani Basharzad, writing at CapX on the often dire predictions economists have made, such as their mockery of Argentina’s reforms, or the old claims from the early 1980s that the Thatcher policy mix could not work.
The author of this article has good things to say about the limitations of the neoclassical school in economics, and the failure to understand that competition is a discovery process. Entrepreneurs make money precisely by acting on the basis of incomplete data and in the hope they get things more right than wrong. And when they succeed, or indeed fail, it generates new knowledge. One of the problems for a centrally planned economy is that in epistemological terms, it is barren. A point for those who see AI as creating some sort of fix for socialism to bear in mind.
Don’t do as I do, but do as a thousand others say they do.
Fraser’s maxims on economics:
Microeconomics is the law; not “law” as in the law that the legislature passes, but the “law” as in the laws of physics.
Macroeconomics is almost entirely bullshit.
The difference between competition and government control is that in competition one has to pay the price for one’s dumb decisions.
There are very few professions that one can be wrong nearly all the time and still keep one’s job. Economist and civil servant, however, are two examples.
(It used be be “weather man” was another, but weathermen are actually pretty good these days.)
I’m sort of in that same space, but not because macro concepts and theories are intrinsically BS.
I always figured that we are just very very bad at sussing out the variables that are needed for macro thought.
We can’t even agree on employment figures for a month.
We can’t even agree on employment figures for a month.
True but the provisional monthly figures, eagerly seized on by 99% of your media with the same unquestioning reverence as the BBC’s daily reports of Gazan casualties supplied by the “health ministry”, have for at least the last decade been consistently bullish or bearish (but rarely accurate) depending on who is in the White House.
“I always figured that we are just very very bad at sussing out the variables that are needed for macro thought.”
Makes me think of something P.J. O’Rourke wrote in the early 1990s:
“Ecology is the science of everything. Nobody knows everything. Nobody even knows everything about any one thing. And most of us don’t know much.“
bobby b:
I would not phrase it this way. “We” more or less know the important variables. The big problems are:
1) We are limited in what we can measure. We’d like to measure wealth, but models use GDP (“If every worker were staffed in the army and fleet we’d have full employment and nothin’ to eat”). We’d like to measure price inflation, but models use CPI change. We’d like to measure the amount of money circulating – good luck picking among the various money aggregates.
2) The time series we do measure are sparse. Most key macro variables are measured on a quarterly basis. It takes 10 years to get 40 data points. A lot changes in 10 years – consumer preferences, investment environment, taxation, geopolitics…
3) We cannot model very well how these variables interact with each other. We build models. They used to be huge, and didn’t work well, but nonetheless a popular joke was “An economist is someone who says, when an idea works in practice, ‘let’s see if it works in theory.'” Then came the Lucas critique, and we switched to models that were not so big, but still a computational PITA. They relied on, shall we say, less-than realistic assumptions, and “predicted nine of the last five recessions.”
4) There is just no escaping Goodhart’s Law.
In a perfect world, with infinite knowledge at one’s fingertips, it would/should be possible to build a theoretical multivariate regression model of how the macro-economy functions. Unfortunately, knowledge is not perfect, and the inputs to such a model change infinitely at timescales impossible to keep track of. So all efforts to build such a model are doomed to fail.
I believe Hayek once said something similar, and it is the principal reason why Government Planning of everything will never work. Society and the economy are the aggregate of billions of micro-decisions made daily by individuals, businesses and organisations beyond the ken or imagination of mere mortals.
Philip Aggrey
Interesting idea. Utterly doomed to fail. In principle. Consider the Three Body Problem…
Plamus’s comment has it.
(some of) Macroeconomics is definitely NOT bullshit-it has significant explanatory power. But to use it predictively at the quantitative level you need to avoid Goodhart’s Law and measure things that we can’t.
Having said that, I had a (very poor quality) PhD student once and we set about trying to do better than the econometricians at predicting things.
We could do a great deal better with much simpler models than they seemed to manage.
I never followed it up because the student was so bad (they wisely opted to graduate with an MSc), but the basic problem the econometricians had, that we avoided, was mindboggling overfitting.
“mindboggling overfitting”
This. So much this. If your data is data is quarterly and you’ve got even 10 years, that’s just 40 observations. Your model really can’t have more than a handful of variables or it will overfit for Britain.
Couple that with the monster dislocation over the Covid years and econometrics is trashed right now IMHO
The American conservative blogger Ace of Spades once mentioned being disappointed in Chaos Theory not resulting in any useful breakthroughs. I wanted to rant at him about how the main breakthrough should have been finding many ways that intricate computation-based models can shatter any connection to reality. I say “should have been” because the consequence of that breakthrough would have been academia and government learning to discount these extravagant models in fields like economics and climate science, and…
I’m not linking back to the Ace of Spades post because
1. It was just a side note to the thesis of the post where he mentioned it.
2. The Ace of Spades site isn’t designed to make it easy to look up old posts.
3. It was several years ago; the only reason I remember it is because I’m just that petty.
Clovis Sangrail, you might get a chuckle out of this econometric tour de force.
@Philip Aggrey
In a perfect world, with infinite knowledge at one’s fingertips, it would/should be possible to build a theoretical multivariate regression model of how the macro-economy functions.
In the computer world we have a saying: “This system would function perfectly were it not for all the damn users screwing things up.”
Macroeconomics fails for the same reason. Businesses and individuals keep failing to follow the dictats of central government. For example, when we adjust tax policy like a finely tuned clock to optimize government revenue these asshole citizens keep adjust their behavior to avoid those new taxes. Damn taxpayers.
I mean there are lots of other reasons that macroeconomics is mostly bullshit, but that is certainly one of the main ones.
In the US we have this organization called the Congressional Budget Office whose job it is to determine the economic impact of new legislation. Congress lives and breathes by their prognostications. Need it be said that they are always, Always, ALWAYS spectacularly wrong about everything?
I think the fact that “economics” is taught under two broad subdisciplines of microeconomics and macroeconomics is kind of like Medical School having two divisions: “Scientific Medicine” and “Voodoo Medicine”, or a university science course having two tracks “Astronomy” and “Astrology”.
Plamus,
I’ll see your dicks and raise you this.
Note that is US margarine consumption and divorce rates just in Maine.
BTW, someone mentioned recessions, The 2008 recession is an example of completely missing the mark. The bottom line of what caused this recession is the fact that no-one seems to suffer the consequences of their poor choices, or at least no one connected to power. Of course we all talk about “too big to fail” and the FDIC, but I think the root of the collapse was around one single thing that is almost never mentioned. Namely that S&P lied about the ratings on the mortgage securities that were zooming around like ninjas in a money printing factory. Had S&P properly rated the bad debt instruments the whole collapse could have been avoided. Nonetheless, surely with such a spectacular failure S&P is now out of business since they have completely lost their one important asset — their reputation for accuracy? Nope. Still flourishing. Surely a whole bunch of people got fired for their mendacity or incompetence? Nope, no-one. Oh, wait, I think they fired the janitor for failing to empty the trash cans attached to the shredders quickly enough though.
Agreed Johnathan Pearce – and the endless funny-money means the Discovery Process either does not happen, or is ignored. It concentrates ownership and control in the hands of people who get the Credit-Money at lower interest rates – or just get the money FIRST (the Cantillon Effect).
Government regulations have the same effect – for example de facto preventing farmers selling more directly to the public, and forcing them to use corporate middle-men (such as the centralized slaughter houses). Almost every industry is regulated to an incredible level – concentrating control, and making “competition” less and less real.
As for economics – read Ludwig Von Mises “Human Action” – or if that is a bit long for you, read Henry Hazlitt’s “Economics In One Lesson”.
Among other things, you can’t predict new scientific/technological advances. How can you possibly know when or if a new invention will be made other than by inventing it? Maybe when there’s nothing left to invent a planned economy could work well.
It is quite likely that the Chief Minister of “the Sun King”, Louis XIV, Colbert (not to be confused with the American television person – although the American Colbert has similar Collectivist opinions to the French 17th century Colbert) would have liked to have regulated farming, the vast majority of the French economy at the time, in the same way he regulated industry and commerce – i.e. control every aspect of it.
But the French Colbert dare not do that – for fear of armed revolt. Today the farming population is small fraction of the population – and governments, not just the United Kingdom – quite a few governments, are gradually moving to destroy family farms – so that people will be forced to depend on governments and “Partner Corporations” (the vast international corporations) for their food.
They will use excuses “Carbon Dioxide and Methane” (the Netherlands), “the need to raise revenue” or whatever – but the real agenda is obvious.
“But I will still be able to walk in the “rewilded” areas”
No you will not – your electric car (assuming you are one of the minority who will be allowed to have one) will not take you there (it is electronic – you can be overridden).
And if you walk from your rabbit hutch “apartment” (likely one room) – you will get a vast fine, as in Canada.
After all “you might cause a fire” – so you must be kept in your rabbit hutch.
And the depopulation agenda will also continue – although only for certain populations (“historically exploitative and oppressive” ones).
Short version – governments and corporations may say they are devoted to “the market” but they do not mean a FREE market.
In a perfect world, with infinite knowledge at one’s fingertips, it would/should be possible to build a theoretical multivariate regression model of how the macro-economy functions.
And even if it worked perfectly (which it wouldn’t) it would generate no actual understanding.
– Jorge Luis Borges.
If you make a map, a model, a formula that is an exact simulacrum of the thing itself then you don’t learn anything that you wouldn’t from just experiencing reality in itself.
― Arthur Stanley Eddington.
george m weinberg:
Among other things, you can’t predict new scientific/technological advances. How can you possibly know when or if a new invention will be made other than by inventing it? Maybe when there’s nothing left to invent a planned economy could work well.
Exactly. George Gilder, in his book Wealth and Poverty (1981) said something along the lines that to forecast an innovation is essentially to make that innovation. While I don’t buy into all of what Gilder wrote or writes, his understanding of information imperfections, risk taking and mistakes people make about so-called monopolies was brilliant. He took Austrian economics and jazzed it up even more.
NickM, challenge accepted. Margarine is passé, Bangladeshi butter is where it’s at. CSB: when I read a part of this story aloud in our office back in the day, I stuttered a little at “butter”, and thus was born a long-running joke about “Bangladeshi butt butter”.