Saturday, September 12, 2020

Theory vs History

Yesterday, Mike Huemer and I had a debate with two anarcho-communists on the topic of left libertarianism vs right libertarianism. It can be viewed on YouTube.

One thing that struck me as interesting was a difference in how we supported our positions. While I made some factual points, my central argument was theoretical, based on the coordination problem. In order for each of us to achieve his objectives he requires the cooperation of thousands, probably millions, of others, as you can see by tracing the supply train for anything we use — what things went into making it, what went into making those things, and so forth — and considering the number of people involved at each stage. Property and trade provide a decentralized way of solving that problem.

My challenge to my opponents was to explain how, without property, they would solve that problem. They never made a serious effort to answer it. Instead, they offered what they claimed was historical evidence that it had been solved, most notably by the Catalonian anarchists during the Spanish Civil War, with a variety of other claimed factual examples.

The problem is that neither I nor they knew if any of that was true. In my experience, if you look carefully at any controversial bit of history, you can find convincing arguments for a range of very different interpretations. I have not tried the experiment, but my guess is that a sufficiently able holocaust denier could offer persuasive arguments that it would take at least an hour or two of research to debunk. I have encountered cases where a false claim, for instance that Hoover's response to the stock market crash was to cut government expenditure or that Adam Smith supported public schooling, is widely believed. Unless one happens to have already spent time and effort investigating a historical claim, there is no way to know if it is cardboard or gospel truth. Even if you have, there is probably no way, in a context like a debate, to prove its truth or falsity to your opponents and audience.

Theory, on the other hand, is something you can actually look at inside your own head, without worrying about biased or otherwise unreliable sources of factual information. My confidence that a minimum wage must always reduce employment opportunities for low skilled workers was answered not by Card and Kruger's very weak factual evidence but by their demonstration, using the conventional price theory I accept, that there was a conceivable situation in which it would have the opposite effect. I doubt that situation has much real world relevance in the U.S. at present, since it depends on the market for such labor being monopsonistic, but it might well be relevant in other environments. 

Similarly here. Persuading me that the coordination problem could be adequately solved without some equivalent of private property and trade on historical evidence would be difficult, since I would have to look at lot of historical claims, find people who disputed them, evaluate their arguments for myself. A clean theoretical argument, possibly inspired by looking at real world examples, could do it in twenty minutes. To the best of my knowledge and belief nobody has produced such an argument yet, which makes me suspect that it cannot be done.

None of that, of course, is an argument against doing empirical work — the real world sometimes has lessons to teach us that we were not clever enough to think of for ourselves. It is an argument against giving much weight to claims about such work that you have not yourself investigated with reasonable care or, short of that, seen confirmed by people who have investigated them and whose honesty and competence you have reason to trust.



P.S. A commenter pointed me at a long essay by Bryan Caplan giving a very negative picture of the Catalonian anarchists:


I posted the link to the FB page of one of my opponents in the debate, and someone responded with a link to an essay rebutting Caplan's:


At this point I have read most of Caplan's essay, only the very beginning of the rebuttal, but I think the existence of two long essays giving opposite interpretations of the same historical episode, at least the first of which is convincing (and the second may well be),  supports my point above.


It might be possible, with enough time and effort, to read both, check against other sources, and reach a reasonable conclusion. Almost nobody who claims to know what happened in Catalonia has really done so — including, I strongly suspect, the gentlemen who wanted me to take the case of Catalonia as evidence that an economic system without private property could solve the coordination problem.


Anonymous said...

Forgive my ignorance, but when would a minimum wage increase employment?

Anna said... This is the study by Card and Krueger. They demonstrated that the short-term effects of minimum wage increases are not as harmful as originally thought, but it was later shown by Meer and West that the long-term effects are.

Anonymous said...

Caplan wrote about the Catalonian episode here:

David Friedman said...

I don't think Card and Kruger "demonstrated" anything, they merely failed to find a harmful effect. What was interesting about the article was the theoretical part.

Suppose you are a monopsonistic employer of unskilled labor — no competitors in bidding for their services. Start with the competitive solution — you are paying the marginal product of your employees, say ten dollars/hr to 110 employees. If you cut the wage to nine dollars/hour you can only get 100 employees so you are losing money on the ten you are no longer hiring, but you are saving a dollar an hour each on the ones you are still hiring, so your profit is probably higher. It's the same argument that shows that a monopoly sells at a price above marginal cost, only applied to a monopsony instead of a monopoly.

Now a ten dollar/hr minimum wage is imposed. Since you have to pay ten dollars an hour you do, and hire ten more workers. Wages have gone up and employment has gone up.

It's logically possible, but I wouldn't expect to see a monopsony, since unskilled labor is not specialized so can do lots of different things, and in a modern society the individual can find jobs over a considerable area.

A Country Farmer said...

At minimum, the burden of proof would be on the ancoms to provide an initial set of evidence for their claim about the Catalonians.

David Friedman said...

I posted a link by Bryan Caplan's piece on FB, and someone posted a link to a rebuttal:

I have not yet read it, but I think the existence of two long essays arguing opposite interpretations of the history, of which at least the first and perhaps both are convincing, supports the argument of my post, the difficulty of reaching firm conclusions on such issues unless one has put a lot of time and effort into making sense of arguments and counter arguments, which very nearly nobody has done.

Anonymous said...

I want to push back against your thesis a bit.
I think empirical examples provide great opportunity to test feasibility of a given theory.
For example, you often bring up saga period Iceland, and it's very informative.

Obviously, if the sides are not familiar with the facts, this is useless, but I'd be very interested in a debate, or an adversarial collaboration (SSC style) about the facts and their interpretation in the Catalonian episode.

My takeaway from watching this and other debates is that it's important to define the scope of the debate in advance, make it as narrow as possible, and ask the participants to prepare (maybe by exchanging some reading material prior to the debate).

Leonid said...

It may sound recursive but there are historical precedents against prioritizing persuasive logical arguments over historical precedents. Consider the Flour War, as an example.

In France, especially from the times of Colbert, the government imposed strict regulations on all areas of economic activity. By the mid-18th century the negative impact of this policy became apparent to many intelligent people and one of them, Jacques Turgot, eventually became the head of the government.

One of the first reforms implemented by Turgot was establishing free trade in grains. Traditionally Ancien Regime aimed to secure cheap food supply for the populace by “restraining the greed of profiteers” (i.e. heavily controlling producers and sellers) which predictably created a strong negative incentive against expanding production. The reformers presented a strong theoretical case for free trade, but just a few months after the reform’s implementation mass riots by hungry people spread around the country. The reform had to be cancelled, the reformer’s cause suffered a huge PR hit and one year later Turgot fell from power.

The economic theories of the reformers were far more sensible than those of their opponents, but their practical implementation still resulted in failure due to an unanticipated “technicality”. Normally when the price of the product rises the demand falls while the production or import increases, so that eventually a new equilibrium between price and demand is established. However, for grain which was the main source of calories in the 18th century France the demand was largely inelastic. Given the poor transport system the imports were also not very elastic (at least on the time-scale of a few months). So, when the harvest during the year of reform proved significantly below average, this created an opportunity for “profiteers” to increase their profits by hoarding flour and waiting until starvation would force the people to pay any price they can afford.

Naturally, there are also cases when incorrect interpretation of history resulted in even worse blunders, so there may be no clear advantage to theory or history in policy decisions. An economics expert like you might find it easier to spot false claims about Anarchism by analyzing theoretical arguments. Most other people are more likely to reach the right conclusions by reading historical accounts like Orwell’s “Homage to Catalonia”.

David Friedman said...

"Most other people are more likely to reach the right conclusions by reading historical accounts like Orwell’s “Homage to Catalonia”."

I don't think so. Lots of people reached incorrect conclusions about communism by reading reports on the early USSR written by sympathizers or about Victorian England by believing Dickens. Orwell was an unusually intelligent and honest man, but if you read him writing about something you actually know a good deal about you discover that he can be confidently and persuasively wrong.

The example I'm thinking of is Orwell's essay on Kipling, where he says some perceptive things but also asserts that _The Light that Failed_ was Kipling's only novel, thus ignoring one really first rate novel (_Kim_) and one pretty good young adult novel (_Captains Courageous_). For detailed criticism, see:

And there is the wonderful exchange between Orwell and the man who not only wrote, under one name, all of the stories that Orwell insisted had to have been written by multiple authors because there were so many of them, but all of the stories written under another name and others as well. _

Although you could argue that, in that case, Orwell was arguing by theory (nobody could write that much) not first hand observation. How could he know that he was referring to the author who the Guinness book of world records would list as the most prolific in history, having produced the equivalent of about 1200 novels.

Leonid said...

“Lots of people reached incorrect conclusions about communism by reading reports on the early USSR written by sympathizers”

I believe the main reason these people reached incorrect conclusions was emotional bias. There was plenty of information coming from the USSR that contradicted these reports. The mere fact that people trying to escape the communist paradize were given official prison sentences could have been a sufficient clue to what really was going on. But people like Orwell who since a very young age felt a lot of antagonism towards the bourgeois society generally ignored any information that did not fit what they wanted to believe.

Whether this information came in the form of factual reports or theoretical arguments did not matter. I suspect that if Orwell had a chance to observe a debate on theoretical feasibility of communism between some marxist professor and Ayn Rand (let alone any mainstream conservative of his period) then he would have been almost certainly persuaded by the marxist arguments.

In practice, a strongly biased person cannot be persuaded neither by data (“history”) nor by logical reasoning (“theory”). However, for a person who is genuinely interested to know whether Communism, Anarchism or any other hypothetical society is feasible, history is probably a more reliable guide than theoretical reasoning. Though all historical accounts tend to be biased, it usually takes little effort to detect author’s agendas and by cross-examining mutually hostile sources one can more or less reliably establish the major facts.

For example, there might be little reliable information in most Francista critical descriptions of Catalonian anarchists. By contrast, in Orwell’s “Homage to Catalonia” one can easily discern dystopia-in-progress despite the author’s obvious bias towards the anarchists. In fact, it’s precisely Orwell’s partiality that makes his account of the events so valuable as evidence against the anarchists.

Unlike retrospective historical analysis, reliably simulating a novel social formation in one’s mind using only abstract reasoning is probably an impossible task. I believe it’s similar in most sciences. Correct theories of complex phenomena are usually produced by fitting models to the existing observations, while most theories developed based purely on abstract reasoning are eventually falsified. Occasionally someone like Einstein with his General Relativity might become an exception but most people are not Einsteins.

P.S. Thank you for the links. To be honest, I also failed to appreciate Kipling’s writing. Perhaps it’s hard to do so without growing up in an english-speaking culture.

Anonymous said...

That was certainly an interesting debate, although it tended to veer off into all kinds of directions. I hope there is some sort of continuation to this in the form of a new debate, where you are able to drill more deeply into some of these issues.

I would be very interested to see more debate that is confined strictly to the economic theory of both RL and LL. Also, the nature of coercion seemed to raise up some very interesting view points.

ASG said...

I have little sympathy for AnComs but I do find the explanation of why Catalonian Anarchists didn't succeed somewhat underwhelming. It mirrors the typical excuses for socialist regimes -- namely there was nothing wrong with the system but it just so happens that skilled saboteurs, wicked wreckers and so on fatally undermined it. Even if true these excuses still seem unattractive as successful systems will be robust systems and systems which aren't robust are likely not good systems.

I'm late to this, sorry.