[Originally published on Mises.ca]
Occasionally, when Austrians try to distinguish their brand of doing economics from the mainstream, they get hit with accusations that they are attacking straw men; that no one believes what Austrians claim is the mainstream approach.
Is this true? Are Austrians attacking enemies that don’t exist anymore? I say no. While it might be true that many of the top economists may in general agree with broad Austrian methodological conclusions, the typical economist is much more likely to either (a) explicitly deny the Austrian criticisms, or (b) implicitly or casually invoke these fallacies during their analyses for reasons I shall explain below. Let’s look at the evidence.
The position often attributed to Austrians regarding econometrics is that Austrians reject the field completely. But this is not true. Austrians criticize econometrics only when it is either (1) trying to prove or disprove (or “falsify”) pure economic theory, (2) trying to establish universal magnitudes between economic phenomena, or (3) trying to forecast economic data. To Austrians, econometrics is only useful as a tool of history: it can tell us quantitative information about a specific period in the past, and that’s it. Econometric findings are not generalizable to all of the past, and cannot be projected to precisely predict the future.
The second case was one of the principle ambitions of the first econometricians. In Human Action, Mises refers to the University of Chicago economist Henry Schultz. Schultz (a founding member of the Econometric Society) had tried to determine “the” price elasticities of demand for a whole bunch of goods. In other words, he wanted to find out exactly how much the how many more potatoes would be sold if their price per kilogram went up $1. Mises correctly demonstrated why this whole endeavour was doomed to failure: all prices are historical, and are subject to change at any time. There are no constants in economics.
Are econometricians still trying to carry on Schultz’s project? Not exactly. But here is how two leading econometricians, James Stock and Mark Watson, in the 3rd edition of their textbook Introduction to Econometrics, define their own field:
“Ask a half dozen economists what econometrics is and you could get a half dozen different answers. One might tell you that econometrics is the science of testing economic theories. A second might tell you that econometrics is the set of tools used for forecasting future values of economics variables…. Another might say that econometrics is the process of fitting mathematical economic models to real-world data…. In fact, all these answers are right.”
So econometrics hasn’t changed much in its scope from the time that Mises was criticizing it. Austrians are still correct to fault it on these grounds.
Indifference curves are a cornerstone of modern mainstream economic education. An indifference curve shows different combinations of bundles of goods to which a consumer is “indifferent.” For example, a consumer may be indifferent between one meal that consists of a burger with two patties and a small box of fries, and another meal that has a burger with one patty but a large box of fries. In other words, he values both the meals exactly the same.
Some Austrians criticize indifference curve analysis by pointing that no action takes place when people are indifferent. To Austrians, the only way to understand preference is through action. So when someone chooses combo A over combo B, they have demonstrated their preference for A. Therefore, it is impossible to demonstrateindifference in the market.
To be sure, many mainstream guys today would in principle agree with most of the above criticism of indifference curves. They would agree that people will not make transactions at indifference. But mainstreamers would say people would make transactions untilthey were indifferent. And the indifference curve represents all the possible combinations at which a consumer could be indifferent.
That’s all well and good. However, to come up with indifference curves, mainstreamers have to make some very specific assumptions that most people (and Austrians) would classify as unrealistic, to put it mildly. For example, all bundles of goods areinfinitely divisible, meaning it’s possible for a farmer to be indifferent between one bundle that contains 2 tractors and 2 horses and another bundle containing 1.732 tractors and 2.896 horses.
Another assumption for indifference curve analysis is that preferences are transitive. That means that if you preferred bundle A over bundle B, and bundle B over bundle C, then you must prefer bundle A over bundle C. This then allows them to make inferences about people’s preferences. For example, if you prefer X to Y, and Y to Z, the mainstream economist would you have revealed your preference of X to Z.
But to Austrians, preferences aren’t “revealed” in the mainstream sense–they can only be demonstrated through action. Austrians acknowledge that preferences can change over time, so that it’s possible for you to prefer A over B at time 1 and B over A at time 2.
Ordinal vs. cardinnal utility
Another big difference between Austrians and mainstreamers is how we understand the very concept of utility. Here, there are in fact two big differences between the Austrian and mainstream conceptions of utility.
First, Austrians vigorously insist that utility is ordinal, meaning that it represents an order. For example, first, second, third, etc. are ordinal numbers. For ordinal numbers, it only makes sense to order them from biggest to smallest, first to last. It doesn’t make any sense to say “first is twice as big as second”; for example, if my top 3 favourite movies were
1. Star Wars
2. Jurassic Park
it wouldn’t make any sense to then conclude I like Star Wars twice as much as Jurassic Park.
This is in contrast with a cardinal view of utility. Cardinal numbers are like 1, 2, 3, and so on. They can be counted and added and multiplied and divided up. Mainstreamers will either insist that utility is ordinal, or say that it doesn’t matter whether it is ordinal or cardinal. But the way mainstreamers use utility, it cannot be ordinal in the same sense as Austrians use it.
When mainstreamers draw several indifference curves, they will assign utility values to them. One might have a utility of 10, another of 20, and the third of 30. A person prefers more utility to less, so a mainstreamer would say that the indifference curve giving 30 “utils” will be preferred to the one giving 20, and the one giving 20 utils will be preferred to the indifference curve giving 10. But where do those numbers come from? And if one bundle of goods provides 30 utils of satisfaction, does that mean it provides 3 times as much satisfaction as a bundle that only provides 10 utils?
The numbers come from “utility functions”: mathematical formulas that define how much utility a person will get from a certain bundle of (infinitely divisible) goods. And while this may sound dubiously close to cardinal utility, a typical mainstreamer will say that these numbers don’t matter, since utility functions are “monotonic”. This means that if you perform the same mathematical operation on all of the functions, then the order of the functions will not change.
For example, if we multiplied all the utility functions by 10, then the utilities from two paragraphs ago would be 100, 200, and 300. This preserves their order, so that the indifference curve that used to be valued at 30 but is now valued at 300 is still the most preferred set.
But as Mises Canada’s own Garrett Petersen pointed out, just because the mainstreamers say that utility is ordinal, doesn’t mean it actually is. This is especially true when talking about expected utility, or the utility associated with lotteries with uncertainty. As Garrett demonstrates, it’s possible to transform an expected utility function in such a way that it isn’t monotonic. This is because expected utility functions sneak in cardinal utility, in an effort to determine “intensities” of preferences.
But cardinal utilities are not only nonsensical (what would you think of a person who asked you “how much more intensely do you prefer your mother to your father?”), they are also, in Garrett’s words, “demonstrably false.” The famous Allais paradox showed via experiment that real people tend to choose lotteries in an way that is inconsistent with the assumptions of expected utility theory.
The reason for this is because real people are not cold, calculating machines–unlike what the mainstream mathematical models want to assume.
The very definition of utility
The big disagreement between Austrians and mainstreamers on utility theory. The reason mainstreamers want to have cardinal numbers for their supposedly ordinal utility functions is simply because they want to do calculus with them. The reason they think this is ok to do is because they have a different understanding of the word utility.
But don’t take my word for it. Here’s economist Donald W. Katzner, from an essay defending the mainstream position on ordinal utility:
Indeed, utility values are not measures, in the theory-of-measurement sense, of anything. Rather the ordinal utility function is simply a numerical, differentiable representation of a preference ordering (that includes the possibility of indifference and) that has no relation to any underlying pleasure structure.
Compare this with Murray Rothbard’s definition of utility in Man, Economy, and State with Power and Market (pg. 18)
These scales of preference may be called happiness or welfare or utility or satisfaction or contentment. Which name we choose for value scales is not important. At any rate, it permits us to say, whenever an actor has attained a certain end, that he has increased his state of satisfaction, or his contentment, happiness, etc.
Assuming that “etc.” includes “pleasure” (and there’s no reason why it shouldn’t), there is a clear contrast between Rothbard’s assertion that by attaining a certain end, one increases their level of happiness or pleasure or whatever; and Katzner’s view that a preference ordering has “no relation to any underlying preference structutre.” Rothbard sees utility as a psychological phenomena, while Katzner (and mainstreamers generally) see utility as having “no intrinsic meaning other than the information [it provides] concerning preferences.”
Interpersonal utility comparisons
Another area that Austrians criticize the mainstream on is the area of interpersonal utility comparisons. In short, Austrians say this can’t be done. Value is subjective, and utility is ordinal. Just because I might have a preference ranking of, say,
- My favourite pair of shoes
- My favourite hat
- My favourite tie
- My favourite belt
And you have a ranking like so:
- Your favourite pair of shoes
- Your favourite tie
- Your favourite belt
- Your favourite hat
it doesn’t mean you like your tie more than I like my tie. I might be much more attached to my clothes than you are.
Mainstreamers, however, will occasionally pay lip service to this but will either sneakily work it in to their analysis, or, as even some libertarians are doing now, brazenly reject the idea altogether.
As we mentioned above, although some mainstreamers would say that even though utility functions pump out things like “the utility of shoes is 20 and the utility of hats is 40,” it doesn’t mean that hats have twice as much utility as shoes. They can only say that hats are preferred to shoes. And because individuals have their own unique utility functions, just because one guy’s utility function said he gets 38 utils for hats compared to another guy’s 40, it doesn’t mean that he values hats any less.
But sometimes they can throw in these interpersonal utility comparisons sneakily. For example, when they want to construct a “social welfare function,” to judge the effects of a new tax policy, for example. They just add up everyone’s utility functions in together, as if they were cardinal or something, and act as if they’ve done nothing wrong.
But a worrying trend is the rise of economists who are brazenly denying that interpersonal utility comparisons could be any sort of problem. Take for example this economist from UC Santa Barbara, who published some class notes that say that interpersonal utility comparisons are possible, since parents make choices for their children all the time:
“You can’t make interpersonal welfare comparisons.” What can we make of such statements? […] Anybody who has ever been a parent of more than one child has plenty of experience with making tradeoffs between the happiness of one person and another. Indeed, anyone who cares at all about the welfare of other human beings has to make some tradeoffs between different people’s happiness. If one’s preferences about the well being of others are transitive and complete, then one must have some kind of a utility function defined over the wellbeing of more than one person.
This is a basic confusion. An interpersonal utility comparison isn’t “a utility funciont defined over the well-being of more than one person.” It is comparing one person’s utility to another. Judging one person’s happiness to another’s. If utility is ordinal, as Austrians insist it is and many mainstreamers at least pretend that it could be, then comparing one person’s ranked preferences to another person’s ranking is nonsensical. Value is subjective.
Even more troubling is a spate of ostensibly free market or libertarian economists who have come out in favour of interpersonal utility comparisons. Here also is Bob Murphy calling out Dan Klein, Bryan Caplan, and Tyler Cowen & David Friedman for casually trying to make interpersonal utility comparisons.
In conclusion, Austrian criticisms of mainstream economics are as relevant as ever.