Ed Stringham and the Information Puzzle

I recently began writing about my interest in The Preference Elasticity of Demand. This article argues that the Austrian School of Economics is a logical theoretical home for the theory of preference elasticity. I also discuss an objection the preference elasticity called the information puzzle.

Ed Stringham is an Austrian economist and author of Private Governance: Creating Order in Economic and Social Life. He also wrote this article in The Quarterly Journal of Austrian Economics, Summer 2010, arguing “most neoclassical economists ignore changing preferences, they too quickly conclude that change is impossible. History shows that social change based on changes in preferences is common.”

Becker and Stigler establish the neoclassical method of the prohibition of preference changes as causal explanations of shifts in supply or demand as part of the paper De Gustibus Non Est Disputandum.

The Austrian notion of demand is teleological. An individual has a goal and consequently demands such things as will help accomplish the goal. What about after the goal is accomplished? In that case the person will have a change in demand. Either they will form a new goal or else they will enter a state of satiation, having no goals or demands at the moment.

The identification of a new goal seems to be a change in preference by definition. Secondly, the satisfaction of different ends seems to provide different levels of utility. Finally, different ends are optimally achieved through the execution of different activities including the use of different tools or intermediate goods and services. This constitutes a change in preferences.

There is another way preferences might change: The introduction of new information. This will later lead to the information puzzle. Suppose I would like chocolate ice cream. I go to the store and find they are out of it. They have blueberry and apple pie ice cream. I haven’t ever tried these flavors at this store, but going to another store is costly and my expected utility from these unknown flavors exceeds the price of going to another store. I take the gamble and order the blueberry.

This store has the worst blueberry ice cream I have ever tried. Holding the price constant, I prefer to consume a far less quantity over time going forward. If preference is defined as a utility curve or an expected utility curve, that curve undergoes a shift, constituting a change in preference.

Or did it? Did my preferences actually shift, or did I simply gain new information? If I had been perfectly informed prior I seem unlikely to have made the same choice.

The information puzzle is a statistical puzzle. It says that in order to measure the effect of a preference change we need to hold everything else constant. Not only prices or even relative prices, that’s not enough. We need to also hold the entire information state constant to prevent improper attribution of an information effect on preferences or vice versa.

I think the puzzle is easily resolved by saying that preferences and information are one in the same. At least in the case I am discussing, where the goal is held constant. In this case we are essentially discussing the neoclassical indifference curve.

Yes, the indifference curve is essentially a set of data, but so is the budget line and so are the prices. Their status as pieces of information does not preclude their usefulness or meaningfulness.

We need not disentangle the problem because there is no distinction and no problem. If I had been able to gain information without eating the ice cream I would still have simultaneously altered my preferences because preferences emerge from a combination of information and goals.

When information is conceived this way we need not hold it constant. Correcting for information would be to correct for the very thing we are interested in measuring, the change in preference which arises over time after eating the ice cream.

I think a better way to establish preference change is through a time-discontinuity model, a synthetic control model, or by an experimental approach. With an experimental approach in particular it is easy to hold prices constant and isolate the effect of a change in preference based on a particular experience in which an individual participates.

What about behavioral economics? I’m not a pro behavioral economist although I’m a big fan of what I’ve seen. What I’ve seen seems to be a large field of empirical facts, generally interpreted and extended on the basis of a neoclassical framework. This makes sense both because Kahneman was influenced by Thaler, and also because neoclassical economics is the main approach to economics, so any interdisciplinary attempt will likely run first into such an approach.

While this makes sense that behavioral has turned out this way, I am not convinced it is optimized this way nor should persist this way. I see no reason not to rebase behavioral economics onto a neo-Austrian framework. Perhaps not exactly as outlined by Hicks, but some kind of post-beckerian synthesis between neoclassical and Austrian economics.

Brief aside (or perhaps a directly related point): George Gilder has stated that “Creativity always comes as a surprise” and “Wealth is essentially knowledge.” I think he has begun an astounding post-Hayekian information-theoretic approach to wealth, and I think he is extremely underappreciated. I’ll try to extend his ideas a bit over time. Like knowledge needs application for example. So in this Gilder view, tasting the terrible blueberry ice cream actually made me richer.


1 thought on “Ed Stringham and the Information Puzzle”

Leave a Comment