In my last blog post here at Mises CA, I picked up on a dispute over interpersonal utility comparisons that involved some prominent free-market economists. David R. Henderson (whose side I had taken) linked to my post with approval at EconLog, and added comments of his own. From the comments at my Mises CA post and especially at David’s, I realized that many people did not appreciate that David and I were talking about “utility” as it is used in modern price theory. This is not the same thing as “happiness” or “pleasure” if those terms refer to physical sensations or psychological states. In the present post I’ll explain the distinction.
Utility vs. Happiness
The critics kept asking Henderson (and me) variants of, “Are you telling me you don’t think a homeless guy gets more happiness from a blanket than a millionaire would?” But in terms of the debate over interpersonal utility comparisons, such a question misses the point.
In modern (meaning since the early 20th century) price theory–both in mainstream neoclassical economics as well as in Austrian theory in the hands of Mises–“utility” is simply a concept tied up with rational choice. To say that Option A gave an individual more utility than Option B really doesn’t mean anything else except that the individual preferred A to B.
For example, suppose Jim and Mary are at a restaurant and must pick between two flavors of ice cream for dessert. Jim chooses vanilla while Mary chooses chocolate. We can say that vanilla gave Jim more utility than chocolate did, while chocolate gave Mary more utility than vanilla did. It is utter nonsense to ask, “What percentage more utility did the vanilla give Jim than the chocolate?” and it is supremely utter nonsense to ask, “Did Jim get more utility from the vanilla than Mary got from the chocolate?”
Now it’s true, there is presumably a lot of biochemical activity associated with Jim and Mary’s choices. Neuroscientists might have many things to tell us about the same situation. But as economists, if we are interpreting their answers to the waiter as voluntary, purposeful behavior (“rational choice” in a neoclassical framework, or “human action” in Mises’ framework), the idea of utility is bound up with the choice itself. We don’t care about the physiological facts that might be involved in an explanation of each person’s preferences. Maybe a scientist can measure the endorphins in Jim versus Mary when each takes the first bite of the dessert, but that doesn’t have anything to do with “utility” as it is used in modern price theory.
Let me try another way to illustrate the distinction: Someone might choose to go to the gym and lift heavy weights, rather than sit on the couch eating pizza. Thus the lifting of the weights gave more utility, even though it was physically painful and very unpleasant per se. Or, a person might choose martyrdom over renouncing her religious or political views. Again, this embrace of death gives more utility to the martyr, but it doesn’t convey happiness or pleasure in any hedonistic sense.
Utility: What Is It Good For?
Confronted with the above explanation, some critics might then wonder: What the heck does this concept of “utility” do for us, if it’s merely a tautology telling us no more information than what the person chose?
The answer is that modern economics–ever since the 1871 revolution in thought–doesn’t just use utility, but specifically it focuses on marginal utility. This was the crucial theoretical advance–along with the derivative concept of diminishing marginal utility–that allows economists to easily explain how prices are formed in a market economy.
Let’s return to my simple example of a shopper seeing a sale of soda bottles for $1 each. Without the idea of marginal utility, it’s surprisingly difficult to explain why the shopper would put two, but only two, bottles into her cart. After all, if trades in an economy are governed by an “equality of values,” or if they’re governed by how much labor effort is congealed in a good, then why doesn’t the shopper fill up her cart with soda until she runs out of money? Looking at it the other way, if the shopper doesn’t think the 3rd soda bottle is worth putting in her cart, why did she have a different opinion regarding the 1st and 2nd bottles?
It’s only with the concept of marginal utility that we focus on the individual bottle and the individual dollar bill, and it’s only with the concept of diminishing marginal utility that we can make sense of the shopper’s behavior. We can then build up more complicated things like the individual’s demand curve and the market demand curve. In this approach, “utility” doesn’t refer to the physiological or psychological motivations for buying soda. For all we know, a person could be buying the bottles of soda for decoration, or for target practice. To say the shopper got more utility from the 2nd bottle of soda than from the 19th dollar bill really doesn’t mean anything deeper than, “The shopper voluntarily chose to buy the 2nd bottle at a price of $1.”
By itself, these concepts don’t give us a a quantitative theory that predicts how many bottles of soda a particular person will buy next Tuesday. Instead, we have a mental framework for organizing our observations and making sense of market phenomena.
Does Diminishing Marginal Utility Imply Cardinality?
Before leaving this topic, I should clarify one ambiguity. Some critics questioned my use of diminishing marginal utility, thinking that this involved a second derivative of the utility function and hence implied cardinal units of utility. Here there is a definite divergence between the neoclassical approach and the Austrian approach.
In standard neoclassical theory, it is true that diminishing marginal utility has been discarded. John Hicks, in Value and Capital, gave an authoritative treatment of the subject. Specifically, if we are going to discard the older notion of utility as units of some psychic substance that is being maximized, then (Hicks argued) we can’t have DMU, because that smacks of saying “marginal utility” is the increase in total utility from one more unit of a good, and thus DMU means that marginal utility goes down with further units. In other words, Hicks argued, DMU might sound like a negative second derivative, and that makes no sense if utility is ordinal. Hicks said that the solution is to use instead the notion Diminishing Marginal Rate of Substitution (DMRS).
However, in Austrian economics, we don’t run into this problem because “marginal utility” does NOT mean “the change in total utility that comes from one more unit of the good.” Instead, in Austrian economics “marginal utility” means “the utility of the marginal unit.” To solve the so-called water-diamond paradox, we just need to realize that successive gallons of water are devoted to less and less important ends. The first gallon of water is used (let us suppose) to quench thirst, the second gallon for cooking, the third through 10th for bathing, and the 1000th gallon and subsequent ones are used for washing the car. The reason a gallon of water is so inexpensive, then, is that on the margin it is devoted to a relatively unimportant end; the marginal utility of the 1000th gallon is much lower than the marginal utility of the 1st gallon of water. This statement doesn’t imply cardinal units of utility, any more than saying “a bowl of vanilla ice cream gives Mary less marginal utility than a bowl of chocolate ice cream.”