Friday, March 18, 2005

Korea University: Daniel McFadden's Lecture

Korea University inaugurated its Nobel Laureate Lecture Series yesterday with Daniel McFadden's talk on "Economic Choices: Consumer Behavior and Economic Policy Analysis."

From the title, one might have expected McFadden to use his research and findings on consumer behavior to analyze economic policies currently being considered in the United States, for example, President Bush's proposed privatization of the social security system. In fact, he does mention this specific issue, but only in very general terms and only in his concluding remarks. Perhaps the title of his lecture should have been slightly different: "Economic Choices: Consumer Behavior and its Implications for Economic Policy."

The greater part of the lecture, however, analyzed the economist's standard model of consumer behavior. According to this model, there are two 'laws' of economics, and I reproduce these here from the lecture handout that McFadden provided for his audience's convenience:

1. "Individuals are consistently rational in their pursuit of self-interest" (p. 6).

2. "Rational consumers gain more from freedom of choice -- the more alternatives, the better" (p. 7).

The term "rational" is key here, so we have to know how economists understand it. McFadden presents three aspects:

1. "Preference-Rationality: Preferences are innate and stable, and do not flicker with experience" (p. 8).

2. "Perception-Rationality: Perceptions are formed consistently using statistical principles to process information and integrate it with experience" (p. 8).

3. "Process-Rationality: Given budget constraints and perceptions of alternatives, choices maximize preferences" (p. 8).

Dumbed down, this means that we know what we want, we know it when we see it, and we know how to get it.

Broadly formulated, this standard model sounds plausible -- though I found myself balking at the statement that the average consumer consistently uses statistical principles in processing information. Strictly speaking, I doubt that many of us do use statistical principles in this way. Isn't this a bit like saying that falling objects use the inverse-square law to guide their fall? Doesn't it make more sense to say that consumers' choices can be analyzed using statistical principles? Perhaps I should have inquired about this (or some of you economists can explain it to me).

Anyway, the model's plausibility proves trustworthy in the empirical realm:

"The economist's standard model of decision behavior is predictive and useful for a broad spectrum of choice behavior" (p. 50).

In short, it generally works.

McFadden used it to develop his own multinomial logit model, a mathematical formula that assigns preference values to discrete choices for calculating the probability of each choice (cf. discrete sets). This application works, too, and was used to correctly predict the number of commuters who would travel on the Bay Area Rapid Transit rail system, also known as BART.

Okay, this is quickly moving beyond my ability to follow -- as the links above will show -- but the point is that the standard model of rational economic behavior works in the real world.

But not always, for "people do not always choose wisely" (p. 19). In a section rather grandly titled "The Architecture of Cognitive Anomalies," McFadden notes that:

1. "Memory is fallible and selective, and recall is influenced by current context" (p. 23).

2. "Availability of information influences perceptions" (p. 23).

3. "Construal, affect, saliency, rules drive behavior" (p. 23).

To dumb down again, this means that we sometimes don't know what we want, don't know it when we see it, and don't know how to get it. We just think that we know.

Incidently, at this point, the standard model often slips from a descriptive rule to a prescriptive one:

"Dammit! Those consumers ought to be more rational!"

McFadden makes a move that bears at least a family resemblance to this, advising:

"Consumers who are aware of their own fallibilities and misperceptions have a better chance of detecting manipulation and avoiding choices they will later regret" (p. 50).

Correct, but I'll bet that McFadden himself wouldn't predict that large numbers of consumers will choose to adopt his advice. During the question-and-answer session following the talk, I had this in mind and asked him:

"Could you tell us any personal anecdotes of how you have used insights gained from your own economic findings for making decisions in your own life?"

Refreshingly frank, McFadden acknowledged that most of his important decisions had been made without really knowing enough to make fully rational decisions.

Those decisions, nevertheless, led to a highly successful career capped by a Nobel Prize, so maybe there's something to be said for intuitive judgements of the sort described by Malcolm Gladwell in Blink: The Power of Thinking Without Thinking.

McFadden didn't delve into that issue but, instead, gave the sensible advice that before making important decisions, one ought, for example, to seek counsel from older people who have faced similar decisions and can draw on their wealth of experience to help.

Most of us won't follow this advice consistently, as McFadden recognizes, so he also had larger advice, formulated in terms of "Implications for Economic Policy":

1. "The broadly rational pursuit of self-interest by consumers justifies in most cases the use of free markets for resource allocation and the principle 'He who governs best, governs least'" (p. 51).

2. "Systematic consumer mistakes in assessing risk and handling intertemporal choice justify public intervention in these areas to educate, regulate markets, and provide a social 'safety net'" (p. 51).

In short, a relatively free market but with some governmental intervention. Thus on the current debate over privatizing social security, McFadden counsels that:

"Unless private retirement accounts are very carefully designed to overcome the deficiencies in consumer rationality for risky intertemporal choice, they will fail to provide an adequate safety net" (p. 52).

Ignorant as I am, I wouldn't presume to risk a judgement on the Bush administration's privatization plan. I suppose that I generally agree with McFadden that government does have a role to play.

The question is how large a role. In the process of maturing and becoming wiser (which includes but is not limited to "more rational"), failure can play an important role. People do make poor decisions, and for some of the reasons noted in the talk. They also are not likely to listen to the advice of economists on how to become more rational. McFadden therefore turns to government as a means of protecting people from the mistakes they make (or would make). The danger here is that because people's poor choices go 'unpunished,' such protection can increase the irrationality of their future decisions because they have too little incentive to learn from (or even take note of) their mistakes. I suspect that the cost of protecting individuals would grow over time, both within the lives of individuals and across generations.

Any thoughts?


Post a Comment

<< Home