Daniel McFadden's strong analytical bent and fascination with human behavior have taken him on a decades-long journey from the University's physics department to the stage of Stockholm Concert Hall. That's where he accepted the 2000 Nobel Prize in economic sciences from Sweden's King Carl Gustav in a glittering awards ceremony last December.
McFadden (Physics '57, Behavioral Sciences Ph.D. '62) and University of Chicago economist James Heckman were honored for their contributions to microeconometrics, the branch of economics that weds economic theory to statistics. Both men developed theories and methods that are widely used by economists and other social scientists to analyze the economic behavior of individuals, households, and firms.
McFadden, currently the E. Morris Cox Professor of Economics at the University of California at Berkeley, describes his work much more succinctly: “I am basically an engineer who designs things. I'm intrigued by puzzles, and I'm a problem solver. In economics, I design
statistical tools that [get] used quite a lot."
Although he's a scientist by training, McFadden has always had an uncommon interest in social issues. In 1953, as a high school junior in his native North Carolina, he was suspended from school for organizing a petition drive supporting what he calls “a civil rights issue” concerning the right of students to leave campus during school hours.
The young McFadden was sent off to an uncle's dairy farm in Buffalo, Minnesota. Shortly thereafter, although he lacked a high school diploma, he entered the University at age 16 by passing an entrance examination.
As a physics undergraduate, McFadden worked in a laboratory headed by the late physics professor emeritus John Winckler, where he was involved in the design and construction of an x-ray telescope. McFadden regards Winckler as an important mentor.
"I was strongly influenced by [Winckler],” says McFadden. “I learned how to do research from him."
In an interview conducted shortly before his death, Winckler recalled that McFadden —“a typical undergraduate”—worked on high-altitude balloon research sponsored by the Office of Naval Research. Winckler described his former student as “a highly goal-oriented and motivated person” but said that he wouldn't necessarily have predicted McFadden would one day win the Nobel Prize.
During those years, McFadden augmented his physics courses with studies in psychology.
"One of my undergrad jobs was to program the card-sorters for the Minnesota Multiphasic Personality Inventory [MMPI],” he recalls. Developed at the University in the 1940s and later revised, the MMPI is the classic standardized psychology test, a diagnostic tool designed to uncover a wide range of mental disorders in adults.
"I got interested in how you could hope to measure people's personalities and [in] what people were thinking when they answered [the MMPI's unusual] test questions."
Working with the MMPI piqued McFadden's curiosity about the statistical measurement of human choices and aspirations. He decided to enroll in an especially rigorous, multidisciplinary doctoral program at the University that was funded by the Ford Foundation.
His graduate advisor in economics, Regents Professor Emeritus Leonid Hurwicz, recalls, “At the time we had a very special Ph.D. program in behavioral sciences. It offered broad connections with everything that has to do with human behavior. The student had to get a Ph.D. in a chosen field but also had to pass the equivalent of a master's exam in five other fields, like sociology, political science, statistics, or child development. The program really suited [McFadden's] personality. He was unusual in the breadth of his knowledge, but he was not a nerd. He was very interested in what was going on in the world."
According to Hurwicz, McFadden was one of only two economics students who completed the extremely challenging degree.
Regents Professor John Chipman of the economics department arrived at the University only a few years before McFadden became a graduate student and Chipman's research assistant.
"McFadden had immense integrity as well as brilliance,” says Chipman. “He was an all-around fine person and very obviously the star that year among all the graduate students."
McFadden says that Chipman and Hurwicz were among the handful of economists and social scientists at the University who “made me what I am."
After completing his doctorate in 1962, McFadden went to the University of Pittsburgh as a Mellon postdoctoral fellow. A year later, he joined the economics department at Berkeley, rising from the rank of assistant professor to associate professor with tenure in just three years. From 1979 to 1991, he was a member of the economics faculty at Massachusetts Institute of Technology. He then returned to UC-Berkeley, where he established the school's Econometrics Laboratory, a world-renowned leader in microeconometric research.
Microeconometrics uses large data sets of economic information about groups of individuals, households, and businesses. Fueled by the development of increasingly powerful computers, the scope of this so-called microdata has expanded rapidly in recent years, creating huge databases that encompass a great range of human activities.
However, studying data derived from individual economic decisions presents the researcher with some thorny issues, which McFadden's work addresses.
Economic decisions often take the form of a “discrete choice"-a decision made from a limited set of observable options. Economists would like to uncover the relationship of discrete choices to factors that economic theory can predict.
For example, a researcher studying the travel choices of commuters might compile
data sets that measure the relative importance of various factors
to their decisions—such as the cost of public transportation,
distance from work, or the availability of free parking.
But the researcher can't possibly observe all the factors that affect an individual's or a household's choices. Life has a way of injecting the unexpected into human decision making. As any commuter knows, a sudden phone call at the last minute, a missed bus, or a dead battery can lead to very different transportation decisions.
As a result, a sample might not be random and therefore not representative. Even when samples are representative, unobservable characteristics make it difficult to explain variable behavior.
One of McFadden's most significant accomplishments was
the development of a statistical method called “conditional
logit analysis.” He used it to create models that
could predict the share of a population that will choose
alternatives to average behavior. McFadden assumed that
these “random errors” have a specific statistical
distribution in the population.
Conditional logit analysis quantifies the role that unobservable factors play in decision making so that the impact of specific measurable variables can be measured accurately.
In the case of commuters, people may use public transportation more frequently if fares are decreased, but there's still a chance every day that each commuter, for unrelated reasons, will choose to drive. Conditional logit analysis will interpret the data in a way that reflects the total variety of actual behavior but permits accurate measurement of the relationship between lower fares and increased ridership.
Although McFadden's methods can be applied to many different economic problems, they've been widely adopted by researchers to analyze what he calls the “big life choices” of work and living habits. Yet he acknowledges that not all of life's decisions are amenable to statistical analysis.
"Where statistics work is where you have an historical record of people responding in analogous situations,” he says. “What [statistical analysis] does is to sort out the 'noise' from the data. But it doesn't work in a really unique situation where you can't find [useful data]."
McFadden himself has studied issues as diverse as residential energy demand, the design of the [San Francisco] Bay Area Rapid Transit system, the cost of the 1989 Exxon Valdez oil spill, and housing choices of the elderly.
But his methods aren't always applied to the most challenging purposes.
"These techniques are used to devise marketing [strategies] for new products. And they often do quite well, but where you need them the least is where they do the best.
"If you want to know whether to make the cereal box different, I can nail the answer,” he says. “But if you want to introduce a new kind of vehicle [or] discover what's the demand for electrical vehicles, for example, that's a lot harder to predict using these models."
In many ways, McFadden the Nobel laureate remains unchanged from the unassuming, hardworking student whom his former University professors remember.
His continuing fascination with people and their choices, which dates back to his undergraduate years, underlies many of his professional accomplishments. The Royal Swedish Academy of Sciences noted that a recurring theme in McFadden's research is “his ability to combine economic theory, statistical methods, and empirical applications, where his ultimate goal has often been a desire to resolve social problems."
In response to a question about the aftermath of winning the Nobel Prize, McFadden
says, “The best thing about winning the prize was [that] it allowed
me to get in contact with students I'd lost touch with, not to mention
a whole bunch of North Carolina cousins that I hadn't seen in years."
