Remembering Hal White
Hal White, Chancellor’s Associates Distinguished Professor of Economics at UC San Diego, passed away on March 21, 2012, at the age of 61. In 1979, Hal moved from the University of Rochester to join the UC San Diego Department of Economics. He was an integral part of the department’s growth and development, and he stayed for his entire career. In his more than 30-year relationship with the campus, he helped transform the department into one of the top econometrics programs in the world. This was a mutually beneficial relationship for Hal, as it allowed him to escape from the blizzards of upstate New York. Hal was a true contender for the Nobel Memorial Prize in Economic Sciences, and he left an indelible mark on students and colleagues. Hal's legacy lives on through the "UC San Diego approach" to econometic analysis he helped to shape and his immense infleunce on the profession.
A True Scholar
Hal began his career studying labor-market issues and was confounded by the disparity between the assumptions he needed to make to use available methods to conduct empirical analysis, and the assumptions he felt were reasonable to make. His discomfort with the conventional methods of empirical analysis led him to search for a new approach. The outcome – his paper on how to construct standard errors for regression with heteroskedasticity – is by some counts the most widely cited paper in economics. Solving a problem in a research field outside his own led not only to one of the most important papers in the field of econometrics but also transformed Hal into one of the field’s leading researchers (he never did get back to labor economics).
Hal’s innovation in econometrics was not a fluke. He had a unique ability to approach new and existing problems in innovative ways. An example is the area of forecast evaluation. A key issue in forecast evaluation is whether we can distinguish if one model’s predictive performance improves upon that of a competing model. Often we observe sequences of forecasts from two competing models generated in such a way that the leaner model is a special case of a larger model. If we observe a very large sample of forecasts from both models, it will be impossible to distinguish between the two models under the assumption that the smaller model represents the true data generating process. This is because the larger model will eventually learn that its additional terms (those excluded by the smaller, true model) are redundant and so, in a large enough sample, will assign zero weight to such terms. This makes the two sets of forecasts identical, and so we cannot really distinguish between them.
In an ingenious paper co-authored with one of his students, Rafaella Giacomini, and published in Econometrica in 2006, Hal came up with a new solution to the forecast-evaluation problem. Rather than comparing the performance of two forecast models in large samples, he proposed comparing forecast methods, the distinction being that forecast methods include not only the underlying models but also a description of how they are implemented (e.g., how their parameters are estimated). Assuming that the parameters are estimated in a way so that estimation error would not vanish, Hal showed that a statistical procedure for conditional forecast performance could be established that would allow us to tell if one method is better or worse than another method because it produces better forecasts.
This is just one example of Hal’s ability to think outside the box. There are numerous other examples (see a previous article in Economics in Action). Abundant innovations help explain why Hal’s work is among the most heavily cited in all of economics, but he is not just known for his research in economics. Throughout his career, Hal managed to publish in an extraordinarily large range of top journals spanning econometrics, statistical analysis, finance, neural networks, medicine and legal studies.
Valued Mentor and Colleague
Hal was extraordinarily productive and accomplished, not only through his own research, but also through the generations of doctoral students whose committees he chaired and whose careers he helped launch. Many of Hal’s students have established their own illustrious careers in academia, finance and industry. His former students can be found at universities such as Yale, Cornell, Rutgers and Michigan State; Montreal and Boston universities; and University College, London University and National Taiwan University.
Hal was a towering academic, even among extremely accomplished colleagues. His near-encyclopedic knowledge of econometrics and statistics came in handy whenever a difficult question arose. He was more than happy to generously share his prolific knowledge. When presented with a difficult technical question at the regular brown-bag econometrics lunch meetings, Hal would regularly walk up to the blackboard, formalize a question into an econometric equation, and show us all how to solve it. As a co-author, he demonstrated the same ability and generosity of time and effort.
Not only did Hal have an amazing career in academia, he also co-founded the highly successful economic consulting firm Bates White with one of his former advisees, Charles Bates. To quote from the firm’s website, “The firm is distinguished by many of Dr. White’s qualities – most notably intellectual and analytical rigor and creativity. [Hal] was especially gratified that Bates White became quickly known for setting new quality standards in the economic and econometric analysis of legal disputes.”
Hal’s unique technical skills and enthusiasm for econometric questions never dampened over the years. To the end, he was involved in numerous econometric projects and remained a cherished co-author by those who were fortunate to work with and learn from him.
On a personal level, Hal was always cheerful and supportive. He shared his hobbies generously with friends and colleagues and would entertain with trumpet songs and big band music after dinner. For Hal, econometrics, trumpet playing, teaching and research were all part of enjoying life to the fullest.
To share memories of Hal, please visit our online memorial.