Econometrics for PhDs

Making sense of data

correlation-300x120


What is econometrics?

Who doesn’t use data in daily life to justify one or more hypothesis? We talk effectiveness of medications, cost of driving a car, measures of behavior, returns of our savings, variance of currency exchange rates. This course is about sorting the wheat from the chaff, the factoids from the actual facts that the data reveals.

In particular, this course will be focused on finding causal relationships either from observational or experimental data. We will start by observing that most empirical questions can be framed as “what is the (causal) impact of X on Y?”. In most situations, a correlation between X and Y is not indicative of a causal relationship between X and Y. For instance, if we see, across PhD classes, a positive correlation between your achievement and the achievement of your peers, is it: (1) a suggestion that your peers have an impact on your achievement? (2) a suggestion that you have an impact on your peers achievement? (3) a suggestion that some professors are better than others? (hence the group level correlation in achievement) or (4) a suggestion that students sort into classes by ability, either because of selection at enrollment or self-selection.

This central question of econometrics, the question of causality, is our approach to the standard tools of linear least squares, the Rubin causality model, panel data analysis, instrumental variable analysis, and natural experiments.


Who am I?

A professor at INSEAD since september 2008, I am doing research using both econometrics and economic theory. I have taught MBAs, EMBAs, LLMs, PhDs, and other programs, with more than 2000 students. My former students are now professors at universities such as the Ross School of Business at the University of Michigan, HKUST, Singapore Management University, the London School of Economics and other universities and business schools. My favorite econometrics language is… statistics, and I use either R, Stata, Mathematica, Fortran, for each specific purpose. My publications touch on policy-relevant topics such as education, urban economics, forecasting of firms' earnings.