How can one evaluate whether a government labour market programme such as the New Deal, or a subsidy to education such as the EMA is actually working? This course deals with the econometric and statistical tools that have been developed to estimate the causal impact on one or more outcomes of interest of any generic ‘treatment’ – from government programmes, policies or reforms, to the returns to education, the impact of unionism on wages, or of smoking on own and children’s health.
After highlighting the ‘evaluation problem’ and the challenges it poses to the analyst, we focus on the empirical methods to solve it: randomised social experiments, naive non-experimental estimator, natural experiments or instrumental variables, regression discontinuity design, econometric selection (or control function) models, regression analysis, matching methods, before-after and difference-in-differences methods.
For each of these approaches, we give the basic intuition, discuss the assumptions needed for its validity, highlight the question it answers and formally show identification of the parameter of interest. There will be plenty of discussion of the relative strengths and weaknesses of each approach, drawing from example applications in the literature. Each method will be implemented ‘hands-on’ in practical Stata sessions.
By the end of the course, participants will be able to:
- frame a variety of microeconometric problems into the evaluation framework, and be aware of the concomitant methodological and modelling issues;
- be discerning users of econometric output – able to interpret the results of applied work in the evaluation literature and to assess its strengths and limitations;
- access the evaluation literature to further deepen knowledge on their own;
- choose the appropriate evaluation method and strategy to estimate causal effects in different contexts; and
- use simple statistical packages (e.g. Stata) to implement the different evaluation methods to real data.
Level of knowledge required:
- This is a course on quantitative empirical methods for policy evaluation. As such, familiarity with basic statistical concepts (e.g. significance testing) and basic econometric tools like OLS regression and probit/logit models is required.
- The practical part of the course will make use of Stata; although the exercises will be guided, basic familiarity with this software is strongly recommended.
Please note that this is an intermediate-level course.
While offering an in-depth and thorough overview and discussion of the various evaluation methods, this is not an advanced course at the post-graduate level. Most emphasis is devoted to understanding the issues, to the choice of the most appropriate method for a given context and to the implementation of evaluation methods in practice. On the other hand, the course does rely on notation and there is a certain degree of formalisation (at the level of this paper), so please consider that PEPA also holds a much less formal introductory course aimed at those who design, commission or manage evaluation work. An Introduction to Impact Evaluation is scheduled once a year. For further information about this course please contact the PEPA Administrator.
Background reading: Though not at all required, some participants prefer to do some background reading in advance; for them, the following references can serve as overview/background:
Blundell, R., Dearden, L. and Sianesi, B. (2005), “Evaluating the Effect of Education: Models, Methods and Results from the National Child Development Survey”, (IFS Working Paper No. WP03/20).
Blundell, R. and Costa Dias, M. (2002), “Alternative Approaches to Evaluation in Empirical Microeconomics”, (Cemmap Working Paper No. CWP10/02).
A very thorough yet accessible book with a clear emphasis on causal effects is Angrist, J.D. and Pischke, J.S. (2009), Mostly Harmless Econometrics, Princeton University Press.