My teaching is mostly aimed at enabling students to conduct empirical projects on their own. I believe that three components are important for reaching this goal:

  1. A (basic) understanding of the available statistical and econometric methods
  2. An appreciation the assumptions necessary for results to admit a causal interpretation
  3. The ability to harness the power of computers for managing projects

The first component is fairly standard in most programs; the second often comes far too short or a one-sided view is given. The third becomes important when projects gain complexity, say at a level of a MSc thesis: Research in computer science has shown that very different techniques are needed for making code manageable when it consists of several hundreds (or thousands) of lines rather than a few dozens. Economics teaching usually stops at the latter, say in examples accompanying applied econometrics classes.

Below are some descriptions of the courses in my current portfolio, it should be fairly obvious to which of the three components for a successful research project each of them caters, unless it is not a methods course (health economics, for example, focuses on economic content).

Effective programming practices for economists (PhD, MSc, next: Winter 2018/19)

Many economists spend much of their lives in front of a computer, analysing data or simulating economic models. Surprisingly few of them have ever been taught how to do this well. Class exposure to programming languages is most often limited to mastering {Stata, Matlab, EViews, …} just well enough in order to perform simple tasks like running a basic regression. However, these skills do not scale up in a straightforward manner to handle complex projects such as a master’s thesis, a research paper, or typical work in government or private business settings. As a result, economists spend their time wrestling with software, instead of doing work, but have no idea how reliable or efficient their programs are.

This course is designed to help fill in this gap. It is aimed at PhD students who expect to write their theses in a field that requires modest to heavy use of computations. MSc students expecting to write on a similar topic are invited to join as well. Examples include applied microeconomics, econometrics, macroeconomics, computational economics — any field that either involves real-world data; or that does not generally lead to models with simple closed-form solutions.

We will introduce students to programming methods that will substantially reduce their time spent programming while at the same time making their programs more dependable and their results reproducible without extra effort. The course draws extensively on some simple techniques that are the backbone of modern software development, which most economists are simply not aware of. It shows the usefulness of these techniques for a wide variety of economic and econometric applications by means of hands-on examples.

You may find the slides here, I will update them irregularly over the course of the semester. Also see the page on Computing resources.


Erik Sørensen got me interested in these topics over many conversations we had while sharing an office in Amsterdam. The course is deeply influenced by Greg Wilson and early versions of the Software Carpentry project.

Microeconometrics (MSc)

In this course, students will acquire the basic skills for finding quantitatively meaningful answers to economic policy questions using data on many individual units (households, firms, countries, …). Specifically, we will look at the two main approaches in the literature:

  • Treatment effects - modelling changes in the policy environment as if they were experiments
  • Classical and structural approaches - modelling the behaviour of the units of observation

In terms of methodology, we will cover the basic theory behind and applications of regression discontinuity designs, instrumental variables with and without heterogeneous effects, GMM estimation, fixed-effect panel data and differences-in-differences estimators, maximum likelihood estimation, probit, multinomial choice models, and provide the basic idea of simulation-based inference in nonlinear models. Lectures will be accompanied by computer tutorials and programming assignments to be solved ahead of classes.

Health economics (BSc)

Students will learn how to use economic arguments in questions regarding individual health and the organisation of the health care system. In the course of this, they will also learn a number of empirical facts in these areas and how to interpret them in the light of economic models.

The first half of the course focusses on the production of health over the life-cycle: How is human health shaped during childhood and adolescence? How do people react to economic incentives later in life and make provisions for the future? What is the role played by socio-economic factors and demographics? What is the role of health care services?

The second half of the course, taught by myself, turn the last question around and consider the organisation of the health care system. Should health insurance be provided by private firms or public entities? What does the optimal insurance contract look like? Who should run hospitals? How should physicians be incentivised?

Throughout the course, we will start from basic empirical facts and then make sense of them using economic theory. Tutorials will consist of reading research papers; grading is based on a term paper.

The Economics of Social Insurance (PhD)

In most developed economies, the cost of social insurance schemes dominate public budgets and have huge implications for distributional outcomes and welfare. We will take the perspective of individual households and read empirical papers on the effects of unemployment insurance, disability insurance, health insurance, social security, etc.. We will emphasize their interactions with the labour market and with private insurance schemes, thinking about how to apply the mainly U.S.-based literature to European and other contexts.

Causality in economics and econometrics (PhD)

Policymakers must be able to predict the effects of potential interventions in order to reach their goals; academic economists have always been striving to provide them with such knowledge. Debates about what constitutes credible evidence for causal relations have been going on for just as long and they are particularly heated these days. We will start by reviewing some causality concepts and relevant debates in the literature, diving into fields as esoteric as the philosophy of science. Thus equipped with some background knowledge, we will discuss a number of empirical research papers. In doing so, we will focus on the plausibility of assumptions that permit drawing causal inferences rather than the statistical properties of the estimators involved.

Due to the interactive nature of the course, the slides are not suitable for being published here. In case you are interested in any, do not hesitate to email me.

Structural microeconom(etr)ic modelling (PhD)

We will focus on the estimation of parameters of dynamic partial-equilibrium economic models by the revealed preference paradigm. After discussing reasons for structural modelling, we read a number of papers that seek to explicitly estimate economic mechanisms. We will pay particular attention to preference heterogeneity and the formation of expectations. Depending on demand, we may also delve into practical implementation issues.