Key methods and aspects of empirical research, with an emphasis on applications in economics and finance. Typical issues associated with the application of (linear) regression models such as residual heteroscedasticity and autocorrelation will be covered. Special emphasis will be put on model specification and selection, the omitted variable bias, and endogeneity. Selected topics of time series analysis may also be covered. 

After passing this course participants will have learned to apply econometric methods based on a range of applications, mainly in economics and finance. Participants will have learned econometric theory on a level which is necessary to understand and conduct independent empirical research. They will know how to appropriately use selected econometric techniques depending on the research question.

I am convinced that learning requires hands-on experience. There are two main elements of the course to facilitate this kind of learning:

1. Participants have to replicate and extend an empirical paper published in a good academic (economics/finance) journal. Ideally, this paper is an important reference of their thesis. The paper has to be extended by adding data to the sample analyzed in the published paper. This implies that the choice of the paper must be made such that (a) the original dataset is available and (b) new/additional data can be collected. As a starting point, participants can inspect and choose papers from the Critical Finance Review on already published replications (http://cfr.ivo-welch.info/home/) or future replication possibilities (http://cfr.ivo-welch.info/home/replication-issues.html). One can also take a look at the special issue on replication in the Strategic Managment Journal (http://onlinelibrary.wiley.com/doi/10.1002/smj.2016.37.issue-11/issuetoc), or The Replication Network (https://replicationnetwork.com/).

• In case you have serious difficulties in finding a suitable paper (despite having tried hard), you can use one of the papers and datasets from the exercises or examples of the lecture notes ( Basic Financial Econometrics; files).
• If you find a suitable paper and dataset, but it is not possible to collect additional data, you can use the first and second half (or other segments) of the existing data to check the stability of results (e.g. fit the first half and predict the second half).
• I recommend choosing a paper which expands your horizon; i.e. do something new, something you are not (too) familiar with.
• In the first unit, you have to present a proposal of the paper you want to replicate, and what you are planning to do. During the course, you have to present an informal (but informative) report and results. Most importantly, you will have the chance to ask questions. The scope of topics and problems encountered in the replications will mainly determine the contents of this course.

2. Participants have to develop a prediction model for a dataset which will be made available after the first unit. Predictions are evaluated on the basis of a holdout dataset. Participants will get feedback whenever they submit a prediction. The submission of predictions, feedback and monitoring are done via a specific website (details will be provided in the first unit). In each unit, the progress made in developing the prediction model will be discussed. I will provide feedback and make recommendations.

Grades are mainly based on my (subjective - what else?) judgment of the participants' contributions during class discussions, of their presentations and the quality of their replication/extension study. These three aspects are equally weighted. Grades can be improved (but cannot get worse) on the basis of the performance of the prediction model.

Some experience in doing empirical research (which includes the ability to apply software such as R, EViews, SPSS, ...). 

Participants should read relevant sections from the lecture notes before each class. They should be well acquainted with the contents of the reviews. The reviews will not be repeated in class (except when specific questions are raised). Participants should review and replicate the examples from the lecture notes. The exercises provide a further opportunity to prepare for the course. Being able to replicate examples and to do the exercises is recommended before participants start to do the replication/extension of a published paper, and to develop the prediction model (see above).