The empirical exercises in this course are created to help you carry out the empirical analyses required for the assignments. The exercises provide step-by-step instructions on how to estimate the various models in OxMetrics, how to interpret the results, etc.

The software package OxMetrics is primarily created for econometric analyses of time series data. All models we consider in this course can easily be estimated using OxMetrics. In particular, we will work with time-series models using the PcGive module.

In the first exercise, you are asked to estimate a univariate ARMA model for Danish macroeconomic time series. The dataset includes several macroeconomic variables from the MONA database for the Danish economy. You can choose the time series you want to work with, but in the tutorial we focus on modeling Danish house prices given by the variable KP. Moreover, we encourage you to mainly focus on autoregressive models which can be estimated using the Single Equation Dynamic Modeling model class in PcGive.

The second exercise is optional. The exercise asks you to simulate various ARMA models for different parameter values. A simulation program is provided, so you can get an idea of how different parameter values result in (very) different time series processes.

Below, you can download a pdf-document with the empirical exercise.

Empirical Exercises 2.pdf.

To help you learn how to use OxMetrics, we have created a number of screencast videos which you can find below. We have also created an online tutorial that takes you through each of the questions in the exercise, shows you which results you should be able to get, comments on the results, etc.

Data and OxMetrics Program

Data files: TimeSeries.in7 and TimeSeries.bn7.

Ox program to simulate ARMA processes: generatearmadata.ox.

Videos

To help you learn how to use OxMetrics in practice, we have created screencast videos which illustrate how to do the various parts necessary to carry out empirical analyses and interpret the results. You can use the videos as a supplement to the empirical exercises or you can watch the videos and skip the empirical exercises if you feel that gives you sufficient understanding of how to carry out the empirical analyses in practice.

On the following pages, you can find more videos illustrating how to load, transform, and graph data and how to estimate linear regression models using the PcGive module.

Video: Estimation of Univariate Autoregressive (AR) Models in OxMetrics (9:59 minutes).
The video shows you how to estimate an autoregressive model in OxMetrics.

Video: Estimation of ARMA Models in OxMetrics (9:27 minutes).
The video shows how to estimate an ARMA model in OxMetrics using the ARFIMA module in PcGive.

Video: Forecasting Based on Univariate Autoregressive Models in OxMetrics (7:34 minutes).
The video shows how to make out-of-sample forecasts in OxMetrics.

Video: How to Get Forecasts When Including Dummy Variables in Autoregressive Models in OxMetrics (6:34 minutes).
If you include dummy or shift-dummy variables in your estimate model, OxMetrics will not report forecasts. But you can get around this using the trick explained in the video.

Online Tutorial

In the online tutorial, we try to guide you through each of the questions in the exercise by combining explanations, graphs, hints, feedback, and quizzes.

Get Started

Quick Menu: Overview Q1 Q2 Q3 Q4 Q5 Q6 Q7 Feedback