30604 - FORECASTING ECONOMIC, BUSINESS AND FINANCIAL DATA. CODING AND APPLICATIONS

It is recommended that students have taken courses on calculus, linear algebra and statistics.

Forecasting is important. Forecasts are routinely made in business, finance, economics and government and they guide many important decisions. In this course, we discuss rigorous and largely-quantitative statistical/econometric methods to produce and evaluate forecasts. All topics are presented in theory and in practice: classroom lectures are mixed with computer lab sessions where applications in micro and macroeconomics are developed using Eviews, an econometrics software.

Forecasting - universal considerations

Regression - curve fitting vs probability model

Regression - violation of BLUE assumptions

Forecasting in a regression framework

Causality in a regression framework

Time series models - trend and seasonality

Time series models - autoregressions and moving averages

Forecasting a time series

Forecast evaluation  and combination

• Being familiar with the concept of regression
• Understand the role of various assumptions and the impact of their violations.
• Being able to apply the theory in concrete cases
• Being able to communicate the output of the forecasting exercise.
• Apply the techniques learned to forecast a time series.
• Being able to evaluate various models and forecasts.

Use a econometric/statistic software to:

• produce descriptive statistics, graphs and other preliminary empirical analyses with any type of data
• use the econometric technique deemed appropriate for the forecasting problem at hand
• estimate the models and build the appropriate forecasts
• evaluate different forecasting models
• write a report based on the outputs of the previous analyses

• Face-to-face lectures
• Exercises (exercises, database, software etc.)

Econometric evaluations must be performed using specific softwares.

In this course, students will learn how to use an econometric software (EViews) and will use it to do empirical analyses.

Eviews is composed by two parts: a menu with windows, useful to do basic operations and a programming menu, where more advanced operations can be developed.

After having introduced and presented the part with windows, we will study the programming menu.

Here, students will learn how to structure a code.

The software will be presented in sessions in the computer lab, where the basic commands will first be introduced.

Once the basic language will be understood, students will be faced with real data and replicate applied works.

At the end of the course, students will be ready to use the software to carry on their own research.

The exams, midterm and general, will assess the knowledge of all aspects of the Intended Learning Outcomes, from the role of the assumptions in a regression, to the ability to communicate the output of the forecasting exercise, to the ability to discuss the pros and cons of different modelling  approaches.

Given that the topics taught in the course will be used in practical situations, questions will be both theoretical and applied.

Students must show to have understood the role of various assumptions and what happen if assumptions are violated.

Students must also show to have understood how different data require different treatment and econometric models in order to produce reasonable forecasts.

Open answer questions will evaluate students’ ability to articulate economic reasoning and to evaluate and communicate the output of a forecasting exercise.

Closed answers questions will in general be used to assess whether students have grasped the theory underlying different models.

The exam is structured in a midterm and a final exam. 

The first midterm is worth 50% of the final grade.

The second midterm is worth 50% of the final grade.

If the grade in the midterm is less than 18, a general exam must be undertook (value 100% of the final grade).

• Diebold, F. X. (2017), "Forecasting in Economics, Business, Finance and Beyond", downloadable from Diebold's webpage (https://www.sas.upenn.edu/~fdiebold/Textbooks.html).

• Diebold, F. X. (2019), "Econometric Data Science: A Predictive Modeling Approach", downloadable from Diebold's webpage (https://www.sas.upenn.edu/~fdiebold/Textbooks.html).

• EViews Manual (online).

• Relevant chapters and additional material will be discussed in class.