ME411 Factor Models in Time Series with Applications in Macroeconomics and Finance

The 2016 Methods Summer Programme is now closed. 
Details for 2017 programme will be available soon.

Large datasets are becoming increasingly available to researchers and practitioners in many disciplines. In particular, during this “big data” revolution the analysis of high–dimensional time series has become one of the most active subjects of modern statistical methodology with applications in various areas of social science including finance, macroeconomics, and econometrics. Although the value of information is unquestionable, the possibility of extracting meaningful and useful information out of this large amount of data is also of great importance. To this end, several new analytical and computational techniques have been developed under the name of factor models.

The aim of this course is to give an introduction to factor models in time series analysis by teaching students the basic analytical methods and their applications to macroeconomics and finance via the use of Matlab software. These models are widely used in central banks for forecasting key macroeconomic indicators such as GDP and inflation. They are also used to study the impact of economic policies on economic activity and in validating models of the economy. Financial institutions adopt factor models for risk management.

This course is designed for postgraduates, academics and professionals with an interest in big data analysis and who have some analytical background in time series analysis.

Course benefits
After successful completion of this course, participants should be able to:

• identify macroeconomic and/or financial policy problems that can benefit from factor analysis and consequently identify the appropriate dataset and methodology to be used
• extract and analyse relevant information from large datasets
• apply the analytical tools of time series analysis to the data using Matlab software
• conduct empirical research in time series, i.e. to interpret the information extracted from the data in a critical way also in relation to the existing literature
• forecast time series using many predictors.

Prerequisites
At least one semester of mathematical statistics with analytical treatment of estimation and inference, and at least one semester of multivariate calculus. Good background in methods of regression modelling and some basic familiarity with the analysis of multivariate time series.

The course consists of five daily lectures of three hours each, supported by four  afternoon computer-based practical classes lasting an hour and a half, which will allow course participants to implement the lecture material in Matlab.

The course covers the following topics:

• Motivations: the availability of large panels of time series and the value of information in macroeconomics, finance and other disciplines.
• Exact and approximate factor models: the curse and blessing of dimensionality. We start by discussing principal component analysis as a useful dimension reduction technique for large panels of time series. This is the most simple example of factor model (the static model) which we then generalize to include all temporal relations among the considered variables (the dynamic model).
• Estimation of factor models: we compare different models and discuss their estimation. The basic tools of multivariate time series analysis will be introduced. 

We then apply these models to three main areas:

• forecasting in real time of macroeconomics indicators such as Gross Domestic Product and Inflation
• policy analysis problems, i.e. the study of the dynamic reaction of observed variables to unexpected changes in policies such as monetary policies
• optimization of financial portfolios.

All topics are of particular relevance for and widely used by researchers in central banks and national or international institutions. Examples based on real-data applications and taken from existing papers are presented and discussed during lectures and some will be replicated during computer workshops.

Main texts
There is no main textbook for this course. Lecture notes and all necessary material will be provided. 

Software used
MATLAB. Some previous experience is desirable.

Teaching schedule
The following teaching schedule is indicative only, and is subject to change.

	Monday	Tuesday	Wednesday	Thursday	Friday
am Lecture	Introduction and principal component analysis	Static and dynamic factor models	Estimation and Kalman filter	Macroeconomics applications	Finance applications
pm	Computer class	Computer class	Computer class	Computer class	Exam

All lectures take place from 10am-1pm. Computer practical classes take place in the afternoon.

A 2-hour final examination will take place on the Friday afternoon.

"Matteo, your course was absolutely mind-blowing. Thank you!"
Participant on the 2014 Methods Summer Programme

"I would like to thank Matteo for some very inspiring and enlightening lectures. They were some of the best I have ever experienced."
Participant on the 2014 Methods Summer Programme

Please note: A full timetable will be provided at registration on Monday 22 August. The below timetable contains approximate hours only.

	M	Tu	W	Th	F
Morning session	3	3	3	3	3
Afternoon session	3	1.5	1.5		Exam