27133 - Analysis of Economic and Financial Time Series

Learning outcomes

At the end of the course the student has the skills to define trading and management strategies based on time series business or financial information.

It is also able to choose critically the statistical method of analysis and prediction on the basis of both available data and operational environement. In particular:

- describe and forecast the dynamic of corporate business and financial price/revenue, through statistical models for time series

- assessing the quality of the available statistical information and relate it to the complexity of the modelling and forecasting methods

- select statistical tools for modelling/forecasting to develop business decisions in different functional business areas and/or considering different risk aversion.

- evaluate the results of alternative management strategies implied by the predictions process outcomes.

Course contents

Preliminary aspects

Stochastic processes and their property. Wold theorem. Linear operators, autocovariance and autocorrelation functions. Elementary stochastic processes examples. The representation of time series and the data quality analysis. Graphical analysis; identification and treatment of outliers and discontinuity

 

Time series m odelling.

Identification and estimation of stationary ARMA processes (also seasonal). The multivariate case: stationary VAR process.

Time series decomposition: Trends, Cycle, Seasonality. Stochastic vs. deterministic trend and seasonality.

Forecasting economic and financial phenomena.
Is it always possible to forecast? Cost functions (and risk aversion) and their role to define the optimal management/investing strategies. The “parsimony principle" for the choice of forecasting techniques.

From statistical models to management/investing decision
Extrapolative methods for trend, cycle and seasonality forecasting. Graphical methods for financial time series. Conjectural forecasting: Delphi and scenarios methods. Forecast encompassing and forecast combination.

Strategies evaluation
The operating cost: bias and accuracy measure. Statistical test for rank forecasting models performances.

 

Readings/Bibliography

F.X: Diebold "Elements of forecasting" South-Western  2001

On-line couse material  (in Italian)

 

Teaching methods

Teacher's lectures supplemented with smaller discussion sections (case study), tutorials, or laboratory experiment sessions.

Assessment methods

oral examination

Teaching tools

Software for time series analysis

Office hours

See the website of Andrea Guizzardi