Course Scope

• Students will be acquainted with specific financial applications which require the use of econometric analysis.
• The students will learn how to use the programming language R.
• The students will learn how to determine whether an event of interest has an effect on the value of a company or an asset price.
• The students will learn how to specify and estimate probit and logit models.
• The students will learn to predict the probability of default of a borrower.
• They will learn how to simulate the returns of an investment strategy based on past historical data, as well as, to evaluate the forecasts generated from this strategy out of-sample.
• They will learn how to use the bootstrap technique to estimate the standard errors of the estimated parameter coefficients of a prediction model, to construct bootstrap confidence intervals without the use of a distribution assumption, and to conduct reliable statistical inference.
• They will learn how to estimate time-varying parameter coefficients of an asset pricing
• They will learn how to implement econometric tests of exuberance (explosive behavior) in asset prices (asset bubble detection).
• They will learn how to implement the variance ratio tests in order to evaluate the random walk hypothesis.
• They will learn to identify when we use a momentum strategy or a contrarian strategy based on the autocorrelation of the financial asset returns.

Course Outline

• Bootstrap: we show how to estimate the standard errors of the estimated parameter coefficients of a prediction model, to construct bootstrap confidence intervals without the use of a distribution assumption, and to conduct reliable statistical inference. The lectures focus on both non-parametric and parametric bootstrap methods.
• Backtesting: simulating the returns of an investment strategy based on past historical data and then evaluating the forecasts generated from this strategy out of-sample.
• Statistical test procedures to evaluate the out-of-sample forecasting performance of the prediction model. The lectures explain the nature of directional forecasting and its importance alongside conventional forecast evaluation procedures.
• Event study analysis: implementation of an econometric procedure in order to investigate the effect of an event on the stock price of a company.
• Binary Choice models: we show how to specify, estimate probit, logit and ordinal models. These models are used when the variable we want to predict is binary. We explain how to use a binary choice model to predict the probability of default of a company.
• Econometric tests of exuberance (explosive behavior) in asset prices: the objective of these tests is to detect periods in which the data generating process of a financial variable is characterized by explosive dynamics. We make use of these methods to investigate for the presence of bubbles in the stock markets, crypto-currencies, exchange rates, etc.
• Estimating models with time-varying parameters: students will code and learn methods such as rolling regressions, expanding regressions, Kalman Filters and Flexible Least squares.
• Variance Ratio tests: we make use of the random walk hypothesis to evaluate the weak efficient markets hypothesis, and consequently, whether markets are predictable or not. We apply the Variance Ratio tests to examine the random walk hypothesis. These lectures focus on various variance ratio tests: the power transformed tests of Chen and Deo (2006), the exact rank and sign-based Variance Ratio tests of Wright (2000), the subsampling method of Whang and Kim (2003), and the bootstrap test of Kim (2006).
•  Investment strategies: contrarian and momentum strategies.

Suggested Reading

Ηλίας Τζαβαλής, 2008. Οικονομετρία. Εκδόσεις ΟΠΑ.

JackJohnston, JohnDinardo. Οικονομετρικές μέθοδοι. Εκδόσεις Κλειδάριθμος.

Dimitrios Asteriou, Stephen G. Hall. Εφαρμοσμένη Οικονομετρία. Εκδόσεις ΠΡΟΠΟΜΠΟΣ.

JamesD. Hamilton. Time Series Analysis.  Princeton University Press.