Many economic and financial issues cannot be answered without analyzing time series. For example, it can be used to find factors that are useful in predicting stock returns, or to estimate the sensitivity of losses on bank loans to business cycle fluctuations. This thesis presents three essays in applied time series analysis, using a variety of methods to answer address several important issues in economics and finance. The first essay is on the cyclicality in losses on bank loans. A unique data set allows the joint modeling of losses and macroeconomic variables, while taking into account key characteristics of the distribution of losses. The second essay is on the impact of parameter instability on the allocation of the long-term investor. The instability is estimated, and the misspecification costs are assessed. The third essay is a forecasting exercise to see whether the relationship between economic uncertainty and macroeconomic output can be exploited in real-time such that it can be used as input for policy makers.

Time series analysis, credit risk, return predictability, forecasting
D.J.C. van Dijk (Dick) , H.J.W.G. Kole (Erik)
Erasmus University Rotterdam
978-90-361-0544-6
hdl.handle.net/1765/115135
Tinbergen Instituut Research Series
Department of Econometrics

Keijsers, B.J.L. (2019, February 22). Essays in Applied Time Series Analysis (No. 731). Tinbergen Instituut Research Series. Erasmus University Rotterdam. Retrieved from http://hdl.handle.net/1765/115135