Research area: econometrics
Department: ORTEC Centre for Financial Research (OCFR)
Supervisor: dr. Hens Steehouwer
Description: Conventional stochastic scenario models that are used for ALM purposes are often based, or at least simulate, at an annual frequency because of the long horizons (say 20 years) of the application of the scenarios. During recent years however the need has come that these scenario models also provide scenarios at higher observation frequencies (say monthly) and possibly also with a shorter horizon (say annual). This need stems for the required consistent integration of long term ALM, medium term implementation and short term monitoring models and from the need to be able to analyze higher frequency policy actions (for example duration matching or rebalancing rules). To be able to model such higher frequency scenarios, it essential to first have a clear picture of the empirical behavior of financial and economic time series in this respect. Besides seasonal developments, one often thinks of volatilities and correlations that vary though time, for example based on more structural business or economic conditions of that moment (business cycle) or more “random” variation due to periods or turbulence on the financial markets. The objective of this project is to collect and study the (empirical) existing literature on the topic of time varying volatilities and correlations, to perform own empirical research based on a specific frequency domain methodology and finally to suggest ways of modeling the typically observed empirical behavior within the context of an existing scenario model.
Background information:
Steehouwer H. (2005), “Macroeconomic Scenarios and Reality. A Frequency Domain Approach for Analyzing Historical Time Series and Generating Scenarios for the Future”, PhD thesis, Free University of Amsterdam.