Search for content and authors
 

How many observations should be taken to obtain appropriate VaR measure using a family Sign RCA models?

Joanna Górka 

Uniwersytet Mikołaja Kopernika w Toruniu, WNEiZ, Katedra Ekonometrii i Statystyki (UMK), Gagarina 13A, Toruń 87-100, Poland

Abstract

Accurate modeling of risk is important in finance. A family Sign RCA models could be used to obtain Value-at-Risk (VaR) measure. In alternative approach to obtain VaR measure, different sizes of sample for rolling estimation of models parameters have been used. In the case when one of the Sign RCA models is used, one does not know what size of sample should be taken to obtain appropriate VaR forecasts. In this paper, to choose the size of sample, a Monte Carlo experiment  was used. The properties of the time series generated by one of the Sign RCA models were analyzed. The results obtained from Monte Carlo experiment and empirical examples were compared.

 

Auxiliary resources (full texts, presentations, posters, etc.)
  1. FULLTEXT: How many observations should be taken to obtain appropriate VaR measure using a family Sign RCA models?, PDF document, version 1.4, 0.2MB
 

Legal notice
  • Legal notice:
 

Presentation: Poster at First International Conference Quantitative Methods in Economics, Poster session, by Joanna Górka
See On-line Journal of First International Conference Quantitative Methods in Economics

Submitted: 2009-04-09 22:30
Revised:   2009-06-07 00:48