×

Restricted VaR forecasts of economic time series with contemporaneous constraints. (English) Zbl 1225.62132

Summary: This paper presents a methodology to incorporate linear constraints into the forecasts of a multivariate time series when some components of the series are also subjected to contemporaneous constraints. This methodology has natural applications when forecasting macroeconomic and financial time series that must satisfy accounting constraints, which are binding. Besides, the methodology has immediate implications on the forecasts of cointegrated systems because the cointegration relationship can be viewed as a stochastic contemporaneous constraint. An illustrative empirical application considers the balance of payments situation where deficit (income minus expenditure) is the variable of interest. A Monte Carlo study is also presented to illustrate the use of the methodology to improve the forecasting performance of a vector time series model when there is a cointegration relationship.

MSC:

62M20 Inference from stochastic processes and prediction
62M10 Time series, auto-correlation, regression, etc. in statistics (GARCH)
91B84 Economic time series analysis
62F30 Parametric inference under constraints
91G70 Statistical methods; risk measures
65C05 Monte Carlo methods
PDFBibTeX XMLCite
Full Text: Link