 Usage:  VaR = VaRest(y{,method}{,opt})  
 
 Input:

  y                      n x d matrix, the returns of d assets. 
                         
  method                 method for VaR, one of 
                         "RMA" (rectangular moving average), 
                         "EMA" (exponential moving average), 
                         "MAD" (mean absolute deviation), 
                         "EDF" (empirical distribution function). 
                         The default method is "RMA". 
                         
  opt                    optional, a list with optional input. The function 
                         "VaRopt" can be used to set up this parameter. 
                         The order of the list elements is not important. 
                         Parameters which are not given are replaced by 
                         defaults (see below). 
                         
  opt.h                  positive integer, window width. 
                         If not given, set to 250. 
                         
  opt.lam                positive scalar, parameter for EMA method. 
                         If not given, set to 0.96. 
                         
  opt.dist               integer, distribution. The default is 0 for 
                         normal, a positive integer denotes the degrees 
                         of freedom when using a t-distribution. 
                         
  opt.alpha              scalar in (0,1), significance level. The 
                         default is 0.01. 
                         
  opt.w                  scalar, 1 x d or n x d, weights for assets. 
                         If not given, set to 1. 
                         
  opt.bw                 scalar, bandwidth for method "KDQ" for 
                         quantiles from a kernel density estimator. 
                         If not given, chosen by Silverman's rule 
                         of thumb. 
                         
 Output:

  VaR                    (n-h) x 2 vector, the VaR for observations 
                         h+1 to n. 
                         
--------------------------------------------------------------
(C) MD*TECH Method and Data Technologies, 17.8.2000
