R/sar.R

In crseEventStudy: A Robust and Powerful Test of Abnormal Stock Returns in Long-Horizon Event Studies

##############################################################################################
## produces standardized abnormal return-series                                              #
## Dependencies: none                                                                        #
## as of October 30, 2018 (by Siegfried Köstlmeier, University of Regensburg, Germany)       #
## Disclaimer: Absolutely no warranty.                                                       #
##############################################################################################

sar <- function(event,          # data, containing event returns
                control,        # data, containing control returns
                logret = FALSE  # logical, TRUE if return series are log-returns
){# asr computes standardized abnormal returns

  if (missing(event) || missing(control)) {
    stop("Return series is missing.")
  }

  if(!is(logret,"logical")){
    stop("The variable 'logret' must be logical.")
  }

  if(!(length(event)==length(control))){
    stop("Return series have different length.")
  }

  if(isFALSE(logret)){
    event <- log(1+event)
    control <- log(1+control)
  }

  AR <- event - control
  SAR <- AR/sd(AR)

  return(SAR)
}