LPM.SE {RPESE}

Standard Error Estimate for Lower Partial Moment (LPM) of Returns

Description

LPM.SE computes the standard error of the LPM of the returns.

Usage

LPM.SE(
  data,
  const = 0,
  order = 1,
  se.method = c("IFiid", "IFcor", "IFcorAdapt", "IFcorPW", "BOOTiid", "BOOTcor")[1:2],
  cleanOutliers = FALSE,
  fitting.method = c("Exponential", "Gamma")[1],
  d.GLM.EN = 5,
  freq.include = c("All", "Decimate", "Truncate")[1],
  freq.par = 0.5,
  corOut = c("none", "retCor", "retIFCor", "retIFCorPW")[1],
  return.coef = FALSE,
  ...
)

Arguments

data

Data of returns for one or multiple assets or portfolios.

const

Constant threshold.

order

Order for the lower partial moment (should be 1 or 2).

se.method

A character string indicating which method should be used to compute the standard error of the estimated standard deviation. One or a combination of: "IFiid" (default), "IFcor" (default), "IFcorPW", "IFcorAdapt", "BOOTiid" or "BOOTcor".

cleanOutliers

Boolean variable to indicate whether the pre-whitenning of the influence functions TS should be done through a robust filter. Default if FALSE.

fitting.method

Distribution used in the standard errors computation. Should be one of "Exponential" (default) or "Gamma".

d.GLM.EN

Order of the polynomial for the Exponential or Gamma fitting. Default polynomial order of 5.

freq.include

Frequency domain inclusion criteria. Must be one of "All" (default), "Decimate" or "Truncate."

freq.par

Percentage of the frequency used if "freq.include" is "Decimate" or "Truncate." Default is 0.5.

corOut

Return correlation of the returns or the influence function transformed returns. Must be one of "retCor", "retIFCor" or "none" (default).

return.coef

Boolean variable to indicate whether the coefficients of the penalized GLM fit are returned. Default if FALSE.

...

Additional parameters.

Value

A vector or a list depending on se.method.

Author(s)

Xin Chen, [e-mail]

Anthony-Alexander Christidis, [e-mail]

Examples

# Loading data
data(edhec, package = "PerformanceAnalytics")
# Changing the data colnames
names(edhec) = c("CA", "CTA", "DIS", "EM", "EMN",
                 "ED", "FIA", "GM", "LS", "MA",
                 "RV", "SS", "FOF")
# Computing the standard errors for
# the two influence functions based approaches
LPM.SE(edhec, se.method = c("IFiid","IFcor"),
       cleanOutliers = FALSE,
       fitting.method = c("Exponential", "Gamma")[1])