lmc_t: LMC Test for Student-t Tail Parameter

In wARMASVp: Winsorized ARMA Estimation for Higher-Order Stochastic Volatility Models

LMC Test for Student-t Tail Parameter

Description

Performs a Local Monte Carlo (LMC) test of the null hypothesis H_0: \nu = \nu_0 for the degrees of freedom parameter in an SV(p) model with Student-t errors. Testing \nu_0 = \infty (or a large value) corresponds to testing for normality.

Usage

lmc_t(
  y,
  p = 1,
  J = 10,
  N = 99,
  nu_null,
  burnin = 500,
  del = 1e-10,
  wDecay = FALSE,
  Bartlett = FALSE,
  Amat = NULL,
  logNu = TRUE,
  direction = c("two-sided", "less", "greater"),
  sigvMethod = "factored",
  winsorize_eps = 0
)

Arguments

y	Numeric vector. Observed returns.
p	Integer. AR order of the volatility process. Default 1.
J	Integer. Winsorizing parameter. Default 10.
N	Integer. Number of Monte Carlo replications. Default 99.
nu_null	Numeric. Value of \nu under the null hypothesis.
burnin	Integer. Burn-in for simulation. Default 500.
del	Numeric. Small constant for log transformation. Default 1e-10.
wDecay	Logical. Use decaying weights. Default FALSE.
Bartlett	Logical. Use Bartlett kernel HAC for weighting matrix. Default FALSE.
Amat	Weighting matrix specification. NULL (default) for identity weighting, "Weighted" for data-driven HAC, or a (p+3)x(p+3) matrix. Takes precedence over Bartlett.
logNu	Logical. Use log-space for nu estimation. Default TRUE.
direction	Character. Test direction: "two-sided" (default), "less" (H1: nu < nu_null), or "greater" (H1: nu > nu_null). Uses signed root of the LR statistic for one-sided tests.
sigvMethod	Character. Method for \sigma_v estimation: "factored" (default), "hybrid", or "direct".
winsorize_eps	Numeric. Winsorization threshold for moment conditions. Default 0 (no winsorization).

Value

An object of class "svp_test".

Examples

y <- sim_svp(1000, phi = 0.95, sigy = 1, sigv = 0.2, errorType = "Student-t", nu = 5)$y
test <- lmc_t(y, p = 1, J = 10, N = 49, nu_null = 5)
print(test)

wARMASVp documentation built on May 15, 2026, 5:07 p.m.