svpSE: Simulation-Based Standard Errors for SV(p) Models

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

Simulation-Based Standard Errors for SV(p) Models

Description

Computes standard errors and confidence intervals for estimated parameters by simulating from the fitted model and re-estimating. Supports all model types returned by svp: Gaussian (with or without leverage), Student-t, and GED.

Usage

svpSE(object, n_sim = 199, alpha = 0.05, burnin = 500, logNu = FALSE)

Arguments

object	A fitted model object from svp. Can be of class "svp", "svp_t", or "svp_ged".
n_sim	Integer. Number of Monte Carlo replications. Default 199.
alpha	Numeric. Significance level for confidence intervals. Default 0.05.
burnin	Integer. Burn-in period for simulation. Default 500.
logNu	Logical. Solve for \nu in log-space for numerical stability (Student-t only). Default is FALSE.

Value

A list with:

CI

2 x k matrix of confidence intervals (lower, upper).

SEsim0

Standard errors relative to true parameter values.

SEsim

Standard errors relative to sample mean.

ISEconservative

Conservative interval-based standard errors.

ISEliberal

Liberal interval-based standard errors.

thetamat

Matrix of parameter estimates from simulations.

Examples

# Gaussian SV(1)
y <- sim_svp(1000, phi = 0.95, sigy = 1, sigv = 0.2)$y
fit <- svp(y)
se <- svpSE(fit, n_sim = 49)
se$CI

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