Generalised moment selection test statistic

t_j = sqrt(n) mbar_j / sigma_j; xi_j = t_j / kappa_n with kappa_n = sqrt(log n); moment j is retained when xi_j > -1, and S = sum over retained of [max(t_j, 0)]^2.

Usage

Gmsbound(mbar, sigma, n, kappa = NULL)

Arguments

mbar: Sample moment means, positive meaning violation.
sigma: Moment standard deviations, strictly positive.
n: Sample size.
kappa: Tuning sequence; NULL uses sqrt(log n).

Value

List with S, t, xi, retained, nretained, kappa, n, J.

References

Andrews and Soares (2010), Econometrica 78(1), 119-157, Sects. 3-4; standard published form.

Examples

Gmsbound(mbar = 5L, sigma = 0.5, n = 5L)
#> $S
#> [1] 500
#> 
#> $t
#> [1] 22.36068
#> 
#> $xi
#> [1] 17.62576
#> 
#> $retained
#> [1] 1
#> 
#> $nretained
#> [1] 1
#> 
#> $kappa
#> [1] 1.268636
#> 
#> $n
#> [1] 5
#> 
#> $J
#> [1] 1
#> 
#> $method
#> [1] "Generalised moment selection (Andrews-Soares 2010)"
#> 
#> attr(,"class")
#> [1] "morie_rich_result" "list"