Stock-Yogo critical values (TSLS maximal size, one endogenous regressor)

The first-stage F thresholds of Stock and Yogo (2005, Table 5.2) for a nominal 5 percent TSLS Wald test whose actual size is at most 10, 15, 20 or 25 percent, for one endogenous regressor and one to three instruments: the rows widely reprinted in textbooks. They assume iid errors. Outside those rows, or with heteroskedastic, clustered or serially correlated errors, use morie_iv_montiel_olea_pflueger (effective F with critical values computed from the data) or morie_iv_kleibergen_paap (rk Wald F, rule of thumb 10).

Usage

morie_iv_stock_yogo(n_endogenous = 1, n_instruments = 1)

Arguments

n_endogenous

Number of endogenous regressors.

n_instruments

Number of excluded instruments.

Value

A named list of critical values (10pct ... 25pct).

References

Stock, J. H. and Yogo, M. (2005). Testing for weak instruments in linear IV regression. In Andrews and Stock (eds), Identification and Inference for Econometric Models, Cambridge University Press, 80-108.

Examples

out <- morie_iv_stock_yogo(n_endogenous = 1, n_instruments = 1)
out
#> $`10pct`
#> [1] 16.38
#> 
#> $`15pct`
#> [1] 8.96
#> 
#> $`20pct`
#> [1] 6.66
#> 
#> $`25pct`
#> [1] 5.53
#>