boot_ssc

set the small sample correction factor applied in <code>boottest()</code>

Implementation of fast algorithms for wild cluster bootstrap
inference developed in 'Roodman et al' (2019, 'STATA' Journal,
<doi:10.1177/1536867X19830877>) and 'MacKinnon et al' (2022),
which makes it feasible to quickly calculate bootstrap test
statistics based on a large number of bootstrap draws even for
large samples. Multiple bootstrap types as described in 'MacKinnon,
Nielsen & Webb' (2022) are supported.
Further, 'multiway' clustering, regression weights,
bootstrap weights, fixed effects and 'subcluster' bootstrapping
are supported. Further, both restricted ('WCR') and unrestricted
('WCU') bootstrap are supported. Methods are provided for a variety
of fitted models, including 'lm()', 'feols()'
(from package 'fixest') and 'felm()' (from package 'lfe').
Additionally implements a 'heteroskedasticity-robust' ('HC1') wild
bootstrap.
Last, the package provides an R binding to 'WildBootTests.jl',
which provides additional speed gains and functionality,
including the 'WRE' bootstrap for instrumental variable models
(based on models of type 'ivreg()' from package 'ivreg')
and hypotheses with q > 1.

Alexander Fischer

fwildclusterboot

Fast Wild Cluster Bootstrap Inference for Linear Models

David Roodman

Achim Zeileis

Nathaniel Graham

Susanne Koell

Laurent Berge

Sebastian Krantz

boot_ssc function

<dl><dt>adj</dt>
<dd>Logical scalar, defaults to TRUE. If TRUE, applies a small
sample correction of (N-1) / (N-k) where N is the number of observations
and k is the number of estimated coefficients excluding any fixed
effects projected out in either fixest::feols() or lfe::felm().</dd>
<dt>fixef.K</dt>
<dd>Character scalar, equal to 'none': the fixed effects
parameters are discarded when calculating k in (N-1) / (N-k).</dd>
<dt>cluster.adj</dt>
<dd>Logical scalar, defaults to TRUE. If TRUE, a cluster
correction G/(G-1) is performed, with G the number of clusters.</dd>
<dt>cluster.df</dt>
<dd>Either "conventional"(the default) or "min". Controls
how "G" is computed for multiway clustering if cluster.adj = TRUE.
Note that the covariance matrix in the multiway clustering case is of
the form V = V_1 + V_2 - V_12. If "conventional", then each summand G_i
is multiplied with a small sample adjustment G_i / (G_i - 1). If "min",
all summands are multiplied with the same value, min(G) / (min(G) - 1)</dd></dl>

Arguments

set the small sample correction factor applied in boottest() — boot_ssc

<dl>

<dt>adj</dt>
<dd>Logical scalar, defaults to TRUE. If TRUE, applies a small
sample correction of (N-1) / (N-k) where N is the number of observations
and k is the number of estimated coefficients excluding any fixed
effects projected out in either fixest::feols() or lfe::felm().</dd>


<dt>fixef.K</dt>
<dd>Character scalar, equal to 'none': the fixed effects
parameters are discarded when calculating k in (N-1) / (N-k).</dd>


<dt>cluster.adj</dt>
<dd>Logical scalar, defaults to TRUE. If TRUE, a cluster
correction G/(G-1) is performed, with G the number of clusters.</dd>


<dt>cluster.df</dt>
<dd>Either "conventional"(the default) or "min". Controls
how "G" is computed for multiway clustering if cluster.adj = TRUE.
Note that the covariance matrix in the multiway clustering case is of
the form V = V_1 + V_2 - V_12. If "conventional", then each summand G_i
is multiplied with a small sample adjustment G_i / (G_i - 1). If "min",
all summands are multiplied with the same value, min(G) / (min(G) - 1)</dd>

</dl>

boot_ssc: set the small sample correction factor applied in `boottest()`

Description

Usage

Value

Arguments

Examples