Skip to contents

Tests the identifying assumption of the shocks route — that shocks are as-good-as-randomly assigned — by regressing the shocks on pre-determined shock-level characteristics, weighted by exposure (Borusyak, Hull & Jaravel 2022). Coefficients near zero and a non-significant joint test are *consistent with* balance: the null hypothesis of balance is not rejected. Note the asymmetry — failing to reject does **not** establish that the shocks are unrelated to observables (the test may simply lack power), whereas a rejection is direct evidence against shock exogeneity.

Usage

ssb_shock_balance(design, shock_covariates, weight = TRUE)

Arguments

design

An [ssb_design()] object.

shock_covariates

A `data.frame` keyed by `sector` (and `time` for panels) holding the shock-level characteristics to test.

weight

If `TRUE` (default) weight by exposure \(s_n\); else unweighted.

Value

A list (class `ssb_shock_balance`) with a coefficient table and the joint Wald test of the null hypothesis of balance.