With thanks to Kit Baum for maintaining the SSC archive, I hereby announce the availability of the package esteta on SSC.

The command esteta estimates the long-run effect of an endogenous contemporary factor instrumented by a historical instrument (see Casey and Klemp, 2021). The estimator requires two variables representing the endogenous factor measured at two different points in time. It uses these two variables to estimate and adjust for the persistency of the endogenous contemporary factor in the estimation of the long-run effect.

The command is designed to estimate the long-run effect in a scenario satisfying the following causal diagram:
Code:
┌──────────┐   ┌───────────────────┐   ┌───────────────────┐   ┌───────┐
│Historical│   │Endogenous variable│   │Endogenous variable│   │Outcome│
│instrument│──▶│   (at time t1)    │──▶│   (at time t2)    │──▶│       │
└──────────┘   └───────────────────┘   └───────────────────┘   └───────┘
                         │               ┌───────────┐             ▲
                         │               │Alternative│             │
                         └──────────────▶│  channel  │─────────────┘
                                         └───────────┘
A conventional IV-regression instruments one of the instances of the endogenous variable. However, as shown in Casey and Klemp (2021), the estimate obtained from such a regression cannot be interpreted as a long-run effect unless the endogenous variable is perfectly persistent. Given measurements of the endogenous variable for two different time periods, the command esteta estimates the long-run effect of the endogenous variable.

Authors:
Gregory Casey (Williams College, USA) and Marc Klemp (University of Copenhagen, Denmark).

Reference:
Casey, G., & Klemp, M. (2021). Historical instruments and contemporary endogenous regressors. Journal of Development Economics, 149, 102586. https://doi.org/10.1016/j.jdeveco.2020.102586

Marc Klemp
web.econ.ku.dk/klemp