PanelMatch - Matching Methods for Causal Inference with Time-Series Cross-Sectional Data

Implements a set of methodological tools that enable researchers to apply matching methods to time-series cross-sectional data. Imai, Kim, and Wang (2018) <http://web.mit.edu/insong/www/pdf/tscs.pdf> proposes a nonparametric generalization of the difference-in-differences estimator, which does not rely on the linearity assumption as often done in practice. Researchers first select a method of matching each treated observation for a given unit in a particular time period with control observations from other units in the same time period that have a similar treatment and covariate history. These methods include standard matching methods based on propensity score and Mahalanobis distance, as well as weighting methods. Once matching is done, both short-term and long-term average treatment effects for the treated can be estimated with standard errors. The package also offers a visualization technique that allows researchers to assess the quality of matches by examining the resulting covariate balance.

Last updated 5 months ago

113 stars 7.36 score 45 dependencies

concordance - Product Concordance

A set of utilities for matching products in different classification codes used in international trade research. It supports concordance between the Harmonized System (HS0, HS1, HS2, HS3, HS4, HS5, HS6, HS combined), the Standard International Trade Classification (SITC1, SITC2, SITC3, SITC4), the North American Industry Classification System (NAICS 2002, 2007, 2012, 2017, combined), as well as the Broad Economic Categories (BEC4), the International Standard of Industrial Classification (ISIC2, ISIC3, ISIC3.1, ISIC4), and the Standard Industrial Classification (SIC). It also provides code nomenclature/descriptions look-up, product code look-up based on user-specified keywords, Rauch classification look-up (via concordance to SITC2), trade elasticity look-up (via concordance to HS0 or SITC3 codes), upstreamness/downstreamness look-up (via concordance to ISIC3 and NAICS codes), and intermediateness look-up (via product descriptions).

Last updated 2 months ago

41 stars 5.83 score 21 dependencies

wfe - Weighted Linear Fixed Effects Regression Models for Causal Inference

Provides a computationally efficient way of fitting weighted linear fixed effects estimators for causal inference with various weighting schemes. Weighted linear fixed effects estimators can be used to estimate the average treatment effects under different identification strategies. This includes stratified randomized experiments, matching and stratification for observational studies, first differencing, and difference-in-differences. The package implements methods described in Imai and Kim (2017) "When should We Use Linear Fixed Effects Regression Models for Causal Inference with Longitudinal Data?", available at <https://imai.fas.harvard.edu/research/FEmatch.html>.

Last updated 5 years ago

20 stars 4.38 score 13 dependencies