Causal Inference by Scott Cunningham

Author:Scott Cunningham
Language: eng
Format: epub
Publisher: Yale University Press
Published: 2020-12-15T00:00:00+00:00

Figure 42. McCrary density test using local linear nonparametric regressions.

Concluding remarks about close-election designs. Let’s circle back to the close-election design. The design has since become practically a cottage industry within economics and political science. It has been extended to other types of elections and outcomes. One paper I like a lot used close gubernatorial elections to examine the effect of Democratic governors on the wage gap between workers of different races [Beland, 2015]. There are dozens more.

But a critique from Caughey and Sekhon [2011] called into question the validity of Lee’s analysis on the House elections. They found that bare winners and bare losers in US House elections differed considerably on pretreatment covariates, which had not been formally evaluated by Lee et al. [2004]. And that covariate imbalance got even worse in the closest elections. Their conclusion is that the sorting problems got more severe, not less, in the closest of House races, suggesting that these races could not be used for an RDD.

At first glance, it appeared that this criticism by Caughey and Sekhon [2011] threw cold water on the entire close-election design, but we since know that is not the case. It appears that the Caughey and Sekhon [2011] criticism may have been only relevant for a subset of House races but did not characterize other time periods or other types of races. Eggers et al. [2014] evaluated 40,000 close elections, including the House in other time periods, mayoral races, and other types of races for political offices in the US and nine other countries. No other case that they encountered exhibited the type of pattern described by Caughey and Sekhon [2011]. Eggers et al. (2014) conclude that the assumptions behind RDD in the close-election design are likely to be met in a wide variety of electoral settings and is perhaps one of the best RD designs we have going forward.

Download

Copyright Disclaimer:
This site does not store any files on its server. We only index and link to content provided by other sites. Please contact the content providers to delete copyright contents if any and email us, we'll remove relevant links or contents immediately.