Testing the Presence of Outliers in Regression Models
Posted: 8 Aug 2018 Last revised: 6 Jun 2022
Date Written: July 20, 2018
Abstract
We propose two sets of tests for the overall presence of outliers in regression models. First, `simple' tests on whether the proportion and the number of detected outliers deviate from their expected values. Second, `scaling' tests on whether the proportion of outliers decreases with the cut-off used to detect outliers. We apply our tests to a panel difference-in-differences model of transport CO2 emissions in response to the introduction of North America's first major carbon tax. Our tests show the presence of significant outliers in the un-taxed control group which results in an over-estimation of the estimated impacts of the tax.
Keywords: misspecification, outlier detection, robust estimation, iterated 1-step Huber-skip M-estimator, indicator saturation
JEL Classification: C12, C52
Suggested Citation: Suggested Citation