White paper
Statistical methods for imputing race and ethnicity
By Larry Baeder, Erica S. Baird (formerly, Erica Rode), Peggy Brinkmann, Joe Long, Caleb Stracke, Kweweli Togba-Doya, Gabriele Usan, Natalie Weaver, and Meseret Woldeyes
29 April 2024
Lack of consistent data collection or reporting is often an obstacle in the study of disproportionate impacts and equity across race and ethnicity cohorts in the insurance industry. This paper, published in April 2024 by the Society of Actuaries Research Institute, describes a range of techniques for developing probabilistic estimates or predictions of individual race and ethnicity. After providing some background on definitions and performance metrics, the article:
- Covers three main types of imputation methods
- Discusses direct data, including challenges and considerations
- Presents a case study on imputation methods
- Provides a tutorial to help readers become familiar with imputation methods
This paper was published by the Society of Actuaries. Read the full paper here.
Explore more tags from this article
About the Author(s)
Caleb Stracke
Statistical methods for imputing race and ethnicity
Imputation is a powerful tool for studying the disproportionate impact of race and ethnicity on insurance, so we discuss some uses and limitations.
Larry Baeder, Erica Baird (formerly, Erica Rode), Peggy Brinkmann, Joe Long, Caleb Stracke, Kweweli Togba-Doya, Gabriele Usan, Natalie Weaver, Meseret Woldeyes