Detecting Person Identifier Misuse: Understanding the Longitudinal Employer Household Dynamics (LEHD) Pseudo-PIK Process

The Longitudinal Employer Household Dynamics (LEHD) program at the U.S. Census Bureau uses primarily administrative data to release public use statistics such as the Quarterly Workforce Indicators (QWI). Administrative data from state unemployment insurance (UI) systems provides extensive coverage of quarterly job earnings, however the person identifier attached to UI earnings reports is susceptible to misuse. The Pseudo-PIK process was developed to detect work histories where multiple people use the same Person Identification Key (PIK). The pseudo-PIK process assigns a new person identifier to each selected job along with a complete set of imputed demographic characteristics. The pseudo-PIK process improves the quality of tabulations by person demographic characteristics, while not affecting tabulations by firm characteristics. For example, in some NAICS 4 industries a relatively large share of job quarters are assigned a pseudo-PIK, greatly increasing the share of Hispanic workers while leaving the total number of jobs in the industry unchanged. The ultimate impact of the Pseudo-PIK process on LEHD demographic tabulations depends primarily on the share of jobs with a pseudo-PIK and the distribution of their imputed characteristics. The higher the share of jobs with a pseudo-PIK and/or the extent pseudo-PIK imputed characteristics differ from the characteristics of the original UI reported PIK the larger the impact.