Method of Combining LP-DID with Matching

Tyler Ludwig

CES Technical Notes Series from Center for Economic Studies, U.S. Census Bureau

Abstract: Many of the difference-in-differences methodologies advanced in the past few years to contend with heterogeneous treatment timing are too computationally taxing to be used in the very large data environment of Census data, such as the LEHD Infrastructure files. One option is local projections difference-in-differences (LP-DiD), put forward by Dube et al (2023). I propose two key modifications for RDC researchers, depending on their use-case. The first, which is applicable across projects, is to separate LP-DiD’s data preparation step from the analysis step. The second, which is relevant depending on the context of the research project, is to perform matching after data preparation and before performing LP-DiD analyses to define more comparable treatment and control groups.

Keywords: LEHD; Decennial Census (search for similar items in EconPapers)
Date: 2026-10
References: Add references at CitEc
Citations:

Downloads: (external link)
https://www2.census.gov/ces/tn/CES-TN-2026-48.pdf Abstract (application/pdf)
https://www.census.gov/about/adrm/ced/apply-for-access.html?CES-TN-2026-48 First version, 2026 (application/pdf)
CES Technical Notes may contain confidential data and, thereby, disclosure is prohibited. Researchers on approved projects (to apply for access, please see https://www.census.gov/ces/rdcresearch/howtoapply.html) with the correct permissions can request full text notes from CES.Technical.Notes.List@census.gov.

Related works:
This item may be available elsewhere in EconPapers: Search for items with the same title.

Export reference: BibTeX RIS (EndNote, ProCite, RefMan) HTML/Text

Persistent link: https://EconPapers.repec.org/RePEc:cen:tnotes:26-48

Access Statistics for this paper

More papers in CES Technical Notes Series from Center for Economic Studies, U.S. Census Bureau Contact information at EDIRC.
Bibliographic data for series maintained by Danielle H. Sandler ().