Multi-Register Administrative Data Linkage for Counterfactual Estimation
Predicting individual potential outcomes across multiple public policy interventions requires granular, long-term employment and earnings histories without survey attrition or reporting errors. Survey data lacks sufficient sample sizes and historical depth to estimate treatment effect heterogeneity across diverse demographic subgroups.
Picture this
Think of building a comprehensive financial history by merging bank transaction logs, tax returns, and employer payroll records into a single master timeline for every citizen. Instead of asking people to remember what they earned ten years ago, researchers connect official government record books to track exact monthly employment histories automatically.
What the evidence says
Linked multi-register database captured 11-year pre-unemployment labor histories and 12-to-24 month post-unemployment follow-up outcomes with zero panel attrition across the entire national population sample.
- Who
- Population database of N = 460,442 Swiss jobseekers aged 25 to 55 registered between January 2001 and December 2003.
- How
- Deterministic linkage of public unemployment insurance records (AVAM/ASAL) from 1998–2004 with central social security pension records (AHV) covering 11 years (1990–2002) of monthly earnings and employment status.
What to do
Link national unemployment benefit administration registers with social security pension records to build longitudinal monthly panel datasets for non-parametric counterfactual prediction.
From the source
"This data set was matched with the complete monthly information from the social security and pensions system (AHV) for the period January 1990 to December 2002. These combined data sources contain very detailed information on registration and de-registration of unemployment, benefit payments and sanctions, participation in ALMP, eleven years employment histories with monthly information on earnings and employment status."
Targeting_Labour_Market_Programmes_Results_from_a_Randomized_Experiment.pdf