On Omitted Variables, Proxies and Unobserved Effects in Analysis of Administrative Labour Market Data

Shihan Du, Pia Homrighausen, Ralf A. Wilke

Research output: Book/ReportReportResearch

Abstract

Empirical research addresses omitted variable bias in regression analysis by means of various approaches. We present a framework that nests some of them and put it to German linked administrative labour market data. We find evidence for sizable omitted variable bias in a wage regression, while a labour market transition model appears to be less affected. Additional survey variables contribute only to the wage model, while the use of work history variables and panel models lead to changes in coefficients in the two models. Overall, unobserved effects panel data models with a restricted regressor set are found to control for more information than cross sectional analysis with an extended variable set.
Original languageEnglish
Place of PublicationNürnberg
PublisherInstitute for Employment Research (IAB)
Number of pages38
DOIs
Publication statusPublished - 2018
SeriesFDZ-Methodenreport
Number6/2018

Keywords

  • Linked survey-administrative data
  • Statistical regularisation

Cite this

Du, S., Homrighausen, P., & Wilke, R. A. (2018). On Omitted Variables, Proxies and Unobserved Effects in Analysis of Administrative Labour Market Data. Nürnberg: Institute for Employment Research (IAB). FDZ-Methodenreport, No. 6/2018 https://doi.org/10.5164/IAB.FDZM.1806.en.v1
Du, Shihan ; Homrighausen, Pia ; Wilke, Ralf A. / On Omitted Variables, Proxies and Unobserved Effects in Analysis of Administrative Labour Market Data. Nürnberg : Institute for Employment Research (IAB), 2018. 38 p. (FDZ-Methodenreport; No. 6/2018).
@book{ac99ab037d0b40aabc23dc3925004b51,
title = "On Omitted Variables, Proxies and Unobserved Effects in Analysis of Administrative Labour Market Data",
abstract = "Empirical research addresses omitted variable bias in regression analysis by means of various approaches. We present a framework that nests some of them and put it to German linked administrative labour market data. We find evidence for sizable omitted variable bias in a wage regression, while a labour market transition model appears to be less affected. Additional survey variables contribute only to the wage model, while the use of work history variables and panel models lead to changes in coefficients in the two models. Overall, unobserved effects panel data models with a restricted regressor set are found to control for more information than cross sectional analysis with an extended variable set.",
keywords = "Linked survey-administrative data, Statistical regularisation, Linked survey-administrative data, Statistical regularisation",
author = "Shihan Du and Pia Homrighausen and Wilke, {Ralf A.}",
year = "2018",
doi = "10.5164/IAB.FDZM.1806.en.v1",
language = "English",
series = "FDZ-Methodenreport",
publisher = "Institute for Employment Research (IAB)",
number = "6/2018",
address = "Germany",

}

Du, S, Homrighausen, P & Wilke, RA 2018, On Omitted Variables, Proxies and Unobserved Effects in Analysis of Administrative Labour Market Data. FDZ-Methodenreport, no. 6/2018, Institute for Employment Research (IAB), Nürnberg. https://doi.org/10.5164/IAB.FDZM.1806.en.v1

On Omitted Variables, Proxies and Unobserved Effects in Analysis of Administrative Labour Market Data. / Du, Shihan; Homrighausen, Pia; Wilke, Ralf A.

Nürnberg : Institute for Employment Research (IAB), 2018. 38 p. (FDZ-Methodenreport; No. 6/2018).

Research output: Book/ReportReportResearch

TY - RPRT

T1 - On Omitted Variables, Proxies and Unobserved Effects in Analysis of Administrative Labour Market Data

AU - Du, Shihan

AU - Homrighausen, Pia

AU - Wilke, Ralf A.

PY - 2018

Y1 - 2018

N2 - Empirical research addresses omitted variable bias in regression analysis by means of various approaches. We present a framework that nests some of them and put it to German linked administrative labour market data. We find evidence for sizable omitted variable bias in a wage regression, while a labour market transition model appears to be less affected. Additional survey variables contribute only to the wage model, while the use of work history variables and panel models lead to changes in coefficients in the two models. Overall, unobserved effects panel data models with a restricted regressor set are found to control for more information than cross sectional analysis with an extended variable set.

AB - Empirical research addresses omitted variable bias in regression analysis by means of various approaches. We present a framework that nests some of them and put it to German linked administrative labour market data. We find evidence for sizable omitted variable bias in a wage regression, while a labour market transition model appears to be less affected. Additional survey variables contribute only to the wage model, while the use of work history variables and panel models lead to changes in coefficients in the two models. Overall, unobserved effects panel data models with a restricted regressor set are found to control for more information than cross sectional analysis with an extended variable set.

KW - Linked survey-administrative data

KW - Statistical regularisation

KW - Linked survey-administrative data

KW - Statistical regularisation

U2 - 10.5164/IAB.FDZM.1806.en.v1

DO - 10.5164/IAB.FDZM.1806.en.v1

M3 - Report

T3 - FDZ-Methodenreport

BT - On Omitted Variables, Proxies and Unobserved Effects in Analysis of Administrative Labour Market Data

PB - Institute for Employment Research (IAB)

CY - Nürnberg

ER -

Du S, Homrighausen P, Wilke RA. On Omitted Variables, Proxies and Unobserved Effects in Analysis of Administrative Labour Market Data. Nürnberg: Institute for Employment Research (IAB), 2018. 38 p. (FDZ-Methodenreport; No. 6/2018). https://doi.org/10.5164/IAB.FDZM.1806.en.v1