Regression with an Imputed Dependent Variable

IFS Working Paper W19/16

38 Pages Posted: 11 Nov 2019

See all articles by Thomas F. Crossley

Thomas F. Crossley

European University Institute - Economics Department (ECO); Institute for Fiscal Studies

Peter Levell

Institute for Fiscal Studies (IFS)

Stavros Poupakis

University College London - Institute for Global Health

Date Written: June 1, 2019

Abstract

Researchers are often interested in the relationship between two variables, with no single data set containing both. A common strategy is to use proxies for the dependent variable that are common to two surveys to impute the dependent variable into the data set containing the independent variable. We show that commonly employed regression or matching-based imputation procedures lead to inconsistent estimates. We offer an easily-implemented correction and correct asymptotic standard errors. We illustrate these with Monte Carlo experiments and empirical examples using data from the US Consumer Expenditure Survey (CE) and the Panel Study of Income Dynamics (PSID).

Keywords: Imputation; Measurement error; Consumption

JEL Classification: C81, C13, E21

Suggested Citation

Crossley, Thomas F. and Levell, Peter and Poupakis, Stavros, Regression with an Imputed Dependent Variable (June 1, 2019). IFS Working Paper W19/16, Available at SSRN: https://ssrn.com/abstract=3478717 or http://dx.doi.org/10.2139/ssrn.3478717

Thomas F. Crossley

European University Institute - Economics Department (ECO) ( email )

Villa San Paolo
Via della Piazzuola 43
50133 Florence
Italy

Institute for Fiscal Studies ( email )

7 Ridgmount Street
London, WC1E 7AE
United Kingdom

Peter Levell

Institute for Fiscal Studies (IFS) ( email )

7 Ridgmount Street
London, WC1E 7AE
United Kingdom

Stavros Poupakis (Contact Author)

University College London - Institute for Global Health ( email )

United Kingdom

Do you have a job opening that you would like to promote on SSRN?

Paper statistics

Downloads
48
Abstract Views
261
PlumX Metrics