Accessibility statement

Ideology and Econometric Output

Wednesday 20 January 2021, 1.00PM to 2.00 pm

Speaker(s): Christopher Hennessy (London Business School)

Host: Kostas Koufopoulos

Abstract: Friedman (1953) outlined ideal economic methodology. First, positive economics provides objective evidence. Then, the objective evidence informs normative decisions. Many, including Friedman, have alleged this ideal has been blocked by partisan econometricians. In this paper, I analyze how a controlling government's ideology inadvertently biases its econometric estimates. An objective econometrician estimates coefficients for an honest, yet ideological, government basing transfers to households on its Bayesian coefficient posteriors, and measured household covariates. The seemingly-objective evidence, MLE/OLS coefficients estimated by the econometrician, changes with: the government's ideological priors; the weight the government's posterior places on incoming evidence; and governmental values mapping estimated causal effects to transfers. Ideological contamination arises because hidden actions of forward-looking transfer-seeking households, not the econometrician, reflect the controlling government's ideology. Thus, elimination of econometrician malfeasance does not ensure evidence objectivity, nor does evidence-ideology correlation imply econometrician malfeasance. A bias formula is provided for applied settings, with regions of coefficient undershooting, overshooting, or sign reversals arising depending on effort incentives implied by the estimation-transfer scheme.


Location: ZOOM (details to follow)

Admission: All welcome