How Should Parameter Estimation Be Tailored to the Objective?

Peter Reinhard Hansen*, Elena-Ivona Dumitrescu

*Corresponding author for this work

Research output: Contribution to journalJournal articleResearchpeer-review


We study parameter estimation from the sample X, when the objective is to maximize the expected value of a criterion function, Q, for a distinct sample, Y. This is the situation that arises when a model is estimated for the purpose of describing other data than those used for estimation, such as in forecasting problems. A natural candidate for solving maxT∈σ(X)EQ(Y,T) is the innate estimator, θˆ=argmaxθQ(X,θ). While the innate estimator has certain advantages, we show that the asymptotically efficient estimator takes the form θ̃=argmaxθQ̃(X,θ), where Q̃ is defined from a likelihood function in conjunction with Q. The likelihood-based estimator is, however, fragile, as misspecification is harmful in two ways. First, the likelihood-based estimator may be inefficient under misspecification. Second, and more importantly, the likelihood approach requires a parameter transformation that depends on the true model, causing an improper mapping to be used under misspecification.
Original languageEnglish
JournalJournal of Econometrics
Issue number2
Pages (from-to)535-558
Number of pages24
Publication statusPublished - Oct 2022

Bibliographical note

Published online: 2. August 2021.


  • Estimation
  • Model selection
  • LinEx loss
  • Multistep forecasting

Cite this