Maximization by parts in extremum estimation

Yanqin Fan, Sergio Pastorello, Eric Renault

Research output: Contribution to journalArticleResearchpeer-review

6 Citations (Scopus)

Abstract

In this paper, we present various iterative algorithms for extremum estimation in cases where direct computation of the extremum estimator or via the Newton–Raphson algorithm is difficult, if not impossible. While the Newton–Raphson algorithm makes use of the full Hessian matrix, which may be difficult to evaluate, our algorithms use parts of the Hessian matrix only, the parts that are easier to compute. We establish consistency and asymptotic efficiency of our iterative estimators under regularity and information dominance conditions. We argue that the economic interpretation of a structural econometric model will often allow us to give credibility to a well-suited information dominance condition. We apply our algorithms to the estimation of the Merton structural credit risk model and to the Heston stochastic volatility option pricing model.

Original languageEnglish
Pages (from-to)147-171
Number of pages25
JournalEconometrics Journal
Volume18
Issue number2
DOIs
Publication statusPublished - Jun 2015
Externally publishedYes

Keywords

  • Extremum estimators
  • Generalized method of moments
  • Implied-states GMM
  • Iterative backfitting methods
  • Maximum likelihood
  • Structural econometrics

Cite this