NISE Estimation of an Economic Model of Crime

TL;DR

This paper introduces a maximum-likelihood estimation method called NISE for an economic crime model, enabling consistent parameter estimation without the need for instrumental variables, addressing a common practical challenge.

Contribution

The paper presents NISE, a novel maximum-likelihood estimation technique for simultaneous linear equations in crime models, eliminating the reliance on instrumental variables.

Findings

NISE provides consistent estimates comparable to traditional methods.

NISE outperforms OLS and two-stage least squares in simulations.

The method addresses the issue of unavailable valid instruments.

Abstract

An economic model of crime is used to explore the consistent estimation of a simultaneous linear equation without recourse to instrumental variables. A maximum-likelihood procedure (NISE) is introduced, and its results are compared to ordinary least squares and two-stage least squares. The paper is motivated by previous research on the crime model and by the well-known practical problem that valid instruments are frequently unavailable.