The problem with accounting partial identities

  1. Francisco Javier Sánchez Vidal 1
  1. 1 Universidad Politécnica de Cartagena
    info

    Universidad Politécnica de Cartagena

    Cartagena, España

    ROR https://ror.org/02k5kx966

Actas:
XVIII Encuentro Internacional AECA

Editorial: Asociación Española de Contabilidad y Administración de Empresas

ISBN: 978-84-16286-48-5

Año de publicación: 2018

Páginas: 15

Tipo: Aportación congreso

GOOGLE SCHOLAR

Resumen

This experiment uses a Monte Carlo simulation designed to test theproblems about the use of accounting identities are present in the model ofFazzari, Hubbard, and Petersen (1988). The Monte Carlo simulationcreates sets of randomly generated cash flows, Tobin’s Q, error termvariables and investments variables that are made dependent on the formerones. The investments and cash flow variables are also related through apartial accounting identity. OLS estimations using investments as thedependent variable and cash flows and Tobin’s Q as independent variablesverify that estimated coefficients do not represent reality. The closer thedata are to the accounting identity, the less the regression will tell about thecausal relation.