Breaking trend, Lagrange multiplier test statistic and the presence of a unit root in the Brazilian gross domestic product

Abstract

Standard unit root tests provided mixed evidence on the stochastic behaviour of the Brazilian gross domestic product series. This study uses the minimum Lagrange multiplier statistic suggested by Lee and Strazicich to test for the presence of a unit root with two endogenously determined structural changes. Contrary to previous works utilizing endogenous break points, the authors were not able to reject the null of unit root.

Acknowledgements

The authors thank Milton Barossi Filho for kindly providing the Gauss codes. They also thank Maximiliano Barbosa, participants in the 10th Time Series and Econometrics School meeting in Brazil and, especially, Afonso H. B. Ferreira and Antonio Aguirre for providing us with data and also for helpful discussions. Financial support from Fundação Instituto de Pesquisa Econômica (FIPE) and Conselho Nacional de Desenvolvimento Científico e Tecnológico (CNPq) is acknowledged.

Notes

¹ See Lee and Strazicich (2001) for critical Monte Carlo evidence on these procedures.

² The first sub-sample was drawn from Haddad (1977), which gives rise to a possible break due to different methodologies.

³ Lee and Strazicich's (2002) test allows us to test for the presence of two endogenous break points.

⁴ Data may be received by anyone interested upon request to the third author.