Testing for the government's intertemporal budget restriction in Brazil during 1823–1889

Abstract

This article tests whether the government's intertemporal budget restriction was fulfilled during the Brazilian imperial period (1823–1889). To accomplish this, newly developed tests for cointegration with unknown structural breaks are applied. It is found that government spending and government revenue are cointegrated if the effect of two unknown structural breaks is taken into account. The estimated parameter vector reveals that the one-by-one relationship that is required for solvency does not prevail for the sub-periods before and after the first break, however it prevails for the sub-period after the second break. We interpret these findings as empirical support for the long-run government solvency in Brazil at the end of the imperial period.

Keywords:

• government budget
• multiple breaks
• cointegration
• Brazil

JEL Classification::

• H21
• H62
• C20

Acknowledgements

We wish to thank the editor of this journal Mark. P. Taylor and the anonymous referee for constructive comments that resulted in improving the article significantly. However, any remaining errors are our own responsibility.

Notes

¹ The long-run variance is estimated by using an automatic bandwidth estimator and a prewhitened quadratic spectral kernel with a first-order autoregression for the prewhitening. For details see Andrews (1991) and Andrews and Monahan (1992).

² These two tests were originally introduced by Phillips (1987) to be used to test for cointegration without any breaks. Gregory and Hansen (1996) extended these tests for cointegration with one unknown break and Hatemi-J (2008) extended them for cointegration with two unknown breaks.

³ It should be mentioned that the critical values are produced via the response surface technique, which takes into account the effect of small sample sizes.

⁴ The Brazilian currency was called mil reis at that time.