This analysis of the results provides good evidence that the composition of fiscal instruments affects economic growth to a certain level. Interestingly, the findings show effects of capital spending on growth to exceed those of GCE in both regressions 4.1 and 4.2, which is in line with commonly held, views. Capital spending has a positive and significant coefficient with the best output elasticity of 0.502. Supporting Luiz R. De Mello (2002) who also found a positive relationship between capital spending and growth in Brazil. Gupta et al. (2005) and Easterly and Rebelo (1993) find similar results in their respective research. It therefore confirms the hypothetical prediction of a positive and significant impact of capital investment on growth. Therefore, it may be preferable to allocate public spending into infrastructures and starting new schemes rather than sustaining and improving existing ones. This implies that Brazil’s economy is likely to perform better if more resources are diverted from government consumption to capital spending. This may reflect capital investment patterns that are optimally growth promoting as described by Gupta et al. (2005).
Unfortunately, current spending data is a sum of all consumption expenditure; thus making it impossible to identify the components of current spending that is detrimental to growth. This suggests that reforming the data collection structure could be valuable for ensuring that “productive” current spending while writing the annual fiscal budget.
Consistent with what was assumed, urbanisation in Brazil was negatively related to growth. All three regressions show consistent negative relationships with a magnitude of its coefficient 1.885 on average. This is consistent with Bloom et al (2008) in which urbanisation was found to be a negative and insignificant determinant of growth.
Trade Openness shared a negative relationship with GDP growth. The Trade coefficient remained consistently negative with a magnitude of -0.383 in regression 4. One assumption why trade openness expressed negative results could be due to its tardy response to growth. Since Brazil has recently undergone a period of hyperinflation together with an increase in tax burden, its sluggish effects on growth is a good argument. The negative relationship with growth is not in line with that of Sachs and Warner (1995). However, it is important to mention Rodriguez and Rodrik (2000) analysis, where he criticise the results of Sachs and Warner (1995) and claims that the results are insignificant due to high collinearity between trade indicators and growth. Nonetheless, it is true that Brazil would be better off if tax burden is lowered. This would perhaps generate growth as stated by Dollar and Kray (2004).
As expected the Taxes on product shares a negative relationship with growth. Nevertheless, in all four regressions, the results on Taxes were all insignificant possessing elasticities of - 0.095. Due to...