The macroeconomic effects of fiscal policy in Spain

Abstract

This paper focuses on the effects of fiscal policy in Spain analysed in a VAR context. Fiscal shocks are found to involve significant effects on GDP, private consumption, private investment, interest rates and prices. Non-Keynesian effects are observed. Moreover, evidence on the channels highlighted in the literature for such effects to arise is found, notably the effects of permanent income on consumption and investment on the demand side, coupled with the response of the equilibrium wage on the supply side affecting entrepreneurial profits and investment. The response of interest rates seems to reinforce both effects. Furthermore, the different readings of spending or taxes do not affect macroeconomic variables homogeneously.

Acknowledgements

I wish to acknowledge the comments received from Javier Andrés, Luis J. Álvarez, Carlos Ballabriga, Ángel Estrada, Jordi Galí, Pablo Hernández de Cos, David López Salido, José M. Marqués, C. Martínez Mongay, José M. González-Páramo, Javier Vallés and an anonymous referee. The views expressed in this paper are the ones of the author and do not necessarily correspond to those of the Banco de España. All remaining errors are exclusively mine.

Notes

¹ Labour-tax cuts would yield similar effects although of lower magnitude.

² The definitions of G_t and T_t are analogous to those in Fatás and Mihov (2000), Blanchard and Perotti (2002) and Perotti (2002).

³ Admittedly, the number of lags is somewhat awkward for quarterly data. However, given the information provided by the tests mentioned it was considered convenient to set it accordingly. On the other hand, the VAR with four lags did not offer different results from those reported in this paper in terms of impulse responses or multipliers.

⁴ This constraint does not seem to be too restrictive in the present case. Alternative models were estimated without it and the results were broadly the same (see Section IV, ‘Other robustness checks’ for further details).

⁵ E.g., Bernanke and Blinder (1992), Bernanke and Mihov (1998) and Christiano et al. (1999).

⁶ Some components of government expenditure are value added produced by the private sector. Thus, a shock to this variable is automatically reflected in households’ disposable income and firms’ profits, which are expected to affect consumption and investment.

⁷ Fatás and Mihov (2000) find effects of a similar magnitude for the US economy, although reaching the peak takes more time.

⁸ One might find examples of rising public expenditure that pays for itself without needing higher taxes. Due to their likely positive spillovers for the private sector, increases in infrastructure spending might be offset by higher tax collections stemming from a positive reaction of output (see ‘Effects of shocks to government expenditure components’ below). In general, such positive spillovers are more likely to arise in the medium term, though.

⁹ The corresponding impulse responses are not presented for the shake of brevity.

¹⁰ This is the approach adopted by Fatás and Mihov (2000). It could be argued that omitting the rest of public expenditure items might bias the results. However, models that included both total expenditure and the specific component in turn were also estimated and led to very similar results, although with more imprecise estimates. Therefore, given the low number of observations compared with the large number of coefficients in the VAR, it made sense to reduce the VAR dimension. The same applies for net taxes.

¹¹ Nevertheless, the estimations in this paper might not fully account for the effects of public investment, since some investment programmes with important spillover effects are carried out by state-owned entities not included in the general government accounts according to ESA-95 definitions.

¹² The cumulative dynamic multiplier at a given quarter is obtained as the ratio of the cumulative response of GDP and the cumulative response of government expenditure.

¹³ Giorgioni and Holden (2003) find similar effects for some countries. They consider these cases to be consistent with the hypothesis of expansionary fiscal contractions.

¹⁴ The inclusion of long-term interest rates in the VAR yielded similar results.

¹⁵ Perotti (2002) and Marcellino (2002) offer mixed evidence on this issue. In the case of Spain, Marcellino finds a significant response in the first two years.

¹⁶ Setting a_1,4 to zero did not produce significantly different results.

¹⁷ Estrada et al. (1997) provide evidence for this hypothesis in the case of Spain.

¹⁸ De Castro (2003) offers a detailed discussion on this result.

¹⁹ Similar effects are also observed in Perotti (2002).

²⁰ The fact that public finances were unsustainable in the previous period (De Castro and Hernández de Cos, 2002) may have contributed for such effects to appear.