Did fiscal institutions affect Wagner’s law in Italy during 1951–2009 period? An empirical analysis

Abstract

In this article, we test Wagner’s assumption of the one-sided directional flow moving from economic growth to public spending in Italy for the 1951–2009 period. We pay particular attention to the impact of certain regime shifts related to changes in Italian budget regulations and procedures and the relevance of fiscal institutions to the fiscal performance equation, i.e. the public spending–national income nexus. The Error Correction Model is estimated to measure short-run dynamic effects and the long-run equilibrium between the two time series. The empirical evidence suggests that Wagner’s law is supported. In regard to policy implications, we find that public spending reacted less to positive changes in economic growth when the strengthening of the Ministry of Finance occurred in 1997 (Ciampi’s reform). Some sensitivity analyses confirm our empirical evidence.

Keywords:

• Wagner’s law
• public spending
• national income
• regime shifts

JEL Classification:

• O40
• H54

Acknowledgements

The authors wish to thank Sergio Destefanis, Paolo Coccorese, Luigi Senatore and anonymous referees for helpful comments on a previous version of the article. The usual disclaimer applies.

Notes

¹ In some studies, Wagner’s law is only supported for wealthier countries (Abizadeh and Gray, 1985).

² In 1979, Beaulieu stated that the maximum amount of public spending is determined by the maximum taxation that citizens believe bear (which binds taxation in public spending). In 1945, Clark instead sustains the idea that government spending should not exceed 25% of the national income. In fact, the increase of fiscal taxation on businesses causes an increase in prices. In 1976, Friedman asserts that the relationship between public spending and national income should not exceed 60%. Finally, Brennan was convinced that the limit of public spending was a social contract between rulers and ruled that establishes the maximum level of taxation and then spending.

³ Wagner’s law may be investigated using other versions of the model proposed by Peacock and Wiseman (1961), Goffman (1968), Musgrave (1969) and Michas (1975).

⁴ The concept of cointegration, originally suggested by Granger (1981) and formulated by Engle and Granger (1987), is that in the long run, a special time-invariant linear combination of nonstationary variables may be stationary.

⁵ 1978 (Law no. 468): Basic Law of the budget. Introduction of the budget bill; 1988 (Law no. 362): Amendments to the 468/1978 law. Strengthening coverage obligation. Planning Document Introduction. Resizing the Finance Act. Edit Regulations parliamentary; 1981: divorce between Treasury and the Bank of Italy; 1997 (Law no. 94 – Ciampi’s reform): Merging the Ministries of Budget and Treasury into a single Ministry: Strengthening the Ministry of Economics.

⁶ As suggested by Hansen (1992), these tests are useful for determining whether the cointegrating relationship was subjected to a regime shift.

⁷ Notice that a key assumption of the ARDL bounds testing is that the errors must be serially independent. We then proceed to test the correlation using LM test, which confirms that the errors in Equation 2 are not correlated. The ‘dynamic stability’ of the model is generally guaranteed by an ‘autoregressive structure’.