Is the Greek budget deficit sustainable after all? Empirical evidence accounting for regime shifts

Abstract

This article re-examines the sustainability of the Greek budget deficit by using a formal framework based on the government’s intertemporal budget constraint. The empirical analysis uses annual data from 1960 to 2011 and employs traditional as well as more recent unit root and cointegration techniques that account for linear and nonlinear effects in fiscal policy actions. Unlike previous studies, the evidence suggests that, allowing for structural breaks, the Greek budget deficit is unsustainable. The parameter after the second detected break reflects the structural deficiencies of the Greek economy.

Keywords:

• fiscal policy
• budget deficit
• sustainability
• structural breaks

JEL Classification:

• C22
• E62
• H61
• H62

Acknowledgements

We are grateful to Professor Dimitris Hatzinikolaou and two anonymous referees for providing valuable comments, which substantially improved this article. The usual caveat applies.

Notes

¹ The Maastricht treaty criteria required the economies of the candidate European countries to converge to low inflation rates, interests rates, stable exchange rates, a budget deficit no greater than 3% of GDP and a debt-to-GDP ratio no greater than 60% of GDP.

² As Cuddington (1997, p. 8) points out, ‘the NPG condition is usually justified by arguing that lenders would presumably not be willing to allow the government to perpetually pay their entire current interest obligation merely by borrowing more’ [i.e., using Equation 1, if , hence , instead of running primary surpluses ()]. If lenders were willing to buy such debt when for all , then Equation 1 implies that the debt would grow at a rate () equal to the interest rate (, not ), and thus the discounted debt in Equation 3 would not converge to zero and the NPG condition would fail (Hatzinikolaou and Simos, 2013, p. 63).

³ From another perspective, Bohn (2007) claims that the intertemporal budget constraint imposes rather weak econometric constraints on the time series properties. In particular, the IBC may well be satisfied even if the components of the budget deficit are not cointegrated and even if neither debt, nor budget revenues or spending are difference-stationary.

⁴ However, in order for the trajectory of the undiscounted debt not to diverge in an infinite horizon, it is necessary to have (Hakkio and Rush, 1991, p. 433).

⁵ According to McCallum (1984) and Hakkio and Rush (1991), the use of ratios is more pertinent for a growing economy. On the other hand, Cuddington (1997) claims that the conversion of the present value budget constraint and the NPG condition into ratios leaves them unaffected.

⁶ This value has been calculated as follows: (Table 7).

⁷ Hatzinikolaou and Simos (2013) using data for the US fiscal and current account deficit employed an innovative test that requires formally that the undiscounted debt be bounded and rejected sustainability. This test is very likely to provide similar inference for the Greek fiscal deficit.

⁸ We are grateful to an anonymous referee of the journal for raising this point.