Effects of parliamentary elections on primary budget deficits in OECD countries – robustness of the results with regard to alternative econometric estimators

Abstract

Using an unbalanced panel of OECD countries for the period 1989 to 2005, the study tests the political budget cycle hypothesis empirically. The econometric approach is based on the equation proposed by Bohn (1998) for testing the sustainability of fiscal policy and various panel estimators. The empirical results strongly support the hypothesis of smaller primary surpluses (only) in election years. Brender and Drazen's (2005) result – an election effect exists only in new democracies – cannot be verified.

Notes

¹ A previous version of this research including preliminary estimation results can be found in Hagen (2007).

² Alternatively elect _it is included as an absolute term. The estimated effect hardly changes.

³ The stata command xtabond2 implemented by Roodman (2006) is applied.

⁴ The stata command xtlsdvc implemented by Bruno (2005) is applied. The estimator by Anderson and Hsiao (1982) is used to initialize the bias correction.

⁵ Mean d_{it −1} in the sample is 63.4% (Table 1). Therefore, the effect is 63.4 × (−0.502/100)/(1 − 0.881) = −2.67.