Government expenditure and national income in Mexico: Keynes versus Wagner

Abstract

This article examines the relationship between government expenditure and national income in Mexico by testing the validity of Wagner's law and Keynes's hypothesis for the period between 1950 and 1999. More specifically, by applying time-series analysis, government-spending and national-income variables were found to be nonstationary and cointegrated, thus satisfying a long-run equilibrium condition. In addition, through the application of Granger causality tests to error-correction models, unidirectional causality, running from Gross Domestic Product (GDP) to government-expenditure variables, could be established between the variables and, therefore, only Wagner's law was found to be valid in Mexico's case for the period of study.

Acknowledgements

The author is grateful to Galindo-Paliza Luis, Katsenos Georgios, Kuroda Yoshimi, Okubo Masakatsu, Shoji Isao, Turnbull Stephen, Yakita Akira and seminar participants at the Nagoya Macroeconomics Workshop for very helpful comments and suggestions. Additionally, the author acknowledges the Ministry of Education, Culture, Sports, Science and Technology (Monbukagakusho), Government of Japan, for the financial support provided for the realization of this study.

Notes

¹ See, for example, the studies by Gupta (1967), Rubinson (1977), Singh and Sahni (1984), Afxentiou and Serletis (1991), Ahsan et al. (1996) and Islam (2001).

²The data set was taken from the Penn World Table Version 6.1 (Heston et al., 2002).

³The Error Correction Model (ECM) technique for evaluating the short-run equilibrium, or dynamic model, and the Granger causality test for determining the causal relationship between the variables are included in this study in accordance to the latest studies made in this area.

⁴With the sole exception of Murthy (1994), where an ECM was determined for the period 1950 to 1980, in no earlier study was an ECM determined or Granger causality test used for determining the actual causal relationship between government spending and Gross Domestic Product (GDP). At the same time, the studies of Mann (1980) and Nagarajan and Spears (1990) may have strong critical deficiencies, like the ones of not testing for unit roots, or for cointegration, which, if present, could make their respective conclusions and results invalid, as the corresponding analysed variables would be nonstationary.

⁵For an elaborate discussion of the unit-root tests applied in this study, see Harris (1995, pp. 27–51) and Enders (2004, pp. 181–206).

⁶For a detailed explanation of the Engle–Granger methodology, see Enders (2004, pp. 335–39).

⁷See Charemza and Deadman (1997, pp. 166–7) and Enders (2004, pp. 283–7) for a detailed explanation of these tests.

⁸The Keynesian hypothesis is tested when we express these four equations in terms of GDP as a function of government expenditure in the ECMs.

⁹See Mann's (1980) study for an interpretation of six different versions of Wagner's law, including the four versions analysed in the present study.

¹⁰Heston et al. (2002).