Military Spending and Economic Growth in the Middle East Countries: Bootstrap Panel Causality Test

Abstract

This study revisits the causal relationship between military spending and economic growth in 10 Middle East countries via a panel causality analysis that accounts for cross-sectional dependence and heterogeneity across countries. Our results indicate unidirectional causality from military spending to growth for Turkey; one-way causality from economic growth to military spending for Egypt, Kuwait, Lebanon, and Syria; bidirectional causality for Israel; and no causality in either direction for Jordan, Oman, and Saudi Arabia. The empirical evidence does not provide consistent results regarding the causal relationship between defense expenditure and economic growth in these countries.

Keywords:

• Military Spending
• Economic growth
• Dependency and heterogeneity
• Panel causality test
• Middle East countries

• JEL Codes
• H56
• O41
• C23

Acknowledgements

We would like to thank the editor and the anonymous referee for their highly constructive comments. All remaining errors are our own.

Notes

¹ There are some reasons for using a Cobb–Douglas functional form as follows. First of all, neo-classical productivity theory specifies the marginal product of labor to be positive, with a second derivative that is first positive and then it turns to negative after reaching a maximum point. This means that there is an initial range of production that is characterized by increasing returns to the labor input. Ideally, an appropriate production function should be engineered or more scientific in nature to obey well-known scientific laws (Aschauer 1989). Next, the Cobb–Douglas production function is much simpler and easier (Binswanger 1974). Moreover, it is more suitable for a time-series model to use the Cobb–Douglas production (Chamarbagwala, Ramaswamy, and Wunnava 2000). Finally, the Cobb–Douglas function fits almost any observed data, without offering an alternative way to analyze macroeconomic data (Shaikh 1974).

² To save space, we refer the readers to Pesaran and Yamagata (2008) for the details regarding the estimators and for Swamy’s test.

³ For details of these tests, see Kar, Nazlıoğlu, and Ağır (2011).

⁴ We refer to Kónya (2006) and Kar, Nazlıoğlu, and Ağır (2011) for details on how the country- specific critical values are generated in the bootstrap procedure.

⁵ The study was limited to only 10 Middle Eastern countries, as data for other Middles Eastern countries were unavailable.