Time Varying Determinants of US Demand for Defense Spending in the post-Cold War Era

ABSTRACT

We adopt the maximum entropy bootstrap methodology in a rolling window framework in order to investigate the time varying determinants of the US demand for defense spending. Our results based on annual data between 1967 and 2018 show that the US defense demand is mainly driven by lagged military burden, economic growth, GDP share of non-military government spending, election cycle, relative costliness of defense as well as the Russian and the Chinese military burdens. Moreover, the signs and the magnitude of the coefficients show significant variations throughout the sample period. The results also provide a strong evidence of the rising rivalry between the US and China, reflecting the developments in the world economy and the global military arena in the last two decades.

KEYWORDS:

• Demand for defense spending
• US military burden
• post-Cold War
• China
• Russia
• maximum entropy bootstrap

JEL CLASSIFICATION:

• F50
• H56
• O51
• C22

Acknowledgments

We are truly grateful for the anonymous reviewers for their extremely useful comments and suggestions.

Disclosure statement

No potential conflict of interest was reported by the authors.

Notes

1. For an extension of the alliance theory using Joint Product Model; see Sandler (1977). For a recent empirical application of alliance theory with extension of spatial weight given to countries based on NATO membership, US contiguity, and geographical propinquity; see George and Sandler (2017).

2. For more information and the use of Security Web concept, readers can refer to Dunne & Perlo-Freeman (2003), Dunne et al. (2008), and Abdelfattah et al. (2014).

3. The effects of incumbency as well as different electoral cycles such as the US Senate elections are considered to be outside the scope of the present paper. For a more political economics oriented approach, the reader is referred to Zuk and Woodbury (1986) and Bove et al. (2017).

4. It is important to note that the US military expenditures can also affect Russian and Chinese military expenditures. However, the instrumental variable approach for taking into account these effects typically requires a larger data set, which is not possible in a rolling window analysis. As a result, such potential endogeneity issues are relegated to a future study.

5. R version 3.5.1 was used. The code and the data are available from the authors upon request.

6. Lag military burden of 2 and 3 years also has been considered in separate estimations. Although similar results were observed, a one-year lag is used in the final model based on the Akaike Information Criterion, and also in order to maintain a parsimonious model important for rolling window analysis.

7. We have also considered different start/end dates and observed that the signs and the magnitudes of the coefficients as well as the confidence intervals exhibit minor changes. We do not report these results for brevity.