Defense Spending and Economic Growth around the Globe: The Direct and Indirect Link

ABSTRACT

In this study, we combine both demand- and supply-side models to investigate the direct and indirect link between defense spending and economic growth. With data from 161 states for 1990–2012, we conduct a comprehensive examination of the defense-growth nexus in the post-Cold War world. Our findings suggest that at the global level, military spending significantly dampens private investment but somewhat reduces unemployment. However, the direct economic effects of defense spending on growth are minimal at best.

KEYWORDS:

• Defense spending
• economic growth
• employment
• investment
• security

Acknowledgments

We would like to thank Marc Ethridge; John Bohte; two anonymous reviewers; and Gerald Schneider, the editor of International Interactions, for their helpful comments and suggestions. All the remaining errors are ours.

Notes

¹ In many countries, security policy is regarded as the province of the central government because it not only has (almost) total control of the military but also takes full responsibility in providing security. Since security issues also involve economic issues, both political elites and the public agree that it is better for the central government to control these issues (Murdock 1977:70–71).

² Lee and Chen (2007) report a positive relationship between defense spending and economic growth for OECD countries but a negative relationship for 62 non-OECD states they examined. Alptekin and Levine (2012) also find a positive relationship between the two variables, but only in developed countries.

³ Sevastianova (2009) as well as Aizenman and Glick (2006) argue that defense spending has harmful effects on economic growth, but defense increases in response to external conflict help economies.

⁴ Exceptions are Mintz and Huang (1990, 1991), Heo (1999), and Heo and Eger (2005).

⁵ Some studies indicate a nonlinear relationship between defense spending and economic growth (for example, Pieroni 2009; Menla Ali and Dimitraki 2014).

⁶ Mintz and Huang (1990, 1991) and Heo and Eger (2005) included capital stock in their investment models, analyzing the defense-investment nexus in the United States. However, we could not create a capital stock variable for this study. The only available capital stock data for the world are the Penn World Tables (PWT 8.1). We cannot use this data set for the following reasons: (1) the PWT data cover 147 of the 161 countries included in our study between 1990 and 2012 (data for 2012 are missing); (2) capital stock data in the PWT are expressed in terms of purchasing power parity (PPP) values, while the data we used for analysis are in constant US dollars. Since we are unable to express all of our economic variables in terms of PPP values, including capital stock in our analysis would introduce significant bias. That said, we ran the analysis with capital stock and show the results in Table A1 in the appendix. Although the effects of some control variables change, our major findings concerning the impact of military expenditures on investment are similar.

⁷ Both studies are from Nincic and Cusack (1979). These studies compared a number of jobs created by $1 billion of expenditures in the United States. The results show that defense spending is a poor mechanism for job creation.

⁸ Those who argue against the reduction of defense spending emphasize the immediate unemployment effects of defense spending, while those who prefer cutbacks in defense spending focus more on its long-term effects on overall economic growth.

⁹ DeGrasse (1983) compared the employment data from the US Labor Department’s employment requirements table, including both direct and indirect employment. The author finds that seven of the 11 defense industries selling the largest amount of goods to the military create fewer jobs per dollar than the median civilian manufacturing industry. Seven of the nine largest military suppliers generate lower employment per dollar than the median nonmanufacturing industry. More importantly, the three largest defense manufacturing industries—those accounting for over 40% of the Pentagon’s total purchases from the private sector—employ fewer people per dollar than the median manufacturing industry does.

¹⁰ A survey of the structure of labor market is provided in Nickell (1990).

¹¹ There are two reasons for using a lagged military expenditure term. First, military procurement contracts may be spread over several years. Thus, their impact on the labor market may be delayed. Second, as discussed in the theoretical explanation section, defense expenditures may have a lagged negative effect because of their opportunity cost.

¹² Most defense expenditures go to salaries for military personnel, military procurement (contract awards), and maintenance of weapons systems. Contract awards generate aggregate demands and salaries paid to military personnel; those who work in the weapons industry create more purchasing power in the market. Thus, when the economy is not in full utilization of its resources, defense spending will stimulate economic growth.

¹³ The SIPRI’s military expenditure data set claims to include 171 countries. However, these 171 countries include states that no longer exist (such as the former Yugoslavia, the Soviet Union, and Czechoslovakia, etc.). Appendix Table A1 offers a complete list of these 161 states and their most recent military expenditures both in current values and as a percentage of their GDPs.

¹⁴ The SIPRI military expenditure data are available at http://portal.sipri.org/publications/pages/expenditures/country-search.

¹⁵ The WDI data are available at http://data.worldbank.org/data-catalog/world-development-indicators.

¹⁶ To confirm the stationarity of each variable included in our analysis, we also conducted other unit root tests, including the Levin-Lin-Chu (LLC) unit root test (2002), the Harris-Tzavalis (HZ) unit root test (1999), the Breitung test (Breitung 2000; Breitung and Das 2005), the Im-Pesaran-Shin (IPS) test (2003), and the Phillips-Perron (PP) test where applicable (some variables have missing data or unbalanced panel issues). All these tests produced consistent outcomes.

¹⁷ This result might have occurred due to the omission of capital stock in the model specification. With capital stock (in PPP value instead of constant US dollars) in the model, economic growth in the previous year shows a positive and significant effect (see Table A2 in the appendix). This change in the sign of the previous economic growth variable implies the theoretical importance of including capital stock in the investment model for empirical analyses.

¹⁸ According to fertility growth theories, increases in family size reduce the average investment parents make in each child; this lowers the quality of human capital. This phenomenon eventually hurts economic growth (see Heo and Hahm 2015; Krieckhaus 2006).

¹⁹ See Table A1 in the appendix for each state’s portion of military expenditures to their GDP.