U.S. Troops and Foreign Economic Growth

Abstract

Do American troops help or hinder economic growth in other countries? We consider a newly constructed dataset of the deployment of U.S. troops over the years 1950–2000 and discover a positive relationship between deployed troops and host country economic growth, which is robust to multiple control variables. Each tenfold increase in U.S. troops is associated with a one–third percentage point increase in average host country annual growth. We explore three possible causal explanations: a Keynesian aggregate demand boost; the diffusion of institutions; and security. Extensive econometric testing, including the use of panel data, confirms the core relationship.

JEL Classification :

• F5
• O2
• O43

Keywords :

• Economic growth
• U.S. military
• National security
• Econometric regressions

ACKNOWLEDGEMENTS

The authors would like to thank Jake Bartolomei, William Beach, Ike Brannon, James Carafano, William Casebeer, James Hamilton, Robert Litan, David Lyle, Greg Mankiw, Donald Marron, Mike Meese, John Nye and seminar participants at West Point for helpful comments. Jon Casale and Christine Leming provided research support. All mistakes are the authors’ alone.

Notes

¹The limiting factor is comparable data on economic growth. For example, since economic growth data on Vietnam is unavailable in the Penn World Tables, that large troop deployment is not included in our analysis. This is especially unfortunate because Vietnam seems to be growing faster than nearby countries Laos and Cambodia which had no comparable U.S. presence.

²Note that the use of log Troops diminishes the power of fast-growing outliers such as Germany which had over 10 million troop-years and Japan which had over four million troop-years.

³The variable log troops was generated by taking the log of Troops plus 1, since log(0) is negative infinity and log (1) is zero. We utilize base 10 logs rather than natural logs so that estimates of the troops-growth relationship can be interpreted in order-of-magnitude terms. Note that whatever log root is used will have the exact same econometric effect.

⁴Likewise, one extra year with over 1000 soldiers has roughly the same effect.

⁵Hoover and Perez (2004) offer a qualified endorsement of extreme bounds in cross-country growth regressions, while Temple (2001) is more enthusiastic. Both offer good discussions of the methodological issues involved.

⁶Sala-i-Martin’s top 21 variables includes the following, in order: Equipment Investment +, Number of Years Open Economy +, Fraction Confucian +, Rule of Law +, Fraction Muslim +, Political Rights +, Latin America Dummy –, Sub-Saharan Africa Dummy –, Civil Liberties +, Revolutions and Coups –, Fraction of GDP in Mining +, Std. Dev. of Black Market Premium –, Fraction of GDP in Primary Exports in 1970 –, Degree of Capitalism +, War Dummy –, Non-Equipment Investment +, Absolute Latitude +, Exchange Rate Distortions –, Fraction Protestant –, Fraction Buddhist +, Fraction Catholic –.

⁷Including a war dummy does shed light on the natural question of whether there are different types of troops, or if their impact varies based on the local environment. Superficially, the regression suggests that the presence of the American military is more useful than peace for enhancing pro-growth institutions. A better interpretation might be that peace is important, but not necessary or sufficient at promoting economic growth.

⁸USAID maintains an online record at http://qesdb.cdie.org/gbk/index.html of all economic and military aid sent overseas, by country and by year, from 1946 to the present.

⁹Nonlinearities may exist for non-troops variables, but the exploration of nonlinearities is both uncommon in empirical growth research and beyond the scope of this paper, the first on the topic of the troops-growth relationship.

¹⁰Easterly (2002) describes the optimism about Ghana, which in 1957 became the first nation in sub Saharan Africa to achieve independence. He writes that “many of the world’s economists [believed that] assistance to Ghana would yield very high returns.” The head of the World Bank’s economics department believed in 1967 that Ghana had a potential annual economic growth rate of 7%.