Reward, punishment or inducement? US economic and military aid, 1946–1996

Abstract

If the US has wielded power as global hegemon, then there should be evidence of a linkage between American resources and the ability to influence behavior. However, there is widespread disagreement on how this power is manifested (see Krause, 1991). Methodological and epistemological issues have hampered empirical studies of US hegemonic behavior. For example, does the US reward past behavior or does it offer inducements for future behavior? We document and discuss these issues in terms of the aid–foreign policy compliance nexus. The empirical portion of our paper tests whether US military and non‐military aid are correlated with foreign policy similarity. The main merits of our study are that: (1) we test a new measure of foreign policy similarity developed by Signorino and Ritter (1999) and compiled by Gartzke et al. (1999); (2) we relax the causality issue and test using vector autoregression (VAR) for 76 developing countries; and (3) our time horizon is a minimum of 30 years for each country. Our central finding is that aid is most often used as a reward. More specifically, foreign policy similarity leads to greater economic aid for most African countries, foreign policy similarity leads to greater military aid for a majority of Latin American states, and there are no overarching patterns for Asia and the Middle East. We discuss the implications of our findings in the context of the bargaining and structural approaches to statecraft.

Keywords:

• US economic and military aid
• Reward
• Punishment

Notes

‡ E‐mail: [email protected]

† Corresponding author. E‐mail: [email protected]

An earlier version of this paper was presented at the Annual Meeting of the American Political Science Association, Boston, MA, 29 August–1 September 2002. We acknowledge the support of Grant R‐6493 from the University of Canterbury and a Research Enhancement grant from Texas State University. We appreciate the comments and assistance of Richard Tucker and Norrin Ripsman. We also appreciate the research assistance of Rayana Gonzales.

† Corresponding author. E‐mail: [email protected]

Although bivariate VAR is criticized for excluding other control variables, the method has been commonly used based on the objective of the study. Our goal is not to maximize explanation of variation in aid but rather to use new data and a compelling method to take a novel look at the reward/inducement puzzle.

According to Lai and Morey (2002), S also outperforms measures of similarity based purely on alliance portfolios. Thus, according to them, S is a theoretical improvement in the measurement of foreign policy similarity.

Since the Granger causality test is based on the F‐statistic, which does not indicate the sign, positive or negative, the impulse response function is often used. In this case, however, the impulse response function is not necessary for both theoretical and practical reasons. Theoretically, it does not make sense that the US will punish a country after the country supported US policies in the UN. If the US does not reward the country, the country is unlikely to continue to support US policies in the UN. In fact, the country is likely to oppose US policies to induce rewards from rivals of the US. By the same logic, a country is unlikely to oppose US policies after receiving aid, and the US will not provide continuous aid to countries that do not support the US. Practically, including 76 impulse response graphs will take a lot of space. Thus, we report only F‐statistics and make the impulse response graphs available upon request.

Additional information

Notes on contributors

Karl Derouen, Jr †

‡ E‐mail: [email protected] † Corresponding author. E‐mail: [email protected] An earlier version of this paper was presented at the Annual Meeting of the American Political Science Association, Boston, MA, 29 August–1 September 2002. We acknowledge the support of Grant R‐6493 from the University of Canterbury and a Research Enhancement grant from Texas State University. We appreciate the comments and assistance of Richard Tucker and Norrin Ripsman. We also appreciate the research assistance of Rayana Gonzales.