Content uploaded by Uk Heo
Author content
All content in this area was uploaded by Uk Heo on Aug 31, 2016
Content may be subject to copyright.
Defence and Peace Economics, 2004,
Vol. 15(5), October, pp. 453–470
ISSN 1024-2694 print: ISSN 1476-8267 online © 2004 Taylor & Francis Ltd
DOI: 10.1080/1024269042000222392
REWARD, PUNISHMENT OR INDUCEMENT?
US ECONOMIC AND MILITARY AID, 1946–1996
KARL DEROUEN, JRa,† and UK HEOb,‡
aDepartment of Political Science, University of Alabama, Tuscaloosa, AL 35487-0213, USA;
bDepartment of Political Science, University of Wisconsin-Milwaukee, PO Box 413,
Milwaukee, WI, USA
Taylor and Francis Ltdgdpe041006.sgm (Received 13 October 2003; in final form 1 April 2004)
10.1080/1024269042000222392Defence and Peace Economics1024-2694 (print)/1476-8267 (online)Original Article2004Taylor & Francis Ltd0000002004KarlDerouen JrDepartment of Political ScienceUniversity of CanterburyPrivate Bag 4800ChristchurchNew Zealandkarl.fpdm@lycos.com
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
INTRODUCTION
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. In his influential work on
the UN, Robert Keohane (1966, p. 19; see also Wang, 1999) argued that the “more dependent
a state is on a great power for trade, aid, or protection, the more responsive it is likely to be to
pressure.” Keohane later broadened his notion of influence to entail the concept of asymmet-
rical interdependence (Keohane and Nye, 1977). Asymmetrical interdependence exists when
one country benefits less than another in a relationship that entails reciprocal flows of money,
goods or people. This advantage gives the dominant state the potential to influence outcomes.
†Corresponding author. E-mail: karl.fpdm@lycos.com
‡E-mail: heouk@uwm.edu
An earlier version of this paper was presented at the Annual Meeting of the American Political Science Associa-
tion, 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.
454 K. DEROUEN, JR AND U. HEO
Questions of dependence and dependency were at the forefront of political science research
in the 1970s and 1980s (see Caporaso, 1978a,b). Various forms of dependence have been
linked to foreign policy compliance. Scholars looked at trade dependence (e.g. Menkhaus and
Kegley, 1988; Richardson, 1976, 1978; Richardson and Kegley, 1980); aid dependence (e.g.
Armstrong, 1981; Rai, 1980; Roeder, 1985; Wang, 1999; Wittkopf, 1973); military depen-
dence (e.g. Armstrong, 1981; Menkhaus and Kegley, 1988; Roeder, 1985; Wang, 1999); and
foreign investment dependence (e.g. Richardson, 1978). Much of the work on this subject took
paths that frequently did not cross and often generated discrepant findings.
Since the US re-emerged as the hegemon, as a consequence of the end of the Cold War, it
is an appropriate time to re-evaluate the ability of the US to link resources to foreign policy
outcomes (Wang, 1999). Before presenting our empirical model, we briefly discuss prior stud-
ies in this vein and explore the theoretical origins of the reward-inducement concepts. After
testing our model of aid and compliance, we discuss the theoretical implications of our findings.
PRIOR STUDIES
Richardson (1976) treated compliance with US foreign policy as a dependent variable.
Richardson finds that export dependence is inversely related to compliance and that the use of
Cold War (salient) votes leads to stronger relationships between dependence and compliance.
In his extended work, Richardson (1978) showed that compliance deteriorated for much of
the Assembly, not just dependents, from 1950 to 1973. Latin American dependencies were
most often in agreement with the US. Richardson (1978, pp. 71–87) proposed four different
responses to hegemonic demands. First, compliance is termed linear when there is a positive
linear correlation between economic dependence and compliance. Second, decreasing
marginal compliance can result if the dependent government maintains compliance to a certain
point, and then diminishes as certain issues prove too incompatible with its wishes. Defiance
can be is fueled by resentment of dependence and the foreign presence of MDCs. The depen-
dent state complies up to a certain point whereupon increases in dependence lead to antipodal
foreign policy stances. Scholars have yet empirically to fit particular compliance situations
into these propositions.
Wittkopf (1973, p. 887), however, complained of the difficulties in determining whether US
aid was closed linked to the foreign policy compliance among dependents. He correlates
compliance data with aid data from the same year, the year before and the year after to
determine whether the dominant state was attempting inducement (in the first two cases),
reward (in the second case) or punishment (in the third case). His findings were largely incon-
clusive but he suggested that the US used aid politically.
Richardson and Kegley (1980) disaggregated the dependencies of the US by regressing
voting agreement of each dependent from 1950 to 1973 with export dependence, and the
dependent’s voting agreement with the US from the previous year. They found that voting
agreement of dependencies with the US was higher than that of non-dependencies. They also
showed a trend of decreasing compliance with the US since the 1960s.
Armstrong (1981) yielded powerful positive correlations between trade importance and
compliance, and military aid and compliance. By using different forms of compliance,
Armstrong showed that salient (Cold War) votes elicit greater compliance. Ray (1981) corre-
lated salient and non-salient votes with the dependence of the Latin American states upon the
US for 1960 and 1970. He found no evidence of the political use of aid by the US in Latin
America.
Defiance refers to a situation in which the dependent’s compliance decreases as a result of
growing resentment, it is assumed that, over time, compliance will take the form of a parabolic
US ECONOMIC AND MILITARY AID 455
curve (Richardson 1978, p. 147). Deteriorating compliance entails a temporal component –
essentially, compliance decreases as the dependent develops rapport with other LDCs and the
populace of the dependent state populace becomes opposed to the dominant state (Richardson,
1978, p. 152).
Defiance and deteriorating compliance may themselves be symptoms of the paradox of power
as described by Maoz (1989). Compliance is expected of dependents – that is dominant states
expect dependent states to make a sacrifice. Thus, if the dependent’s foreign policy preferences
are antipodal to those of the dominant state, or compliance decreases over time, “a causally
induced contradiction between expectations and the consequences of behavior resulting from
them” may have occurred (Maoz, 1989, p. 240). Biddle and Stephens (1989, pp. 412–414)
describe such a situation in Jamaica from 1962 to 1982. They note that while the domestic
business elite supported a pro-US foreign policy throughout the period, the Jamaican govern-
ment took extreme anti-US and anti-TNC stances from 1972 to 1980. The US was unable to
supplant the anti-imperialist foreign policy orientation of Jamaica as evidenced by UN votes
on salient issues in the early 1970s.
Menkhaus and Kegley (1988) characterized the relative differences between military
import, export, and aid dependence. Kegley and Hook (1991) focused on Reagan’s 1986
explicit attempt to tie aid to UN votes. They found that threats of economic sanctions did not
lead to greater compliance in the 1980s.
Wang (1999), on the other hand, argues that one of the reasons for not finding a systematic
pattern of relationship between US aid distribution and UN vote similarity is because all
issues are not important to the US. Based on this argument, he investigated how US aid affects
aid recipients’ UN votes using a pooled cross-sectional and time-series research design for
1984–1993. He found that compliance levels lead to greater aid.
Summarizing, significant methodological and theoretical advances have been made over
the decades. However, previous studies still have not provided clear evidence of whether the
US uses aid as an inducement or reward. To this end, we empirically reassess the linkage
between US aid and foreign policy similarity using a more sophisticated method, vector
autoregression. We use new data on foreign policy similarity compiled by Gartzke et al.
(1999) and based on a method from Signorino and Ritter (1999). In the research design section
that follows, we discuss that two otherwise distinct literatures agree on the point that the US
is capable of linking resources and compliance. We then move to a test of US aid and foreign
policy similarity for 76 countries.
RESEARCH DESIGN
Keohane (1966, p. 18) asserts that, “the more dependent a state is on a great power for trade,
aid, or protection, the more responsive it is likely to be to pressure.” The reason, according to
him (Keohane, 1966, p. 19), is that, “it is sufficient that the smaller state is aware that ‘Big
Brother’ is watching.”
Based on this logic, Keohane and Nye (1977) and Krause (1991) develop an argument that
a strong state can link its policies on certain issues to weaker states’ policies on other issues.
In other words, the dominant state is considered capable of rewarding or inducing compliant
behavior and punishing non-compliant behavior in the peripheral state.
Menkhaus and Kegley (1988, p. 322) further the argument by contending that the unit of
analysis chosen to study this issue has implications for discerning causality of dependence. By
using the dependent state-centric approach “the ability of alternate types of economic depen-
dence to account for variations in the state’s compliant behavior can be estimated by a model
that permits comparisons across types to be made” (Kegley and Menkhaus, 1989, p. 322).
456 K. DEROUEN, JR AND U. HEO
Therefore, the probability of determining whether inducement, reward or punishment is at
work is increased. The authors also state that “since competing dominant states will vary in
the degree and nature of their relationship with a weak state (one may trade heavily with it,
whereas another may primarily employ aid or military assistance to influence the weak state),
this circumstance, when it arises, provides an opportunity to compare the relative influence of
these variables” (Kegley and Menkhaus, 1989, p. 322).
According to Wang (1999, pp. 200–201), most previous studies on this issue have two
common shortcomings: (1) the time period covered in the analysis is short; (2) when the
period is long, the statistical method used in the analysis is simple bivariate correlations or
regressions rather than sophisticated time-series methods. Following Wang (1999), we
employ a longer time-series data (1946–1996) and a sophisticated time-series method,
Vector Autoregression (VAR). Since VAR is a dynamic method capturing all the individual
countries’ UN voting behavior as well as changes in US military and economic aid, we
employ an individual country time-series design. This approach will allow us to determine
whether US aid is used as an inducement, punishment, or reward for political purposes.
Moreover, the findings of this study will reveal what types of strategy the US use towards
other countries.
Steven Lukes (1974) describes a form of power that is directly relevant to this analysis. The
core of his argument is that observable conflict need not necessarily be manifest when power
is wielded. Power might not be used by a hegemon in a simple off-on fashion. Instead, it
might be more of a continual phenomenon that is maintained over time, i.e. “power over”
rather than “power to” (Lukes, 1974, p. 31). In short, a hegemon can attain “compliance by
overcoming or averting … opposition” (Lukes, 1974, p. 31). The implication for our analysis
is clear. Because power in this sense can be manifested over time and in the absence of
observable conflict, it makes sense to look at the broad dynamics motivating aid such as
inducement or reward. The US might not only be giving aid based on discrete events
surrounding observable conflict, but rather might use aid to maintain power. By not consider-
ing power in this context, one might risk mistakenly characterizing a relationship between a
hegemon and another state as one of consensus if foreign policy similarity is high and there is
no observable conflict.
Method
As discussed thus far, the direction of causality between aid and foreign policy compliance
is unclear because the US may punish, reward, or offer inducement. For this reason, we
employ VAR using Granger causal analysis since it makes no a priori assumption about
causal direction. The concept of causality used by the Granger analysis in VAR is empiri-
cal rather than theoretical. It requires two conditions for X to cause Y: changes in X must
precede changes in Y and X must help predict Y. In other words, if incorporating the
history of X can predict Y better than the past of Y alone, then it can be said that variable X
causes variable Y. Since this concept is slightly different from what we usually mean by
causality, we use the term “Granger cause” (see Freeman, 1983). We carry out our tests on
76 developing countries. We only tested countries that have received aid and had been in
existence for at least 30 years prior to 1996. Since theoretical guidance is not available and
there is a lack of a priori information about the proper lag length, we rely on standard
information criteria suggested by Schwarz (1978) and recommended by Geweke and
Meese (1981).
While some might argue that our method and simple model lack depth and potentially
important controls, Wang (1999, p. 207) shows that democracy level, military strength and
GNP of a recipient state are not important determinants of a developing country’s UN vote
US ECONOMIC AND MILITARY AID 457
coincidence with the US.1 Moreover, the objective of this study is testing whether US aid
is used as an inducement or reward for foreign policy compliance. Adding a suite of
control variables would prevent us from doing a dynamic VAR analysis due to degrees of
freedom and lack of feedback relationship between our endogenous variables and control
variables.
VAR is also a compelling method in terms of Lukes’ (1974) three-dimensional view of
power mentioned above. If we accept that conflict between the US and developing states need
not be observable or discrete, then it follows that aid cannot be looked at only in the short-term.
VAR allows us to relax causality and allow for the presence of latent conflict. The US might
be controlling the agenda in its relations with developing states by providing aid years before
or after certain forms of compliance are manifested. In other words, aid may not be directly
tied to an observable conflict between the US and another state, but rather it may be used by
the US to prevent the other state from even identifying its real long-term interests. The US may
be using power in the form of aid to prevent “conflict from arising in the first place” (Lukes,
1974, p. 23). Aid may be used to shape “…perceptions, cognitions, and preferences in such a
way that they accept their role in the existing order of things …” (Lukes, 1974, p. 24). VAR
will allow us to assess these claims.
Foreign Policy Similarity
Richardson (1981, p. 102) defines compliance as “one party acceding to the preferences of a
second party, thereby acting contrary to what it would have done in the absence of the second
party’s influence. In other words, compliance is a sacrifice, wherein actors abandon their
preferences as they conform to another’s dissimilar foreign policy wishes.” Keohane (1966,
pp. 5–6) justifies using UN data to measure foreign policy influence as “political influence in
the General Assembly … [is] the ability of a state or a group of states to achieve its ends by
altering the behavior of other states without the exertion of physical force”.
Foreign policy similarity data are taken from Gartzke et al.’s (1999) Similarity of UN Policy
Positions project (SUN). SUN data are created using UN roll-call votes on resolutions in the
UN General Assembly. Since the UN General Assembly discusses and passes a number of
resolutions annually, Gartzke et al. using Signorino and Ritter’s (1999) S-score method,
created an annual data set indicating how each nation has voted in the UN General Assembly
and how similar votes are between certain nations. In other words, if nation A voted the same
way that the US did repeatedly in the same year, the S-score will indicate that their similarities
are high. We consider that Nation A was compliant in terms of US foreign policy.
Signorino and Ritter (1999) describe their measure (known as S score) in technical detail.
They discuss how their measure is superior to tau-b used in previous studies. The reason is that
the tau-b score is a measure of association. Therefore, tau-b, in estimating ordinal data, is a
measure of rank-order correlation, which is not sufficient to measure “similarity”. According
to Signorino and Ritter (1999, p. 121),
Kendall’s tau-b reflects the extent to which states i and j rank their alliance commitments to paired members of
the international system in the same order, whereas we would like to measure the extent to which states i and j
have the same type of alliance commitments to each of the individual members of the international system.
1 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.
458 K. DEROUEN, JR AND U. HEO
In other words, not all alliance portfolios can be measured with an association measure such
as tau-b. Thus, Signorino and Ritter (1999, pp. 126–127) develop the similarity S of states i’s
and j’s policy portfolios Pi and Pj respectively as follows.
where d(Pi, Pj, W, L) represents the distance between the points Pi and Pj, L denotes the scoring
rules; and W indicates dimension weights. Intuitively, d(Pi, Pj, W, L) becomes 0 when states i
and j have the identical alliance portfolio whereas d(Pi, Pj, W, L) becomes equal to dmax (W,
L) when their policy portfolio is as far apart as possible. Thus, S scores range from 1 to −1. A
score of 1 between two nations’ policy portfolio means that they are in near-perfect foreign
policy commitment and −1 represents complete discordance between the two nations. Inter-
ested readers can consult Signorino and Ritter (1999) for examples of S using UN votes.2
Economic and Military Aid
Bilateral (non-military) loans and grants from a core country to a peripheral one can lead to
foreign aid dependence. Richardson (1978, pp. 41–54, 46) notes that the immediate “purpose
of aid is to supply recipients with an ability to pay for imports that they cannot afford solely
on the basis of export earnings and capital transfers by direct foreign investment.” The aid is
needed by the poor country to enable it to pay for development – thus, the donor country
controls something desired by the recipient country, and can therefore be used as a reward or
inducement by the donor. Aid can be used to induce or reward foreign policy compliance, but
it is also often tied to stipulations that the recipient uses the aid to buy donor goods (examples
of studies investigating the foreign aid–compliance nexus include Wittkopf (1973), Rai
(1980), and Menkhaus and Kegley (1988).
Some states are thought to comply with the wishes of core states because of the presence of
military dependence. Krause (1991, p. 314) asserts that the US has used military aid as a vital
foreign policy tool since the Second World War. Moon (1983) shows that military grants are
the best predictor of UN voting in both cross-sectional and longitudinal studies, although he
does not attribute this to the presence of any short-term bargaining, but rather, regime changes
and/or changes in the dependent’s relationship with the US (see also Armstrong, 1981).
US foreign aid and military aid data used in this study are from the US Agency for Interna-
tional Development (USAID, various years). We use the USAID’s measure of “total economic
assistance loans and grants”. These data include loan and grant commitments made by USAID
and its precursors, and all Food For Peace loans and grants. It also includes ‘other economic
assistance loans and grants’. Military aid consists of USAID’s “total military assistance loans
and grants”. Essentially, this entails Military Assistance Program grants, Foreign Military
Credit Financing, International Military Education and Training Program, Transfers of Excess
Defense Articles, and Other grants. The appendix describes in detail the economic and military
aid used in this analysis.
FINDINGS
The VAR with Granger test results are summarized in Table I by region, and individual results
are reported in Table II. Overall, the results are fairly convincing: 65 out of 76 countries show
2 According to Lai and Morey (2002), S also outperforms measures of similarity based purely on alliance portfo-
lios. Thus, according to them, S is a theoretical improvement in the measurement of foreign policy similarity.
SP P W L dP P W L
dWL
ij
ij
(,,,) (,,,)
(,)
max
=−12
US ECONOMIC AND MILITARY AID 459
a significant Granger causal relationship between aid and foreign policy similarity based on
the F-statistic. The most common pattern is that similarity leads to greater aid. Aid leads to
greater similarity in 27 cases. This implies that the US is more likely to use rewards rather than
inducements.3
In terms of regional pattern, Africa shows the most consistent pattern between aid and
foreign policy compliance. In 16 out of 30 countries, similarity in UN voting “Granger
caused” economic aid. This means that the US has heavily used economic aid as a reward to
supporting the US in the UN. The US has also used economic aid to draw support in Africa on
the grounds that economic aid “Granger causes” similarity in UN voting in seven countries.
The strong relationship between economic aid and similarity in UN voting may, in part, be
because most African countries are in dire need of economic aid.
In contrast, the US used military aid as a foreign policy compliance tool for Latin American
countries more than economic aid. In 11 out of 21 Latin American countries, the US rewarded
military aid for their support in UN voting, and induced the support in five countries.
Economic aid was also used for reward in eight countries and for an inducement in one
country. This may be due to the military authoritarianism common in Latin America for much
3 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.
TABLE I Summary by Region
Region Causal direction Number of significant
relationships
Latin America SUN→USMA 11
USMA→SUN 5
SUN→EA 8
EA→SUN 1
Africa SUN→USMA 5
USMA→SUN 0
SUN→EA 16
EA→SUN 7
Middle East SUN→USMA 3
USMA→SUN 2
SUN→EA 5
EA→SUN 3
Asia SUN→USMA 3
USMA→SUN 3
SUN→EA 4
EA→SUN 6
Countries showing no significant relationship:
Latin America: Mexico, Paraguay
Africa: Congo, Uganda, Tanzania, Zambia, Lesotho
Middle East: Sudan, Egypt, Israel
Asia: Pakistan
460 K. DEROUEN, JR AND U. HEO
TABLE II Results of the Vector Autoregression by Country and Region
Nation (period) Lag Causal direction F-statistics
Latin America
Haiti (46–96) 2 SUN→USMA 2.85*
USMA→SUN 1.41
SUN→EA 1.51
EA→SUN 0.55
Dominican Republic (46–96) 5 SUN→USMA 3.27***
USMA→SUN 0.64
SUN→EA 0.76
EA→SUN 0.75
Jamaica (62–96) 2 SUN→USMA 2.75*
USMA→SUN 0.24
SUN→EA 11.13***
EA→SUN 1.34
Trinidad and Tobago (62–96) 8 SUN→USMA 3.27**
USMA→SUN 0.84
SUN→EA 0.28
EA→SUN 0.41
Mexico (46–96) 7 SUN→USMA 0.51
USMA→SUN 1.12
SUN→EA 1.30
EA→SUN 0.29
Guatemala (46–96) 2 SUN→USMA 0.24
USMA→SUN 1.05
SUN→EA 3.70**
EA→SUN 0.54
Honduras (46–96) 2 SUN→USMA 0.33
USMA→SUN 0.76
SUN→EA 2.57*
EA→SUN 0.23
El Salvador (46–96) 2 SUN→USMA 1.31
USMA→SUN 0.27
SUN→EA 3.78**
EA→SUN 0.27
Nicaragua (46–96) 4 SUN→USMA 2.68**
USMA→SUN 3.60***
SUN→EA 1.00
EA→SUN 0.40
Panama (46–96) 6 SUN→USMA 2.83**
USMA→SUN 2.07*
SUN→EA 0.92
EA→SUN 0.13
Colombia (46–96) 3 SUN→USMA 2.27*
USMA→SUN 0.18
SUN→EA 1.31
EA→SUN 0.24
US ECONOMIC AND MILITARY AID 461
TABLE II Continued
Nation (period) Lag Causal direction F-statistics
Venezuela (46–96) 4 SUN→USMA 2.09*
USMA→SUN 0.34
SUN→EA 5.11***
EA→SUN 0.66
Guyana (66–96) 5 SUN→USMA 1.00
USMA→SUN 3.41**
SUN→EA 2.11
EA→SUN 2.40*
Ecuador (46–96) 4 SUN→USMA 0.45
USMA→SUN 1.15
SUN→EA 2.54**
EA→SUN 0.80
Peru (46–96) 2 SUN→USMA 0.34
USMA→SUN 0.47
SUN→EA 5.85***
EA→SUN 0.09
Brazil (46–96) 4 SUN→USMA 2.61**
USMA→SUN 2.85**
SUN→EA 1.53
EA→SUN 0.68
Bolivia (46–96) 2 SUN→USMA 2.27
USMA→SUN 0.86
SUN→EA 6.74***
EA→SUN 0.16
Paraguay (46–96) 5 SUN→USMA 1.11
USMA→SUN 0.14
SUN→EA 0.90
EA→SUN 0.60
Chile (46–96) 2 SUN→USMA 3.32**
USMA→SUN 0.05
SUN→EA 1.17
EA→SUN 0.72
Argentina (46–96) 6 SUN→USMA 1.17
USMA→SUN 1.94*
SUN→EA 0.96
EA→SUN 0.55
Uruguay (46–96) 4 SUN→USMA 2.75**
USMA→SUN 0.69
SUN→EA 0.32
EA→SUN 0.38
Africa
Gambia (65–96) 3 SUN→USMA 1.95
USMA→SUN 0.97
SUN→EA 5.15***
EA→SUN 0.05
462 K. DEROUEN, JR AND U. HEO
TABLE II Continued
Nation (period) Lag Causal direction F-statistics
Mali (60–96) 4 SUN→USMA 1.41
USMA→SUN 0.64
SUN→EA 1.23
EA→SUN 6.94***
Senegal (60–96) 4 SUN→USMA 2.01
USMA→SUN 0.77
SUN→EA 0.88
EA→SUN 8.10***
Benin (60–96) 5 SUN→USMA 0.20
USMA→SUN 0.62
SUN→EA 2.38*
EA→SUN 0.68
Mauritania (61–96) 6 SUN→USMA 0.54
USMA→SUN 0.71
SUN→EA 3.03*
EA→SUN 1.47
Niger (60–96) 2 SUN→USMA 1.58
USMA→SUN 0.87
SUN→EA 11.72***
EA→SUN 0.85
Ivory Coast (61–96) 2 SUN→USMA 4.93***
USMA→SUN 0.19
SUN→EA 1.67
EA→SUN 0.30
Guinea (58–96) 3 SUN→USMA 2.06
USMA→SUN 1.04
SUN→EA 3.19**
EA→SUN 0.29
Liberia (47–96) 3 SUN→USMA 1.04
USMA→SUN 1.11
SUN→EA 2.75**
EA→SUN 2.06
Sierra Leone (61–96) 3 SUN→USMA 1.53
USMA→SUN 1.53
SUN→EA 2.29*
EA→SUN 1.40
Ghana (57–96) 2 SUN→USMA 1.93
USMA→SUN 0.58
SUN→EA 2.66*
EA→SUN 0.27
Togo (60–96) 2 SUN→USMA 1.59
USMA→SUN 0.02
SUN→EA 5.66***
EA→SUN 0.14
US ECONOMIC AND MILITARY AID 463
TABLE II Continued
Nation (period) Lag Causal direction F-statistics
Cameroun (60–96) 4 SUN→USMA 0.78
USMA→SUN 1.66
SUN→EA 0.41
EA→SUN 2.61*
Nigeria (60–96) 2 SUN→USMA 0.11
USMA→SUN 0.90
SUN→EA 1.41
EA→SUN 3.58**
Gabon (60–96) 5 SUN→USMA 0.91
USMA→SUN 1.31
SUN→EA 2.34*
EA→SUN 0.78
Central African Republic (61–96) 2 SUN→USMA 4.03**
USMA→SUN 0.11
SUN→EA 4.66***
EA→SUN 0.25
Chad (61–96) 2 SUN→USMA 1.06
USMA→SUN 0.88
SUN→EA 4.91***
EA→SUN 0.34
Congo (60–96) 6 SUN→USMA 0.46
USMA→SUN 0.58
SUN→EA 0.61
EA→SUN 0.98
Uganda (62–96) 2 SUN→USMA 1.52
USMA→SUN 0.22
SUN→EA 1.64
EA→SUN 0.03
Kenya (63–96) 2 SUN→USMA 0.53
USMA→SUN 1.71
SUN→EA 0.64
EA→SUN 3.84**
Tanzania (61–96) 6 SUN→USMA 1.38
USMA→SUN 0.25
SUN→EA 1.35
EA→SUN 1.11
Burundi (62–96) 4 SUN→USMA 2.55*
USMA→SUN 0.60
SUN→EA 3.55**
EA→SUN 1.77
Rwanda (62–96) 6 SUN→USMA 0.63
USMA→SUN 0.12
SUN→EA 2.34*
EA→SUN 3.59***
464 K. DEROUEN, JR AND U. HEO
TABLE II Continued
Nation (period) Lag Causal direction F-statistics
Somalia (60–96) 4 SUN→USMA 0.53
USMA→SUN 1.82
SUN→EA 0.90
EA→SUN 4.38***
Ethiopia (46–96) 8 SUN→USMA 3.84***
USMA→SUN 0.95
SUN→EA 0.84
EA→SUN 0.31
Zambia (65–96) 3 SUN→USMA 0.53
USMA→SUN 1.70
SUN→EA 0.60
EA→SUN 1.39
Malawi (65–96) 4 SUN→USMA 1.89
USMA→SUN 0.69
SUN→EA 3.76**
EA→SUN 0.69
Lesotho (66–96) 5 SUN→USMA 1.32
USMA→SUN 0.63
SUN→EA 0.17
EA→SUN 2.12
Botswana (66–96) 2 SUN→USMA 2.66*
USMA→SUN 0.10
SUN→EA 2.15
EA→SUN 1.06
Malagasy Republic (60–96) 3 SUN→USMA 2.04
USMA→SUN 0.09
SUN→EA 2.48*
EA→SUN 1.42
Middle East
Morocco (56–96) 2 SUN→USMA 0.20
USMA→SUN 1.18
SUN→EA 7.58***
EA→SUN 0.97
Algeria (62–96) 4 SUN→USMA 2.15*
USMA→SUN 0.20
SUN→EA 0.49
EA→SUN 3.08**
Tunisia (56–96) 8 SUN→USMA 1.20
USMA→SUN 1.05
SUN→EA 0.95
EA→SUN 2.11*
Sudan (56–98) 6 SUN→USMA 1.00
USMA→SUN 0.97
SUN→EA 0.39
EA→SUN 1.79
US ECONOMIC AND MILITARY AID 465
TABLE II Continued
Nation (period) Lag Causal direction F-statistics
Iran (46–96) 3 SUN→USMA 4.17***
USMA→SUN 1.29
SUN→EA 2.45*
EA→SUN 3.24**
Turkey (46–96) 2 SUN→USMA 3.49**
USMA→SUN 1.03
SUN→EA 1.05
EA→SUN 1.30
Iraq (46–96) 4 SUN→USMA 0.13
USMA→SUN 0.06
SUN→EA 9.61***
EA→SUN 0.24
Egypt (46–96) 4 SUN→USMA 1.24
USMA→SUN 0.56
SUN→EA 0.75
EA→SUN 1.55
Syria (48–96) 3 SUN→EA 2.15*
*Syria has received no U.S. military aid. EA→SUN 0.05
Lebanon (46–96) 2 SUN→USMA 0.24
USMA→SUN 3.41**
SUN→EA 3.33**
EA→SUN 2.22
Jordan (56–96) 9 SUN→USMA 0.57
USMA→SUN 2.53*
SUN→EA 0.36
EA→SUN 1.14
Israel (49–96) 2 SUN→USMA 0.48
USMA→SUN 0.01
SUN→EA 1.77
EA→SUN 0.14
Afghanistan (46–96) 4 SUN→USMA 2.68
USMA→SUN 0.41
SUN→EA 1.05
EA→SUN 1.49
Asia
India (48–96) 5 SUN→USMA 3.82***
USMA→SUN 0.96
SUN→EA 0.60
EA→SUN 1.01
Pakistan (47–96) 5 SUN→USMA 0.22
USMA→SUN 0.72
SUN→EA 0.30
EA→SUN 1.17
Burma (48–96) 4 SUN→USMA 1.35
USMA→SUN 2.26*
SUN→EA 0.37
EA→SUN 3.73***
466 K. DEROUEN, JR AND U. HEO
of this study period. The Middle East and Asia do not exhibit overarching patterns, but a
majority of the countries revealed significant relationship between these variables. US strate-
gies may be more ad hoc in these regions.
Interestingly, Israel, Egypt and Mexico do not exhibit patterns. Perhaps high levels of US
aid are not tied to recipient behavior. Israel gets high aid but also votes with US most of the
time. Egypt gets tremendous aid as a result of the Camp David Agreement.
TABLE II Continued
Nation (period) Lag Causal direction F-statistics
Sri Lanka (55–96) 5 SUN→USMA 0.75
USMA→SUN 0.05
SUN→EA 4.10***
EA→SUN 0.01
Nepal (56–96) 3 SUN→USMA 0.32
USMA→SUN 0.07
SUN→EA 3.83**
EA→SUN 0.72
Thailand (47–96) 6 SUN→USMA 2.63**
USMA→SUN 1.07
SUN→EA 0.96
EA→SUN 1.42
Cambodia (56–96) 2 SUN→USMA 0.02
USMA→SUN 11.31***
SUN→EA 0.10
EA→SUN 7.01***
Laos (56–96) 6 SUN→USMA 1.56
USMA→SUN 2.62*
SUN→EA 4.98***
EA→SUN 3.39***
Malaysia (57–96) 2 SUN→USMA 1.15
USMA→SUN 0.11
SUN→EA 7.53***
EA→SUN 3.41**
Singapore (65–96) 3 SUN→USMA 0.46
USMA→SUN 1.55
SUN→EA 0.11
EA→SUN 4.24***
Philippines (46–96) 6 SUN→USMA 3.42***
USMA→SUN 0.67
SUN→EA 1.85
EA→SUN 0.90
Indonesia (50–96) 3 SUN→USMA 0.61
USMA→SUN 1.07
SUN→EA 0.67
EA→SUN 2.21*
Note: SUN = Country A’s similarity with the US in UN Roll Call votes in the General Assembly; USMA = US military aid to Country
A; EA = US economic aid to Country A; * significant at 0.1; ** significant at 0.05; *** significant at 0.01.
US ECONOMIC AND MILITARY AID 467
DISCUSSION
These findings can be discussed in light of an important paradigm distinction in the dependence–
compliance literature. One path follows the traditional realist school notion that dominant states
use power to achieve outcomes. This is the bargaining model. The second model depicts a long-
term dominance relationship based on economic dependence. This model is known in the liter-
ature as the structural model. Our purpose in this discussion is not to distinguish between the
two models. In fact, since both models contain elements of bargaining and political manipu-
lation, some (e.g. Krause, 1991) treat the structural model as an extension of the bargaining
model (although its time-frame is “medium-term” as opposed to “short term”). Rather, we are
concerned with using these models as a heuristic to examine further our empirical findings.
The bargaining model is rooted in the power perspective as it pertains to asymmetrical
interdependence (see Goldstein, 1995; Keohane and Nye, 1977; Lee and Heo, 2001; Palmer,
1990). When two countries are asymmetrically interdependent, one country is in a position to
benefit from trade and other flows more. The country in the position to benefit the least is
dependent to the extent that it costs to compensate for any change in the flow of trade or aid
from the dominant country (Richardson and Kegley, 1980, p. 192). Thus, the dominant state
is in a position to influence the foreign policy preferences of the dependent state through a
bargaining mechanism that either rewards or punishes past behavior, or induces future
behavior.
Most of the dependence–compliance studies mentioned earlier used the bargaining
approach as a point of departure. For instance, Richardson (1976) applies the bargaining
model in order to show that countries dependent upon the US agree with the latter more than
countries not dependent upon the US in the mid-1960s. Rai (1980) analyzes UN vote compli-
ance and finds that the US gives aid as an inducement. Although not always stated explicitly,
Wittkopf (1973), Richardson (1978, 1981), Ray (1981), Armstrong (1981), Roeder (1985) and
Sislin (1994) used the bargaining model approach.
Krause (1991) has also detailed the structural approach in terms of superpower arms trans-
fers. He suggests that a structural relationship can obtain when the superpower uses military
assistance to prevent regional shifts away from the US sphere. This approximates the situation
between the US and Latin America during much of the Cold War – especially after the revo-
lutions in Cuba and Nicaragua. Latin America was of greater strategic value than Africa,
where the US more often responded to specific situations such as Angola during the civil war
or Egypt after Camp David. US military aid to Latin America is also an element of dependency
theory. The US supplies military aid to maintain stability for its capital investments in the
region and to prevent the emergence of socialism (see DeRouen and Heo, 2001).
Thus, it makes sense that military aid is more important in Latin America and that our find-
ings for that region seem to fit within the confines of a structural relationship. The US has
been interested in imposing general regional strategic relations in Latin America since the
Monroe Doctrine. The subsequent use of structural statecraft eventually resulted in Brazil and
Argentina trying to reduce their vulnerability to US military aid and technology through indig-
enous production. The bargaining model would apply in Africa where the US used aid to meet
immediate and specific goals. Again, it is important to keep in mind that both models involve
bargaining and threat-reward behavior. The US’s clear preference for using military aid in
Latin America lends evidence to a structural dynamic and fits in to Lukes’ (1974) conception
of three-dimensional power.
Although the results are strong, interpretation needs to be done with caution because they
are based on bivariate VAR using the Granger Analysis. Factors that may affect economic or
military aid, such as poverty, the Cold War, regime types and so on are not considered in this
analysis due to theoretical and methodological reasons explained earlier.
468 K. DEROUEN, JR AND U. HEO
CONCLUSIONS
We have explored the linkage between US aid and foreign policy similarity using VAR methods
and a new measure of foreign policy similarity based on S scores. Our sample consisted of 76
developing countries between 1946–1996 (varying by country). We find that the US and many
developing countries under investigation in this paper interact with respect to inducing and/or
rewarding aid. For instance, in Africa and Latin America, the US is more likely to reward past
behavior. In general, the US rewards African countries that have foreign policy similarity with
non-military aid. Similarly, the US rewards Latin American states with military aid. In Asia
and the Middle East, patterns are less clear-cut. Asia appears to be the region where aid as
inducement is most common.
Overall, US aid seems to be used as a reward determined by past behavior rather than induce-
ment. This finding is consistent with the recent work of Lai and Morey (2002) who examined
how the US influences votes in the UN and found that aid does not induce votes in the UN.
The reason may be that the US uses aid as a reward more often than inducement. While our
findings cannot assess the key votes in the interesting manner Wang (1999) does, we do
observe that the US seems to induce similarity from important countries. For instance, aid had
a significant Granger causal impact on votes from Argentina and Brazil in Latin America,
Nigeria in Africa, and Indonesia and Malaysia in Asia.
Krause (1991) also reports that superpowers were more likely to use reward strategies when
doling out military aid. He suggests (Krause, 1991 p. 317) that this is because the “costs of
complying with or fulfilling promises of rewards is likely much lower for clients and patrons.”
The use of threats undermines the future of the relationship for the patron, and bowing to a
threat undermines client sovereignty.
A final consideration is that the nature of US hegemony has changed since the Cold War
ended and 9/11. Recent studies have looked at whether US military aid is tied to human rights
in recipient countries (e.g. Blanton, 2001); aid and recipient regime type and level of develop-
ment (e.g. Meernik et al., 1998); and aid and conditionality (e.g. Emmanuel, 2001).
Emmanuel (2001) looks at US aid to Sub-Saharan Africa and finds that, in the post-Cold
War era, the US has tied aid to political reform. He also demonstrates the importance of trade
ties, mineral resources, and humanitarian crises. Blanton (2001) explores the linkage between
US arms transfers and level of democracy in the recipient regime. Controlling for the presence
of US troops, conflict, trade and population, she finds that democracy is an important deter-
minant of US arms transfers in the post-Cold War era. A promising area of research would be
to search for patterns in the wake of the 9/11 terrorist attacks. For example, how did the US
reward countries that supported the ensuing war in Iraq. Not long after the war began, for
example, the US began negotiations to provide trade concessions for Australia. This type of
short-term reward is likely to differ from the three-dimensional view of power conceptualized
by Lukes (1974). It is, in fact, more akin to the second dimension where coercion and threat
are more apparent.
References
Armstrong, A. (1981) The political consequences of economic dependence. Journal of Conflict Resolution 25
401–428.
Biddle, W.J. and Stephens, J.D. (1989) Dependent development and foreign policy: the case of Jamaica. International
Studies Quarterly 33 411–434.
Blanton, S. (2001) Arms transfers, democracy, and human rights: comparing U.S. foreign policy in the Cold War
and post-Cold War periods. Paper presented at the Annual Meeting of the Peace Science Society (Interna-
tional).
Caporaso, J. A. (1978a), Introduction. International Organization 15 1–12.
US ECONOMIC AND MILITARY AID 469
Caporaso, J. A. (1978b) Dependence, dependency, and power in the global system: a structural and behavioral
analysis. International Organization 15 13–43.
DeRouen, K. and Heo, U. (2001) The military and modernisation in Latin America. British Journal of Political
Science 31 475–496.
Emmanuel, N. (2001) Is political conditionality a fallacy? The determinants of U.S. foreign assistance in the Post-
Cold War era. Paper presented at the Annual Meeting of the ISA.
Freeman, J. (1983) Granger causality and the time-series analysis of political relationship. American Journal of
Political Science 27 337–358.
Gartzke, E., Jo, D. and Tucker, R. (1999) The similarity of UN policy positions, 1946–96, Version 1.17. http://
www.vanderbilt.edu/∼rtucker/data/affinity/un/similar.
Geweke, J. and Meese, R. (1981) Estimating regression models of finite but unknown order. International Economic
Review 22 55–70.
Goldstein, A. (1995) Discounting the free ride: alliances and security in the Postwar world. International Organization
49 39–71.
Kegley, C. and Hook, S. (1991) U.S. foreign aid and UN voting: did Reagan’s linkage strategy buy deference or
defiance? International Studies Quarterly 35 295–312.
Keohane, R. (1966) Political influence in the General Assembly. International Conciliation #557.
Keohane, R. and Nye, J. (1977) Power and Interdependence: World Politics in Transition. Boston: Little, Brown.
Krause, K. (1991) Military statecraft: power and influence in Soviet and American arms transfer relationships.
International Studies Quarterly 35 313–336.
Lai, B. and Morey, D. (2002) Aid and influence: the role of domestic politics. Paper presented Annual Meeting of the
APSA. Boston, MA, 29 August–1 September 2002.
Lee, J. and Heo, U. (2001) The U.S.–South Korea alliance: free-riding or bargaining? Asian Survey 41(5) 822–845.
Lukes, S. (1974) Power: A Radical View. London: Macmillan Education.
Maoz, Z. (1989) Power capabilities and paradoxical conflict outcomes. World Politics 42 239–266.
Meernik, J., Krueger, E. and Poe, S. (1998) Testing models of U.S. foreign policy: foreign aid during and after the
Cold War. Journal of Politics 60 63–85.
Menkhaus, K., and Kegley, C. (1988) The compliant foreign policy of the dependent state revisited. Comparative
Political Studies 21 315–346.
Moon, B. (1983) The foreign policy of the dependent state. International Studies Quarterly 27 315–340.
Palmer, G. (1990) Corralling the free rider: deterrence and the Western Alliance. International Studies Quarterly 34
147–164
Rai, K. (1980) Foreign aid and voting in the U.N. General Assembly, 1967–1976. Journal of Peace Research 17
269–276.
Ray, J.L. (1981) Dependence, political compliance, and economic performance: Latin America and Eastern Europe.
In C. Kegley and P. McGowan (eds.) The Political Economy of Foreign Policy Behavior, pp. 111–136. Beverley
Hills: Sage.
Richardson, N.R. (1976) Political compliance and U.S. trade dominance. American Political Science Review 70
1098–1109.
Richardson, N.R. (1978) Foreign Policy and Economic Dependence. Austin and London: University of Texas Press.
Richardson, N.R. (1981) Economic dependence and foreign policy compliance: bringing measurement closer to
conception. In C.W. Kegley Jr. and P.J. McGowan (eds.) The Political Economy of Foreign Policy Behavior, pp.
87–110. Beverley Hills: Sage.
Richardson, N.R. and Kegley, C.W. (1980) Trade dependence and foreign policy compliance: a longitudinal analysis.
International Studies Quarterly 29 191–216.
Roeder, P.G. (1985) The ties that bind: aid, trade and political compliance in Soviet-Third World relations. Interna-
tional Studies Quarterly 29 191–216.
Schwarz, G. (1978) Estimating the dimension of a model. Annals of Statistics 6 461–464.
Signorino, C., and Ritter, J. (1999) Tau-b or not tau-b: measuring the similarity of foreign policy positions. Interna-
tional Studies Quarterly 43 115–144.
Sislin, J. (1994) Arms as influence. Journal of Conflict Resolution 38 665–689.
USAID, (Various years) U.S. overseas loans and grants and assistance from international organizations: obligations
and loan authorizations. Washington DC. Available at http://qesdb.cdie.org/gbk/concepts.html.
Wang, T. (1999) U.S. foreign aid and UN voting: an analysis of important issues. International Studies Quarterly 43
199–210.
Wittkopf, E. (1973) Foreign aid and U.N. votes: a comparative study. American Political Science Review 67 868–888.
470 K. DEROUEN, JR AND U. HEO
APPENDIX
These definitions of U.S. economic and military aid are taken from USAID. The selected
descriptions below are taken directly from http://qesdb.cdie.org/gbk/concepts.html.
Economic aid is broken down into three main categories:
(1) US Agency for International Development (USAID): these data include loan and grant
commitments for development assistance.
(2) Food for Peace: Loans and grants under the P.L. 480 program – Agricultural Trade Devel-
opment and Assistance Act of 1954. This entails:
(a) Title I – Sales Agreements: sales on credit terms.
(b) Title II – Food Aid and Section 416 Program: funds for the transfer of agricultural
commodities by the Commodity Credit Corporation (CCC) to meet relief require-
ments or to promote economic development.
(c) Title III – Food for Development Program: for the transfer of agricultural commodi-
ties by the CCC.
(3) Other Economic Assistance:
(a) Loans: loans authorized by the Inter-American Development Bank, US Export-Import
Bank loans and other loans, principally short-term credits under the CCC Charter Act,
Overseas Private Investment Corporation (OPIC) direct loans, and private trade agree-
ments under Title I of PL 480.
(b) Grants: these include:
(i) Contributions to International Financial Institutions: the World Bank and the
regional development banks.
(ii) Peace Corps: data include country budgets for regular Peace Corps operations.
(iii) Narcotics: grants for International Narcotics Control.
Military aid from 1964 is broken down into five sections:
(1) Military Assistance Program (MAP) Grants: primarily grants of military equipment,
supplies and services purchased with funds appropriated under the MAP.
(2) Foreign Military Credit Financing: through FY 1968 this aid represents credit financing
under the MAP program; data for FY 1969–1993 represent US government credits
provided under the Foreign Military Sales (FMS) Credit and the Foreign Military Financ-
ing programs.
(3) International Military Education and Training (IMET) Program: the data represent grants
of instruction and training services to military and related civilian personnel of friendly
countries.
(4) Transfers of Excess Defense Articles (EDA): these data reflect the original acquisition
value of equipment and supplies.
(5) Other Grants: primarily aid provided under authorities of the Arms Export Control Act;
also included are transfers of vessels.
