To what extent does Japanese foreign aid reflect alignment with US foreign policy interests? We know that the United States has called on Japan to use aid and other resources to support US security interests and alliances (Hook Reference Hook, Picard and Groelsema2015, 98; Ikenberry and Inoguchi Reference Ikenberry, Inoguchi, Ikenberry and Inoguchi2003, 12; Jain Reference Jain2015; Reveron, Gvosdev, and Owens Reference Reveron, Gvosdev and Owens2014, 181). And some analysts suggest that the defense alliance shaped Japan's responsiveness to US diplomacy concerning international security (Uchiyama Reference Uchiyama2010; Osius Reference Osius2002). By the 1970s and 1980s, for example, Japan provided aid to promote some US security interests, including containment of communist regimes in Africa and other regions (Orr Reference Orr1990; Chan Reference Chan1992; Tuman and Ayoub Reference Tuman and Ayoub2004; Tuman, Strand, and Emmert Reference Tuman, Strand and Emmert2009). As Arase (Reference Arase and Arase2005, 11) concludes, “Generally speaking, there is ample evidence of Japan using ODA flexibly to support US allies and strategic aims, and this has helped Japan maintain a stable alliance relationship with the United States.”
Many scholars claim that Japan also aligned its foreign policy with the US War on Terror (Tuke Reference Tuke2013, 2; Smith Reference Smith, Ikenberry and Inoguchi2003; Wakabayashi Reference Wakabayashi2008, 3–4; Ashizawa Reference Ashizawa2014). Most studies point to three factors to explain the alignment. First, enjoying broad domestic support, former Prime Minister Koizumi was willing to engage in a global agenda that converged with the United States’ interests in terrorism (Shinoda Reference Shinoda2007, 85–86, 91–97; Uchiyama Reference Uchiyama2010, 16, 25, 158–159).Footnote 1 Second, due to the legacy of security relations with the US, Japanese officials viewed the US as Japan's most important ally and key for its national defense. Koizumi considered alignment to be in the national interest, and he was eager to avoid US criticism that Japan was not supporting the War on Terror (Uchiyama Reference Uchiyama2010, 87, 93–94; Shinoda Reference Shinoda2007). Third, US diplomacy contributed to Japan's foreign policy response. As early as the 1990s, US diplomats suggested that Japan should punish sponsors of terrorism through a reduction of ODA (Watson Reference Watson1996). Following the attacks of 9/11, however, terrorism became more salient for the US In this context, the Bush administration signaled to Japan (and other allies) the importance of cooperating to counter “terrorist” states (Newnham Reference Newnham2008; Japan Times 2002).
Scholars argue that Koizumi's 9/11 response illustrates aspects of Japan's support for the War on Terror. Koizumi secured passage of legislation permitting deployment the Self-Defense Forces (SDF) in non-combat support roles in Afghanistan and Iraq, while also offering strong support for the US invasions of both countries (Uchiyama Reference Uchiyama2010, 86–88; Shinoda Reference Shinoda2007, 113–114). But equally important was his emphasis on using economic resources to support the US-led efforts (Prime Minister of Japan 2001, as noted in Horiuchi Reference Horiuchi2016; see also Uchiyama Reference Uchiyama2010, 92). His position on foreign aid was articulated in a plan for Japanese–US cooperation to fight terrorism, where he noted: “to the peripheral and the countries concerned, humane and economically necessary support, as part of such efforts, and emergency economic assistance for Pakistan and India to cooperate with the U.S.” (Prime Minister of Japan 2001, 2; authors’ translation, and emphasis added). Thus, Koizumi raised the possibility of disbursing aid to countries cooperating with US military operations. Subsequently, official documents framed Japanese aid projects in Iraq and Afghanistan as examples of Japan–US security cooperation (MOFA 2002; Japan International Cooperation Agency 2010, 2013).
The claim that Japan aligned its foreign aid with the War on Terror has now diffused widely in the literature (Amakasu Raposo and Potter Reference Amakasu Raposo, Potter, Brown and Grävingholt2015, 101–103; Tuke Reference Tuke2013; Soderberg Reference Soderberg, Asplund and Soderberg2016; Wakabayashi Reference Wakabayashi2008; Ashizawa Reference Ashizawa2014; Lancaster Reference Lancaster, Picard and Groelsema2015, 54; Jain Reference Jain2015, 11). For example, case studies of Japanese ODA to Afghanistan and IraqFootnote 2 argue that Japan's aid was significant and intended to help the United States achieve security through reconstruction and economic stabilization (Wakabayashi Reference Wakabayashi2008, 3, 9–10, 19; Ashizawa Reference Ashizawa2014; Tuke Reference Tuke2013; Amakasu Raposo and Potter Reference Amakasu Raposo, Potter, Brown and Grävingholt2015, 101–103).Footnote 3 Reflecting a common interpretation, Stallings and Kim (Reference Stallings, Kim, Shimomura, Page and Kato2016, 125) conclude:
The Japanese government sees its ODA partially as a way to mollify the United States for its small defense budgets and its reluctance to use Self-Defense Forces abroad. This can be seen in the fact that Afghanistan and Iraq were among Japan's largest recipients in 2012.Footnote 4
Further, scholars argue that Japan withdrew aid to states linked to international terrorism (Vreeland and Dreher Reference Vreeland and Dreher2014, 156), while boosting ODA to states that supported the US military, including with the War on Terror (Arase Reference Arase and Arase2005, 11, 18, n 2; Soderberg Reference Soderberg, Asplund and Soderberg2016; Vreeland and Dreher Reference Vreeland and Dreher2014, 157).
To summarize, the recent literature suggests that Japan's foreign policy commitments involved foreign aid policy, and specifically increased aid disbursements to post-war Iraq and Afghanistan (Lancaster Reference Lancaster, Picard and Groelsema2015, 54; Ashizawa Reference Ashizawa2014; Sato and Asano Reference Sato, Asano, Sato and Hirata2008, 112; Stallings and Kim Reference Stallings, Kim, Shimomura, Page and Kato2016). Alignment with the US may have even led Japan to increase ODA to countries that cooperated with US military operations in Iraq and Afghanistan (Prime Minister of Japan 2001; Tago Reference Tago2008; Newnham Reference Newnham2008; Arase Reference Arase and Arase2005).Footnote 5 Conversely, given the salience of terrorism after 2001, Japanese alignment may have reduced ODA to countries designated as state sponsors of terror (Vreeland and Dreher Reference Vreeland and Dreher2014; Tuman, Strand, and Emmert Reference Tuman, Strand and Emmert2009).
Yet prior empirical literature has not examined whether Japanese aid disbursements to post-war Iraq and Afghanistan or other supporters of US war efforts were significantly different from expected disbursements. In this paper, we take up this question by modeling: (1) Japanese aid disbursements to post-war Afghanistan and Iraq; (2) Japanese aid disbursements to developing countries that formed part of the US coalition in Iraq, Afghanistan, and other regions covered by Operation Enduring Freedom (OEF); and (3) Japanese ODA flows to countries that are alleged by the US to be state sponsors of terror. We use a pooled cross-sectional time-series data set that covers 133 countries for the period of 1995 through 2008, with controls for many other possible influences on Japanese aid.
DATA, VARIABLES, AND METHODS
We model aid disbursements by examining real net Japanese ODA to each recipient country for each year. The measure of the dependent variable is normalized as a ratio of real net Japanese ODA (constant 2005 US dollars) to recipient real GDP (constant 2005 US dollars).Footnote 6 The data for net Japanese ODA were obtained from the Development Assistance Committee, OECD (various years), while data for normalization and conversion to constant 2005 US dollars figures were drawn from the World Bank (various years). As noted, the data set for this study covers 133 countries for which there was complete data located in Africa, the Middle East, Asia, Latin America and the Caribbean, and Central and Eastern Europe for the period 1995 to 2008.Footnote 7
To understand the effects of security interests related to the War on Terror, we employ several measures.Footnote 8 First, we employ dummy variables for Afghanistan and Iraq. These are coded 1 for all years of the US invasion and occupation, and 0 otherwise. Second, we include several covariates to control for security interests specific to the Coalition of Willing (COW)-Operation Iraqi Freedom in Iraq, and International Security Assistance Force (ISAF)-Operation Enduring Freedom in Afghanistan (and other areas). These covariates include all developing and transitional countries with military troops in Iraq and Afghanistan. This is an ordinal variable coded for approximate sizes of force structures (see Meernik, Krueger, and Poe Reference Meernik, Krueger and Poe1998). We coded a country 1 for any year in which a country contributed 1–79 troops (small detachments); 2 for any year in which a country contributed 80–299 troops (infantry company); 3 for any year in which a country contributed 300–1,299 troops (battalion); and 4 for any year in which a country contributed >1,300 troops (regiment, brigades); and 0 in those years that troops were withdrawn, and 0 for all other countries that were not contributors. We also include separate covariates (coded with the same criteria) for Central and Eastern European troop contributors.Footnote 9
We also consider the effects of countries that provided the US military use of bases for operations in Iraq, Afghanistan, or other areas covered by Operation Enduring Freedom. The US base covariate is also a dichotomous measure (1 for all countries that contributed bases; 0 otherwise). The data for the US bases variable, and all measures for troop contributors to Iraq and Afghanistan were collected from the US Department of Defense (various years), Belasco (Reference Belasco2009), US Central Command (Reference Command2012), NATO International Security Assistance Force (2012), Global Security (various years), and Johnson (Reference Johnson2004).
Finally, we include a covariate for alleged state sponsors of terrorism. We constructed this measure from the US State Department list of “state sponsors of terror” (US State Department, various years). For this variable, a country was coded 1 during years it was included as a state sponsors of terrorism, 0 for years a country was removed from the list, and 0 for all other countries not on the list.Footnote 10
Because previous studies find evidence of Japan's economic interests and humanitarianism, we include controls for Japanese trade with each potential recipient, oil exporters, sectoral interests (e.g., Japanese whaling), and poverty levels and human rights conditions in recipients (Sato and Asano Reference Sato, Asano, Sato and Hirata2008; Hirata Reference Hirata, Sato and Hirata2008; Miller and Dolšak Reference Miller and Dolšak2007; Strand and Tuman Reference Strand and Tuman2012; Arase Reference Arase and Arase2005; Tuman, Strand, and Emmert Reference Tuman, Strand and Emmert2009; Neumayer Reference Neumayer2003; Anderson Reference Anderson, Koppel and Orr1993; Hook Reference Hook and Zhang2000; Yasutomo Reference Yasutomo1989/1990).Footnote 11 Table 1 provides detailed information on each measure and the data source.
Control Variables
The models are estimated with OLS regression and panel-corrected standard errors (PCSE) to adjust for contemporaneous correlation of the errors and heteroscedasticity (see Beck and Katz Reference Beck and Katz1995).Footnote 12 In addition, to address serial correlation we include a lagged endogenous covariate, Japanese real ODA/GDP in the prior year (see Beck and Katz Reference Beck and Katz2004). We examined summed residuals and residual variance ratios for each country (Stimson Reference Stimson1985) and identified three unit effects (that are modeled).Footnote 13 Finally, we include regional dummy variables to control for regional bias. Central and Eastern Europe (transition states), which received the lowest mean aid during the study period, is the reference. Separate trials indicated that the results were robust to mixed-effects linear regression with a random intercept.Footnote 14 However, trials with a full fixed-effects model were unviable due to a high degree of multicollinearity among the country dummies and other covariates.Footnote 15 Regardless, given that diagnostics identified only three influential units, we remain confident that a full fixed-effects model was unnecessary.
RESULTS
Table 2 presents the results of the statistical models. We begin with a discussion of the covariates for Iraq and Afghanistan. In both models 1A and 1B, the coefficient for Iraq is positive and statistically significant. This suggests that Japanese ODA to Iraq was significantly higher in comparison to pre-war levels and compared to all other recipients, even when controlling for other determinants. While the coefficient for Afghanistan's post-invasion years is positive in both models, it failed to achieve significance. In the conclusion, we offer some tentative reasons for the difference observed between Iraq and Afghanistan.
The War on Terror and other Determinants of Japanese ODA, 1995–2008
Note: Entries are unstandardized OLS regression coefficients; panel-corrected errors in parentheses. ***p < .01; **p < .05; *p < .10 (two-tailed test).
Next, we examine the results for countries that provided troops in support of US operations in Iraq and Afghanistan. The coefficient for Central-Eastern European contributors to ISAF-Operation Enduring Freedom is positive and significant. During the period in question, then, Albania, Bulgaria, Estonia, Poland, and other Central and Eastern European states that were the largest contributors in Afghanistan were rewarded with more Japanese ODA.Footnote 16 However, the coefficients for troop contributors to the COW-Operation Iraqi Freedom, ISAF-Operation Enduring Freedom, and Central-Eastern European contributors to Iraq failed to achieve statistical significance. In separate trials, we also estimated the models with a covariate for recipient countries that provided the US military access to bases (as noted, due to collinearity the base variable was omitted from the original models). In these trials, the coefficient for US bases was small and insignificant.Footnote 17
The coefficient for alleged state sponsors of terror is negative, but the effect is not significant. A more fine-grained analysis suggests that after 2002, Iran and Syria experienced a large reduction in ODA, and this influenced the direction of the coefficient. However, among other alleged state sponsors, there was no clear pattern. In the conclusion, we elaborate on some of the possible reasons for this finding.
The effects of the remaining control variables are straightforward. In both models, the coefficient for Trade/GDP, and microstate members of the IWC, are both positive and significant, which suggests the influence of national and sectoral interests in the aid program.Footnote 18 In addition, in both models, the coefficients the recipient real GDP per capita (log) and human rights abuse are negative and significant. Consistent with the prior literature, our results show that humanitarianism is also at work in the aid program.
DISCUSSION AND CONCLUSION
This study has examined Japan's foreign policy alignment with the US War on Terror, with a focus on its aid policy. The findings suggest that Japan's ODA to Iraq was, in comparison to other aid recipients (and to pre-war aid to Iraq), significantly different and positive. Central and Eastern European states that contributed troops to Afghanistan were also more likely to receive Japanese ODA. Taken together, these findings suggest a degree of alignment between Japan's aid and security interests related to the War on Terror. However, in contrast to the claims made by several scholars (Ashizawa Reference Ashizawa2014; Lancaster Reference Lancaster, Picard and Groelsema2015; Stallings and Kim Reference Stallings, Kim, Shimomura, Page and Kato2016), we found that Japan's ODA to Afghanistan was not significantly different from what we would expect given Afghanistan's characteristics. In addition, there was no significant effect for contributors to Iraq, for states that allowed the US to use bases, or in aggregate for state sponsors of terror.
What might explain the difference in aid pattern observed between Iraq and Afghanistan? First, Afghanistan is a poor Asian country, with low levels of GDP per capita. We controlled for the effects of GDP per capita and Asia in the models, which were both significant and in the expected direction. Once these controls were modeled, there may not have been any additional effect for Afghanistan. Second, it should be recalled that there were several facets to Japan's policy, including not only aid but also deployment of the SDF (Uchiyama Reference Uchiyama2010; Shinoda Reference Shinoda2007). Japanese officials may have decided that, compared to other recipients, no significant increase in ODA was necessary in Afghanistan because Japan had already committed resources for the SDF non-combat deployment.Footnote 19 Third, Japan may have coordinated aid with a group of donors who were supporting the US (Jain Reference Jain2015, 11). Japan's role might have been to increase aid to Iraq, and to Central/European states that contributed to the ISAF operation, while other donors ramped up aid to Afghanistan or other developing-country governments who contributed troops. The “donor coordination and burden sharing” hypothesis is beyond the scope of this study, but it is a potentially fruitful line of inquiry (Kisangani and Pickering Reference Kisangani and Pickering2015).Footnote 20
In addition, the findings suggest that Japan may have adopted a nuanced approach to alleged sponsors of terror. As noted, on average, aid flows were negative to countries alleged to be state sponsors of terror, but the effect was not statistically significant. Further analysis indicated that net Japanese aid flows turned sharply negative in two cases, Iran and Syria, but only after 2002. In other cases, such as Cuba and Sudan (and Iraq, prior to the US invasion),Footnote 21 Japanese aid flows remained positive and similar to recipients who were not on the US State Department list. Japanese officials may have targeted Iran and Syria for cuts because of their proximity to Afghanistan and Iraq, and because the Bush Administration was increasingly critical of both states after 2001 (Miyagi Reference Miyagi2008, 157–58; Japan Times 2002).Footnote 22 By contrast, Cuba and Sudan had lower profiles. Cuba's alleged links to terrorism involved sheltering Colombian FARC rebels in the early 1980s, while in Sudan, US recognition of the humanitarian crisis in Darfur, and improved cooperation between the US and Sudan (after 2001), may have mitigated concern about providing aid (Woodward Reference Woodward2006, 120–122; Amakasu Raposo Reference Amakasu Raposo2013, 13–24). Indeed, the United States was also an aid donor to Sudan in this period. As a result, Japanese aid officials may have perceived that there was flexibility to take other factors into account in Cuba and Sudan despite their designation as state sponsors of terror.
This study does not directly examine the factors that might explain Japan's interest in supporting the War on Terror. Although most scholars argue that Japan's alignment was partly rooted in the long-standing security alliance with the US (Uchiyama Reference Uchiyama2010, 93–94; Shinoda Reference Shinoda2007), domestic politics may have also played a role. Moreover, Japan might have had its own interests in Iraq, Afghanistan, or with global terrorism more generally. A full assessment of Japan's foreign policy alignment will therefore require additional research. In particular, a detailed, historical analysis of Japanese and US diplomacy, and official documents and archival materials, would help in clarifying Japan's motivations. However, it is first important to establish whether Japan's aid to Iraq, Afghanistan, members of the Coalition of the Willing or state sponsors of terrorism did in fact deviate from past patterns and what might have been expected. As we have shown, Japanese ODA to Iraq and to European military contributors to ISAF was significant during the period in question. In contrast to expectations, however, we did not find a significant effect for Afghanistan, for most military contributors, or for state sponsors of terror. This suggests that empirical claims about the alignment of Japan's ODA program with the War on Terror should be qualified somewhat.
