Overseas Economic Aid or Domestic Electoral Assistance: The Political Economy of Foreign Aid Voting in the U.S. Congress

Overseas Economic Aid or Domestic Electoral Assistance: The Political Economy of Foreign Aid Voting in the U.S. Congress Ryan M. Powers David Leblang Michael J. Tierney September 1, 2010 Abstract Each year, Congress authorizes the transfer of approximately $27 billion to foreign countries in the form of foreign development assistance. Despite the foreign aid label, much of this money remains in the United States, flowing to individuals and corporations who contract with the U.S. Agency for International Development (USAID) and other government agencies to implement development projects. But despite its long history, little is known about the domestic political economy of U.S. foreign aid. We extend the literature by showing that the domestic political economy of foreign aid spending is more significant and dynamic than the current literature suggests. We find that members of Congress are more likely to support foreign aid bills when foreign aid beneficiaries are present in their district. In addition, we show incumbents are less likely to be reelected if they vote for foreign aid bills while in office. Using new data, however, we find that when a given member s vote for foreign aid is coincident with his or her district receiving contracts from USAID, the negative effects of supporting foreign aid are mitigated. Prepared for the 2010 American Political Science Association Annual Meeting in Washington, DC. We are grateful to Helen Milner and Dustin Tingley for graciously sharing their data. We thank John Kirn, Sneha Raghavan, and Michael Weissberger at the College of William & Mary for research assistance. University of Wisconsin Madison. Email: rpowers@wisc.edu University of Virginia. Email: leblang@virginia.edu College of William & Mary. Email: mjtier@wm.edu 1 Electronic copy available at: http://ssrn.com/abstract=1670682

2 1 Introduction Despite its label, U.S. foreign aid spending with an average annual size of $27 billion over the last five years is an important source of income for many U.S. citizens. Indeed, between 2004 and 2008, an average of 46.9 percent of the U.S. foreign aid budget has been classified as tied aid meaning recipient countries are required to spend the funds on goods and services produced in the United States. 1 The potential domestic economic benefits of foreign aid spending are not lost on members of Congress and the recent political economy literature suggests that these potential economic benefits affect how members of Congress vote on foreign aid packages. Milner and Tingley (2010) find that representatives of Congressional districts with high-levels of human capital are more likely to support foreign aid. They tell a simple and compelling distributional story: The domestic geographic distribution of human capital in donor countries creates winners and losers in relation to foreign aid spending. These winners and/or losers exercise influence over how members of Congress are likely to vote on foreign aid bills. Where foreign aid winners form a sufficiently large coalition, their interests are an important component in determining a legislator s support for foreign aid. We extend this logic and argue that if members 1 We calculated this statistic by using AidData Research Release 1.9.2 where the last five years of complete data run from 2004 to 2008. The portion of tied aid varied between 24 and 69 percent over that five year span. (Nielson et al., 2009) Electronic copy available at: http://ssrn.com/abstract=1670682

3 of Congress are voting with the economic interests of their constituents in mind, we expect that in economic hard times legislators representing large numbers of winners will be even more likely to favor foreign aid. Further, we expect that in districts that are sufficiently dependent on the domestic economic benefits of foreign aid, legislators will be rewarded or punished for their foreign aid voting record at the polls. We test these expectations using new data on domestic contractors with the U.S. Agency for International Development, the principal foreign aid agency of the U.S. government. We begin by replicating Milner and Tingley s findings. We then extend their analysis by testing for whether economic hard times are an important determinant of foreign aid support or opposition by individual legislators. The data show that economic conditions matter little for liberals. Liberal legislators tend to favor foreign aid more than conservatives and neither economic recessions nor rising unemployment have much effect on their voting behavior. For conservatives, however, the higher unemployment in his or her district, the more likely the member is to support foreign aid. The fact that members of Congress respond so strongly to the interests of their constituents suggests that they might be concerned about reelection. But given the comparatively small size of the domestic constituency for foreign aid, it is possible Electronic copy available at: http://ssrn.com/abstract=1670682

4 that these members of Congress are overly sensitive to the concerns of a narrow and inconsequential group. Could a member s voting record on foreign aid bills really affect their chances of reelection? The data suggest that members of Congress understand their districts quite well and that foreign aid votes do shape their prospects at the ballot box. Using a straightforward reelection model, we show that votes in support of foreign aid spending have a negative impact on a member s chances for reelection. But when these votes coincide with more contracts from USAID to firms in a member s district, the negative effects of supporting foreign aid are mitigated. The effect of USAID contracts flowing to a district is magnified when unemployment in a member s district is high. We proceed in four parts. First we review the extant literature and discuss the theoretical grounding for our hypotheses. Second, we explain our sample choice, describe our data sources and discuss how we chose to operationalize each hypothesis. Third, we present our empirical findings and discuss how they relate to our hypotheses. Finally, we suggest some directions for future research.

2 The Domestic Political Economy of Foreign Aid 5 Conceptualizing foreign aid appropriations as payoffs for domestic constituencies has garnered comparatively little attention in the empirical literature. 2 While spending in issue areas such as agriculture, transportation, and education is commonly analyzed for its domestic distributional (and electoral) consequences, few people think of aid as a domestic political payoff...and nobody thinks of it as electorally meaningful. Instead, aid is typically conceptualized as a tool to promote development (Sachs, 2006; Tarp and Hjertholm, 2000) or as a foreign policy instrument to be exercised for the strategic benefit of the state (Dreher et al., 2009; Baldwin, 1985; Vreeland, 2010; Alesina and Dollar, 2000). In addition, the amount of money allocated to foreign aid each year is generally assumed to be small compared to other U.S. government programs like defense ($533 billion in 2010), Medicare ($453 billion in 2010), or transportation ($72.5 billion in 2010). But as Milner and Tingley (2010) note, Aid is not an insignificant part of American foreign policy. Nor is aid spending small compared with several major domestic policy areas. Indeed, while the U.S. committed $32 billion in foreign aid in 2008, the U.S. Congress passed a budget that included $56 2 As far as we know, Fleck and Kilby (2006; 2001) and Milner and Tingley (2010) are the only efforts to consider how domestic constituencies impact foreign aid votes in Congress. Broz and Hawes (2006) look at how domestic constituencies impact votes for funding of international financial institutions (IFI) but the indirect benefits of support for IFIs would suggest a more diffuse political logic at work.

6 billion for Department of Education and $20.2 billion for the Department of Agriculture. Moreover, the U.S. has a reputation of tying significant portions of its foreign aid. 3 According to the OECD s Creditor Reporting System, the U.S. tied an average of 46.9 percent of its total committed foreign aid over the 2004-2008 period. Therefore, a significant portion of the funding labeled foreign aid remains in the U.S. economy, providing immediate and direct economic benefits to individuals, universities, non-profit organizations, and corporations across the U.S. Of course, there are also indirect benefits to individuals and corporations that supply goods and services to the aforementioned direct beneficiaries. We refer to the individuals who benefit directly and indirectly from domestic foreign aid spending in a given representative s district as the foreign aid constituency. Members of the House of Representatives may have a variety of personal and policy goals they would like to achieve, but in order to realize any of these goals, they must first stay in power. Hence all votes are taken in the shadow of an upcoming election where, every two years, a member s constituents can fire said member. This provides a powerful constraint on the legislative behavior of an ambitious politician. The best way to stay in office is to faithfully represent the interests of one s constituents. Despite the strength of party loyalty and presidential prerogatives that 3 In the parlance of international development, tied aid is that aid which is committed for the benefit of another country, but which must be spent on goods or services in the donor country. For more see Jepma (1991).

7 are often said to dominate foreign policy voting in Congress (Kesselman, 1961; Mc- Cormick and Wittkopf, 1990; Canes-Wrone, 2006), Milner and Tingley show that constituent interests are also a significant determinant of support or opposition to foreign aid bills. Using the percentage of a district s workforce that is employed in high skill jobs as a proxy for human capital, they conclude: An identifiable and theoretically predictable group of legislators who support foreign aid exists. Domestic political and economic factors systematically influence American legislators when they cast their votes on foreign aid.... Legislators appear to understand the distributional implications of aid and to vote in accord with the preferences of their constituents, even though they are not organized and lobbying for such aid.... On economic aid votes that have domestic distributional consequences, the Stolper-Samuelson predictions provide a strong explanation for patterns of support and opposition to such aid. Controlling for a wide variety of factors, districts that are better endowed with capital (labor) are more (less) supportive of economic aid, as the theory predicts. Milner and Tingly have taken a crucial first step toward a domestic political economy explanation for foreign aid expenditures. However, in order to ground these results in a micro-foundational political logic, we would ideally show that politicians are self-consciously voting on foreign aid bills to gain electoral support or avoid electoral losses in upcoming elections. 4 Further, we would show that when members of Congress fail to do this, they are punished by voters. We pick up where Milner and 4 Other observable implications may also follow from this domestic political economy model, but are not addressed in this paper. For example, the outsourcing of foreign aid work to U.S. corporations, NGOs, and universities may follow a similar political logic as the money. If most foreign aid activities were carried out by employees of the federal government living in Washington D.C. or foreign capitals, this would certainly provide fewer opportunities to pay off foreign aid constituencies back in the home district.

8 Tingley left off. We extend their logic to identify several more empirical implications that might follow from their model and then test these using new data. So, are members actually responding the interests of their constituents? We need a series of more refined tests that include additional variables and explicitly explore the impact of votes on electoral outcomes. How do varying economic conditions at the national and district levels affect support or opposition to foreign aid? Are there conditions under which legislators normally predisposed to vote against aid would support it instead? Finally, how strong is this effect? How do foreign aid votes impact a member s chances for reelection? And when constituents benefit from foreign aid spending, do they reward their representatives at the polls? In general, we expect economic hard times to negatively impact a member s likelihood of supporting foreign aid. In hard times, the federal budget is constrained and representatives are leery of appearing to send taxpayer dollars abroad while the country struggles to meet its own needs. In February 2010, for example, the Chairwoman of the House Appropriations State and Foreign Operations subcommittee, Rep. Nita Lowey (D-NY), explained to Foreign Policy magazine how it might be difficult to meet the Obama administration s 2011 foreign aid budget request (Rogin, 2010): We re going to be as strong an advocate as we can be, but with 10 percent unemployment, urgent needs at home, a trillion-dollar budget deficit, and focus on creating jobs, there is no doubt that these factors make it

9 a difficult political environment for expanding our foreign assistance and development budgets. Note that the amount of money in question is not large and this cut will not help much to balance the budget. However, Lowey s quote above suggests that spending abroad is more salient in hard times. Such spending opens up legislators to criticism by their partisan opponents on the national level and challengers on the district level. Echoing Lowey, the U.S. Senate approved a budget resolution that cut the foreign aid budget by nearly $4 billion in May 2010. Congress read the writing on the wall and with an election less than 6 months away, they cut foreign aid spending. An April 2010 Economist poll showed that 71 percent of Americans believed foreign aid should be cut in order to balance the federal budget (The Economist, 2010). So, anecdotally, members of Congress appear to be concerned about the amount of money that they spend during economic hard times. However, it is not simply the amount of money the politically savvy politicians ought to be concerned about, but also the delivery mechanism. Recall from our discussion above that one way to garner political support is to provide resources to local constituents through government contracts. A closer look at the type of foreign aid being cut illustrates the same domestic political logic. May 2010 was not the end of foreign aid cuts in the U.S. Congress. In fact, a special session of Congress was called in August of 2010 when the Democratic leadership decided to pass a special

10 appropriations bill in order to ensure that states could avoid firing teachers, health workers, and public safety workers during the recession. However, given the heat Democrats were taking about the budget deficit, they believed that this spending increase would have to be off-set by cuts in other programs. An early target and eventual victim of those cuts was the Millennium Challenge Corporation (MCC), a U.S. government aid agency which lost $50 million out of a $1.5 billion budget to help offset the cost of funding additional teachers at home. Incidentally, the USAID budget, which is much larger, was not cut in August. Tellingly, none of the aid allocated by the MCC is tied to the purchase of U.S. goods and services. Instead, it is closer to pure foreign aid and thus more difficult for legislators to use if they were interested in paying off constituents in their districts. We expect that members of Congress who represent districts receiving contracts from the USAID budget will be less likely to vote for spending cuts and more likely to vote for increases in foreign aid. While conservatives have been shown to be less supportive of foreign aid when all else is equal (Lancaster, 2007; Milner and Tingley, In Press), we are particularly interested in the impact of ideology on support for aid when all else is not equal specifically, in economic hard times. While conservative members of Congress may have an ideological predisposition to be budget hawks, when an election is looming

11 and their district s population is hurting, they may be more willing to change their position on foreign aid (or any other spending program) if they are in a position to steer resources toward their constituents in need. Most conservative members of Congress are not in such a position, because they do not represent districts that get a lot of foreign aid contracts, but some do. Our political economy logic suggests that a constituency s economic need will trump conservative ideology when an election is looming. For members of the House, an election is always looming. Liberal members of Congress are not as ideologically constrained in the first place and should favor the distribution of foreign aid money in good times or bad. Hence, when times shift from good to bad, we ought not observe a major change in the behavior of liberals since there was never much of an impetus to restrain foreign aid spending in the first place. 3 Sample, Data, and Methods Our main dependent variable in the first part of our analysis is a given member s support or opposition to foreign aid spending bills in Congress. We employ a dataset of roll call votes on foreign aid assembled by Milner and Tingley. Their dataset includes roll call votes on military, economic, and agricultural aid in the U.S. Congress between the 96th and 108th Congresses. Because we are primarily concerned with

12 the political economy of economic, rather than military, agricultural, or other forms of foreign assistance, we limit our sample to the 13 roll call votes in their dataset that have a high focus on foreign economic aid. The resulting dataset includes 5,721 observations spanning 24 years. High focus economic aid votes are those votes that most directly reveal a given member s preferences for or against sending economic assistance abroad. As Milner and Tingley explain: These votes all concern economic aid amendments, which sought a general increase or decrease in foreign aid appropriations.... These votes had clear financial consequences for economic aid distributed through key foreign aid programs, such as the main US bilateral aid agency, USAID, or key multilateral organizations. These votes satisfied our four main criteria: 1) they had unambiguous financial effects on aid flows, 2) they were not related to aid directed at a particular country, 3) they did not deal with other key issues such as AIDS, labor rights or abortion, and 4) they did not concern complicated procedural issues. We expect these votes to exhibit the most important distributional effects. We matched each amendment back to the original bill to ensure that we correctly interpreted the high focus coding criteria employed by Milner and Tingley. Each observation in this dataset records a specific member s vote on a specific bill along with district-level, state, and member-level data including among other things unemployment rate, geographic area, ethnic and religious makeup, labor skill level, date of vote, and member ideology (DW-NOMINATE). Their sample begins in the 96th Congress and extends to the 108th Congress. We chose not to extend the dataset backward in time because the quality of district-level control variables declines dra-

13 matically prior to 1979. In order to ensure that our control variables and thus our extended analysis were directly comparable to Milner and Tingley, we chose not to extend the dataset beyond the 108th Congress. We augment their dataset with a dichotomous variable for whether or not the roll call vote was taken during a recession as determined by the National Bureau of Economic Research (NBER). For the Milner and Tingley replication model, the main independent variable is the percent of a given Congressional district s workforce employed in high-skill jobs. As Milner and Tingley define it, the percentage of working age persons in a district employed in executive, managerial, administrative, and professional occupations. This is a fairly standard proxy for a district s level of human capital and is employed by other scholars for similar purposes. Their model also includes alternative measures of human capital from Ladewig which measures the value of physical, fixed capital in manufacturing industries in a district. We retain these measures as well. Recall that Milner and Tingley expect representatives of districts with relatively higher levels of human capital to be more likely to vote in favor of foreign aid. Milner and Tingley include control variables for the influence of the president, ideology, ethnicity and religion of a district s constituents, organized economic interests (labor, corporate political action committees, etc.), and economic conditions in a district. We retain their model specification.

14 While Milner and Tingley find little evidence that economic conditions in a district have significant direct effects on a member s support for or opposition to foreign aid, we believe that testing for the marginal effect of economic hard times provides a better hard test for the existence of a domestic political economy of foreign aid. If members are motivated by the economic interests of their foreign aid constituencies in good times, they should be even more motivated by those economic interests in hard times. Given the nature economic hard times, we suspect that members who would otherwise oppose foreign aid might moderate their opposition when unemployment in their districts is high. We operationalize this test by interacting ideology with unemployment. 5 To ensure that we re capturing district-level (rather than national-level) economic conditions, we control for whether or not the U.S. was in a recession at the time of the roll call vote. We operationalize this using the recession determinations made by NBER. A significant but untested assumption in the Milner and Tingley argument is that representatives are responding to their constituents economic and political interests with regard to foreign aid out of fear that voters may punish them at the polls. We test this assumption using electoral outcome data in combination with relative change in foreign aid dollars flowing to a given district. We use new data on the number of USAID contracts that each district was awarded prior to a given vote. This data 5 Recall that Milner and Tingley find that ideology is a significant determinant of foreign aid.

15 is from the Federal Procurement Data System Next Generation (FPDS-NG) and provides a very direct measure of economic benefits that each district receives from the foreign aid spending over any arbitrary period of time. For our model, we use a rolling count of the number of contracts awarded to a given district in the 24 months prior to any single vote. In all, we have data on over 80,000 individual contracts awarded by USAID to individuals and corporations in the United States from 1980 to 2006. Each observation contains the value of the contract in nominal U.S. dollars, information about who won the contract, a short description of the contract, and the contractor s 5-digit ZIP code. We used the ZIP code to match this contracting data to Congressional districts. 6 If the interests of contractors have no effect on electoral outcomes, increasing or decreasing the number of contracts won by a given district from one Congress to another should have no effect on a member s likelihood for reelection. 6 Matching ZIP codes to Congressional districts is an inherently messy business. Recognizing the drawbacks and the small potential for some error, we used the ZIP-to-CD mapping for the 104th Congress to match ZIP codes to Congressional districts for all Congresses. The 104th Congress is the earliest Congress for which we could find a readily available ZIP-to-CD mapping. Data for several more recent Congresses is available. In later revisions, we plan to use Congress-specific mappings where possible. In any event, redistricting is a comparatively rare event and thus the error rate is quite small. This mapping is provided by the Missouri Census Data Center at Missouri State University (http://mcdc.missouri.edu/).

16 4 Empirical Findings In the first column of Table 1 we successfully reproduced Milner and Tingley s finding. Like Milner and Tingley, we show that levels of human capital in a representative s district are a strong determinant of his or her support for foreign aid. Representatives of districts with high levels of human capital foreign aid winners are more likely to support foreign aid spending. Similarly, we find that ideology is an important determinant for foreign aid support. As one moves right along the ideological spectrum, support for foreign aid declines significantly. The average liberal in our sample voted for aid 73.16 percent of the time, while the average conservative voted for aid 26.8 percent of the time. Also like Milner and Tingley, we find that district-level economic indicators are not in themselves a significant determinant of the average representative s foreign aid vote. In the second through fifth columns of Table 1 we extend Milner and Tingley s model to see whether economic hard times have an influence on the voting behavior of members of Congress. The second column adds the measure of recession based on NBER s determination and we find that this variable does not have a statistically significant and independent effect on the foreign aid votes. Our hypothesis, however, is that economic hard times are conditional on the partisanship of a member of Congress. In column three we model this effect by interacting recessions with the

17 ideology of the representative. And the results in the third column support the conditional hypothesis: all else equal, more conservative members of Congress tend to vote against foreign aid bills and this effect is magnified during recessionary periods. Likewise more liberal members of Congress tend to vote in favor of foreign aid but this likelihood decreases significantly when the economy is in recession. This finding suggests that when times are tough elected officials are disinclined to vote for measures that appear to send money out of the country. In this respect, our findings are broadly similar to Dang, et al. (2009) who find that national-level economic crises have negative impacts on expected foreign aid levels. The fourth and fifth columns of Table 1 extend our argument about the conditioning effects of the economic environment by including an interaction between district-level unemployment and ideology. Here the results are a bit more interesting: the interaction between unemployment and partisanship is positive. This means that, all else equal, more members of Congress will be compelled to vote in favor of foreign aid bills as unemployment in their district increases. While this may be counter to our expectations, it does make sense if foreign aid is tied to firms or products that are located in conservative districts. We explore this effect in more detail in Table 2, where we test for electoral effects. Existing theory suggests that members of Congress vote for legislation that will

18 benefit their constituents. Receiving these benefits the constituency will, in turn, be more likely to return its representative to Congress. The results in table 1 provide evidence supporting the first part of this conjecture. In Table 2,we provide some of the first empirical tests of the second. For this set of tests our dependent variable is whether a sitting member of the House of Representatives is returned to Congress. We exclude those members who decided not to run for reelection so that they could pursue higher office. Our initial specification, contained in the first column of Table 2, models reelection prospects as a function of legislator characteristics, district economic characteristics and their vote on foreign aid bills. We find, all else equal, that partisanship matters as the measure of ideology based on DW-Nominate scores and the dummy variable for membership in the Democratic party are both statistically significant. Both of these variables are negative which suggest that democrats are less likely at least for the time period under examination to be returned to Congress. But the same can be said for the most conservative members of Congress. The results also indicate that constituency economic conditions as measured by the unemployment in the district do not play an important role in reelection. Consistent with conventional wisdom we find that members of Congress that received larger vote shares in their prior election will be more likely to be returned to elective office.

19 In the first column we also include the number of times during the past Congressional term that the member of Congress cast a yes vote on a piece of foreign aid legislation. This variable is statistically significant and negative. In our sample there are a maximum of two foreign aid votes per member and when we calculate the marginal effects, we find that each yes vote decreases the probability of reelection by a little over four percent. The question, of course, is why do yes votes on foreign aid have such negative consequences. We argue that this effect is based on the electorate s antipathy for redirecting tax dollars to targets outside of the country. In the second column of Table 2 we include a variable that measures the change in the number of foreign aid contracts brought to the district during the previous Congressional term. That variable is positive and statistically significant and shows that increasing contracts enhances reelection prospects. In the third column we generate an interaction between yes votes and the change in contracts and find that the joint effect of yes votes, contracts and the interaction is statistically insignificant. This means that although the electorate punishes a representative for voting yes on foreign aid bills, the distaste they have is mitigated when the member of Congress brings the bacon back home to the district. Finally, in the fourth column of Table 2 we include an interaction between the district s unemployment rate and the change in foreign aid contracts awarded to firms

20 in the district. And again we find that the joint effects of contracts, unemployment and their interaction is statistically insignificant. We interpret this as support for the idea that bringing home benefits to the district outweighs negative economic conditions and overall increases the probability that a member of Congress will be reelected. 5 Conclusion Because the U.S. grants such a large amount of tied aid each year, U.S. foreign aid spending represents a significant source of income for many Americans. As such, some Americans the foreign aid winners benefit significantly more than other Americans the foreign aid losers from each dollar spent on foreign aid. Replicating the findings of Milner and Tingley 2010, we demonstrate that when representatives have a sufficiently large foreign aid constituency in their district they are more likely to support foreign aid spending. But the domestic political economy of U.S. foreign aid spending is more complex and dynamic than that. Indeed, we find that liberals support for foreign aid is unrelated to economic conditions in their districts. Conservatives, however, are much more likely to support foreign aid in times of high unemployment in their district. Further, we find that representatives are responding to the interests of their foreign aid constituencies with good reason.

21 While votes in favor of foreign aid have a negative impact on chances for reelection, increases in contracts with USAID mitigate this effect.

22 Marginal Effect of Recession on Foreign Aid Support Marginal Effect of Recession on Yes Vote -3-2 -1 0 1-1 -.5 0.5 1 1.5 Ideology Dashed lines give 95% confidence interval. Figure 1: Marginal Effect of recession on likelihood to support foreign aid across the ideological spectrum (DW-NOMINATE).

23 Marginal Effect of District Unemployment on Foreign Aid Support Marginal Effect of Pct Unemployment on Yes Vote -20 0 20 40 60-1 -.5 0.5 1 1.5 Ideology Dashed lines give 95% confidence interval. Figure 2: Marginal Effect of district unemployment on likelihood to support foreign aid across the ideological spectrum (DW-NOMINATE).

24 Marginal Effect of Change in USAID Contracts on Relection 0.005.01.015.02.025 Marginal Effect of USAID Contracts on Reelection 0.5 1 1.5 2 Number of Yes Votes on Foreign Aid Dashed lines give 95% confidence interval. Figure 3: Marginal Effect of USAID contracts on reelection as the number of supported bills increases.

25 Marginal Effect of USAID Contracts on Reelection Marginal Effect of Change in USAID Contracts on Relection -.005 0.005.01-2 0 2 4 Two-Year Change in State Unemployment Dashed lines give 95% confidence interval. Figure 4: Marginal Effect of USAID contracts on reelection as change in unemployment increases.

26 Table 1: Determinants of Foreign Aid Votes Vote on Aid Vote on Aid Vote on Aid Vote on Aid Vote on Aid (Milner (+ NBER) (+ NBER * (+ Unemp. (+ NBER & Tingley) Unemp. * Ideol.) * Ideol.) * Ideol.) Pct High Skill 3.146 3.146 3.821 3.652 3.980 (3.28) (3.28) (3.82) (3.72) (3.94) Ideology 2.558 2.558 2.623 2.462 2.783 ( 14.52) ( 14.52) ( 15.76) ( 14.11) ( 16.45) Pct Corp 0.605 0.605 0.580 0.617 0.565 PAC Contrib. (2.38) (2.38) (2.22) (2.37) (2.19) Pct Lab 0.190 0.190 0.00905 0.117 0.0178 PAC Contrib. (0.84) (0.84) ( 0.04) (0.51) ( 0.08) Presidential GOP 0.0710 0.0710 0.142 0.0198 0.142 Vote Pct (0.81) (0.81) ( 1.47) (0.20) ( 1.45) Pct Unemp. 0.272 0.272 2.872 3.214 2.450 ( 0.10) ( 0.10) (1.35) (1.46) (1.15) Log Median 0.423 0.423 0.400 0.478 0.380 Income (1.71) (1.71) (1.60) (1.95) (1.49) Pct Foreign 1.760 1.760 1.799 1.891 1.812 Born (3.25) (3.25) (3.63) (3.67) (3.65) West 0.143 0.143 0.179 0.168 0.192 ( 1.33) ( 1.33) ( 1.69) ( 1.57) ( 1.80) Midwest 0.0400 0.0400 0.0647 0.0772 0.0685 (0.38) (0.38) (0.60) (0.71) (0.63) South 0.332 0.332 0.354 0.359 0.351

27 ( 3.15) ( 3.15) ( 3.36) ( 3.37) ( 3.31) Pct Black 0.609 0.609 1.066 1.132 1.072 (1.91) (1.91) (3.41) (3.58) (3.41) Value of Ag. 0.557 0.557 2.495 2.263 2.361 Production (0.11) (0.11) (0.47) (0.43) (0.44) Recession 0.0788 0.248 ( 1.13) ( 2.63) Recession 1.481 2.369 * Ideology ( 6.60) ( 6.32) Unemp 19.96 23.69 13.68 * Ideology (5.47) (6.38) (3.18) Recession * Unemp 45.54 * Ideology ( 3.07) Constant 5.590 5.590 5.448 6.409 5.251 ( 2.26) ( 2.26) ( 2.19) ( 2.64) ( 2.08) Observations 3953 3953 3953 3953 3953 t statistics in parentheses p < 0.10, p < 0.05, p < 0.01

28 Table 2: Reelection Models Reelection Reelection Reelection Reelection (Basic) (+ Contracts) (+ Votes Unemp.) (+ Contacts Unemp.) Ideology 0.686 0.671 0.812 0.661 [ 1.65] [ 1.46] [ 1.71] [ 1.44] Democrat 0.611 0.781 0.818 0.778 [ 2.04] [ 2.27] [ 2.35] [ 2.26] Share of Pop. Vote 0.0156 0.0186 0.0182 0.0186 [3.53] [3.77] [3.71] [3.77] District Unemployment 2.845 [ 1.21] Number of Yes Votes 0.197 0.214 on Foreign Aid [ 2.36] [ 2.12] Change in USAID 0.000243 0.000150 0.00140 Contracts [1.80] [ 0.36] [2.26] Two-Year Change in 0.210 0.213 0.215 State Unemployment [ 4.09] [ 4.15] [ 4.15] Yes Votes * 0.00585 State Unemployment [1.72] Change in USAID Contracts 0.00118 * State Unemployment [1.98] Constant 1.452 1.014 1.138 1.009 [3.82] [2.43] [2.69] [2.42] Observations 3420 2500 2500 2500 t statistics in brackets p < 0.10, p < 0.05, p < 0.01

29 References A. Alesina and D. Dollar. Who gives foreign aid to whom and why? Journal of economic growth, 5(1):33 63, 2000. David A. Baldwin. Economic statecraft. Princeton Univ Pr, 1985. J. Lawrence Broz and Michael B. Hawes. Congressional politics of financing the International Monetary Fund. International Organization, 60(02):367 399, 2006. B. Canes-Wrone. Who leads whom?: presidents, policy, and the public. University Of Chicago Press, 2006. Hai-Anh Dang, Steve Knack, and Hasley Rogers. International aid and financial crises in donor countries. Policy Research Working Paper Series, 5162, 2009. Axel. Dreher, J.E. Sturm, and J.R. Vreeland. Development aid and international politics: Does membership on the UN Security Council influence World Bank decisions? Journal of Development Economics, 88(1):1 18, 2009. R.K. Fleck and C. Kilby. Foreign Aid and Domestic Politics: Voting in Congress and the allocation of USAID contracts across Congressional districts. Southern Economic Journal, pages 598 617, 2001. R.K. Fleck and C. Kilby. How do political changes influence u.s. bilateral aid allocations? evidence from panel data. Review of Development Economics, 10(2): 210 223, 2006. C.J. Jepma. The tying of aid. Development Centre of the Organisation for Economic Co-operation and Development; Washington, DC: OECD Publications and Information Centre [distributor], 1991. M. Kesselman. A Note: Presidential Leadership in Congress on Foreign Policy. Midwest journal of political science, 5(3):284 289, 1961. C. Lancaster. Foreign aid: Diplomacy, development, domestic politics. University of Chicago Press, 2007. James M. McCormick and Eugene R. Wittkopf. Bipartisanship, partisanship, and ideology in congressional-executive foreign policy relations, 1947 1988. The Journal of Politics, 52(04):1077 1100, 1990.

30 Helen Milner and Dustin Tingley. Who supports global economic engagement? the sources of preferences in american foreign economic policy. International Organization, forthcoming, In Press. Helen V. Milner and Dustin H. Tingley. The political economy of u.s. foreign aid: American legislators and the domestic politics of aid. Economics Politics, 22(1): 95 94, Summer 2010. Daniel L. Nielson, Ryan M. Powers, and Michael J. Tierney. Broad Trends in Foreign Aid: Insights from PLAID 1.9. 1. Conference paper presented at Aid Transparency and Development Finance: Lessons and Insights from AidData. Conference paper presented at Aid Transparency and Development Finance: Lessons and Insights from AidData, 2009. James Rogin. Lowey: State Department budget faces a difficult political environment. Foreign Policy, 2010. URL http://thecable.foreignpolicy.com/ posts/2010/02/03/lowey_state_department_budget_faces_a_difficult_ political_environment. Jeffrey Sachs. The end of poverty: Economic possibilities for our time. Penguin Group USA, 2006. Finn Tarp and P. Hjertholm. Foreign aid and development: lessons learnt and directions for the future. Routledge, 2000. James R. Vreeland. Buying Bretton Woods. Conference paper presented at Aid Transparency and Development Finance: Lessons and Insights from AidData, 2010.