The Alma Mater effect: Does foreign education of political leaders influence UNGA voting?

According to a 1984 New York Times article, he described the Polish leader as a “patriot”, while criticizing Poland’s Solidarity trade union as “negative and dangerous” (see https://​www.​nytimes.​com/​1984/​10/​26/​world/​papandreou-calls-jaruzelski-a-patriot.​html, date of last access: March 18, 2019).

See http://​www.​merriam-webster.​com/​dictionary/​affinity (date of last access: August 27, 2015).

See http://​www.​nyasatimes.​com/​2014/​02/​06/​us-clear-mutharika-on-green-card-gives-up-mp-seat-come-may-21-im-malawi-president/​ (date of last access: August 27, 2015).

See http://​www.​huffingtonpost.​com/​leon-t-hadar/​john-kerry-diplomacy_​b_​4251473.​html (date of last access: August 27, 2015).

See http://​trailblazersblog​.​dallasnews.​com/​2014/​06/​no-canada-sen-ted-cruz-has-formally-shed-his-dual-citizenship.​html/​ (date of last access: August 27, 2015).

A skeptical reader might object that voters hardly know or care about voting in the UN’s General Assembly. Plenty of evidence exists to the contrary. Using response rates to a World Value Survey question about confidence in the United Nations, we find that respondents trust the United Nations to about the same degree as they trust their parliaments or governments. That conclusion holds true for any of the last four waves of the World Values Survey. United Nations General Assembly meetings (where the votes are taken) do not pass unnoticed, but are accompanied by regular protests. Such protests are reported widely in national newspapers, informing voters about the contents of UN sessions, as well as their governments’ stance on them. (See, as one example, the following (US) National Public Radio coverage of the 2012 protests at the General Assembly: http://​www.​npr.​org/​2012/​09/​25/​161767185/​protestors-out-in-full-at-u-n-general-assembly.) The same conclusion holds, in particular, with respect to voting on human rights and with respect to specific domestic leaders. For example, as NYCity Lens reports on October 5, 2015, “the first day of the United Nations General Assembly, hundreds of protesters came from every corner of the globe—Syria, Somalia, Ukraine, Egypt, Korea, China, Tibet, Iran—to raise their voices against oppression, injustice and brutality” (see http://​nycitylens.​com/​2015/​10/​13752/​). In 2015, Ghanaians living in the United States threatened to demonstrate against their President John Mahama’s speech at the UN General Assembly for his human rights violations (see http://​pulse.​com.​gh/​politics/​voter-demo-aftermath-ghanaians-in-us-to-demonstrate-against-mahama-at-un-assembly-id4200943.​html). Similar protests have been recorded and discussed against Indian Prime Minister Modi (see http://​indianexpress.​com/​article/​india/​india-news-india/​sikhs-patels-protest-against-pm-narendra-modi-at-united-nations-headquarters/​). Rwanda’s voting on a human rights resolution in the 2015 UNGA meetings is just one example of the press coverage voting over human rights resolutions achieves (see for example, http://​ktpress.​rw/​2015/​11/​rwanda-votes-for-human-rights-resolution-russia-china-protest/​). Another example is a 2014 resolution against human rights violations in North Korea, when both supporters and opponents received substantial press coverage (see, e.g., http://​www.​aljazeera.​com/​news/​asia-pacific/​2014/​12/​un-north-korea-icc-human-rights-2014121823436300​711.​html).

See the “Protocol for the Modern Diplomat”, https://​www.​state.​gov/​documents/​organization/​176174.​pdf: p.9 (last accessed March 15, 2019).

See https://​www.​reuters.​com/​article/​us-trump-israel-un/​trump-threatens-to-cut-aid-to-u-n-members-over-jerusalem-vote-idUSKBN1EE2A3 (last accessed March 15, 2019).

See https://​usun.​state.​gov/​remarks/​8411 (last accessed March 15, 2019).

Also see Rommel and Schaudt (2016).

We do not use the ideal point estimates calculated by Bailey et al. (2017) because we are not interested in estimating countries’ preference similarities, but rather in whether or not they vote alike on any given set of votes in a particular year.

The difference is significant at the 1% level. It is the largest difference among all topics voted on in the UNGA, as defined by Voeten (2013).

Specifically, we rely on “Civil rights protect people’s liberty against oppression”. Our results also hold when we use the World Values Survey items “Women have the same rights as men” and “Importance of Democracy”.

Following Barro and Lee (2013), we coded the following countries as industrialized (or advanced) economies: Australia, Austria, Belgium, Canada, Denmark, Germany, Finland, France, Greece, Ireland, Iceland, Italy, Japan, Luxembourg, the Netherlands, Norway, New Zealand, Portugal, Spain, Sweden, Switzerland, Turkey, the United Kingdom and the United States. Our main results are similar when we enter those countries into our empirical model.

Following much of the previous literature, our analysis is at the country-year level (Potrafke 2009a; Dreher and Jensen 2013; Mattes et al. 2015; Kahn-Nisser 2019). Given that most of the variables in our regressions do not vary between resolutions within years, we prefer to follow that approach over running regressions at the level of each vote (and thereby increasing the number of observations artificially). As we show in the robustness section, the results are, however, almost identical at the vote level, as one would expect.

To identify leaders receiving foreign educations, and where they did so, the priority order for the sources of our data collection is: (1) Ludwig (2002; 2) Encyclopedia Britannica; (3) individual biographies and government websites; (4) www.​ruler.​org; (5) Wikipedia. We compared our list of foreign-educated leaders with Besley and Reynal-Querol (2011) and double-checked the entries that differed. We also used Besley and Reynal-Querol (2011) to fill in gaps in our dataset.

Faye and Niehaus (2012) use a similar identification strategy. They identify the causal effect of UNGA voting similarities between donors and recipients of foreign aid, comparing election years to non-election years. Nizalova and Murtazashvili (2016) and Bun and Harrison (2019) provide the econometric background. Also see Appendix S.4 in Dreher et al. (2018).

Potrafke (2009a) provides an interesting analysis of the role of a government’s ideology on UNGA voting. According to his results, left-wing governments systematically vote less in line with the United States. The effect is stronger when the US president is Republican. The ideology index employed there (taken from Potrafke, 2009b) cannot be used in our analysis because it is available only for OECD countries. The same is true for the index developed in Bjørnskov (2008).

These data are unlikely to measure ideology in a way that makes governments’ ideological positions perfectly comparable. However, better data do not exist for the wide range of countries and years included in our sample, which explains their wide use in comparable work. Dreher et al. (2015) compare these data with Swank’s (2009) Comparative Parties Data Set as well as with the Comparative Political Data Set from Armingeon et al. (2011), and find them to be correlated at the 1% level.

Dreher and Sturm (2012) suggest GDP per capita and GDP growth as well as aid as proxies for dependence, expecting more dependent countries to vote more in line with the G7. While those expectations receive support from the negative effect of growth in some specifications, it seems that the effect arising from the similar preferences of countries with comparable levels of GDP dominate the dependency effect in our sample. The negative effect of imports on voting similarity is consistent with Kim and Russett (1996), who argue that strong economic ties with developed countries can create feelings of exploitation, which could make them vote against their trading partners. Note that the statistical significance of the control variables typically is lower in the sample focusing on the United States, as could be expected given the sample’s smaller size.

In a previous version of the present paper, we also investigated the effect of studying abroad on voting with each individual G7 country (see Dreher and Yu 2016). According to those results, US-educated leaders are more likely to vote with the United States, on average. While we also obtained positive and significant coefficients for Germany, Italy and Japan, note that the number of leaders who have studied there is rather small—only 12 leaders in our sample have studied in one of these countries. We also tested our theory with respect to leaders who studied in China or the Soviet Union, but found no significant results. Again, the number of leaders in our sample who have studied there is small, however. Detailed results are available on request.

We take Islam from Maoz and Henderson (2013) and CulturalDistance from Spolaore and Wacziarg (2015). It is based on data from the World Values Survey and measures the difference on 98 values-related survey questions across 74 countries. Political ideology is from Beck et al. (2001), as described above.

We also ran these regressions for all G7 countries, including students that have studied in (1) any G7 country or (2) in another G7 country. The results indicate that the effect of foreign education depends on neither religion, cultural distance, nor political color.

The issues include votes involving Israel (and the Middle East more broadly), which provoke controversy as well (Becker et al. 2015).

We also gauged the importance of omitted variable bias following the approach of Altonji et al. (2006) and Bellow and Miguel (2009), comparing the relative influences of those unobserved variables with our coefficients of interest to those of the observed ones. Our results suggest that omitted variable bias would need to be almost 14 times larger than—and in the opposite direction of—the impacts of the observed variables to explain away the full effect of Foreign Education. The working paper version of this paper shows the details of those tests (Dreher and Yu 2016).

One might be concerned about the drop in significance of some of the control variables in the Cold War sample. However, that is unsurprising given the substantial drop in sample size. Note that political ideology is significant in the Cold War sample only, suggesting that cohesion among countries governed by parties of the same color was stronger then. Also note that imports are correlated positively with voting with the United States during the Cold War, and negatively so thereafter. While we have no specific explanation as to why the end of the Cold War marks a structural change in that correlation, theories predict either relationship (negative if strong economic ties with developed countries create feelings of exploitation and positive if countries trade in larger volumes with countries that have similar preferences on human rights). Given that we include those control variables to increase the efficiency of the estimator, but have no specific theory as to how they should affect biases in our variables of interest, we consider changes in the signs of control variables to be of minor importance. When we include “robust” control variables only, the results are unchanged (see Tables A.7, A.8 and A.9).