Who runs the international system? Nationality and leadership in the United Nations Secretariat

National governments frequently pull strings to get their citizens appointed to senior positions in international institutions. We examine, over a 60-year period, the nationalities of the most senior positions in the United Nations Secretariat, ostensibly the world’s most representative international institution. The results indicate which states are successful in this zero-sum game, and what national characteristics correlate with power in international institutions. The most overrepresented countries are small, rich democracies like the Nordic countries. Democracy, investment in diplomacy, foreign aid, and economic/military power are predictors of senior positions—even after controlling for the U.N. staffing mandate of competence and integrity. National control over the United Nations is remarkably sticky; however the influence of the United States has diminished as U.S. ideology has shifted away from its early allies. In spite of the decline in U.S. influence, the Secretariat remains pro-American relative to the world at large.


Introduction
International organizations have played an important role in the process of development, and the global policy approach to issues of world development and international security is crafted in no small part by international institutions. Not surprisingly, senior positions in international institutions are highly contested by governments seeking to place their nationals into office, and once in these positions, there is evidence that officials act in the interests of their home countries. By studying the outcomes of this contested process, we can identify which nations have been most successful at securing top positions, and how this has changed over time. To make this possible, we assembled new data on the nationalities of the most senior officials in the United Nations Secretariat since the founding of the United Nations. We analyze the determinants of control of the United Nations Secretariat, and as an example, use these data to shed light on the role that the United States has played in its history. We provide new evidence both on the internal dynamics of the United Nations, arguably the world's most representative international institution, and we contribute a new measurement of power in international institutions.
Our research informs two major literatures. First, we provide new evidence on the question of which nations exercise influence in international institutions. While the United Nations ostensibly represents the shared interests of all countries, it was nonetheless established by a particular group of countries-the victors of the Second World War-with the goal of sustaining a certain kind of world order (Cronin 2001;Hoopes and Brinkley 2000). A wide and rich literature has examined, for example, the formation and development of the European Union, trade agreements, and the Bretton Woods financial organizations, with the goal of understanding how nations exercise influence through international institutions. Different scholars have brought different theoretical perspectives to this task, emphasizing the role of state preferences (Moravcsik 1993), state power (Gruber 2000), legitimacy (Schimmelfennig 2001), and their interactions with institutional design (Steinberg 2002). Our paper does not attempt to offer a parsimonious theory to explain outcomes in international institutions, but instead to demonstrate a way to track relative influence, using the U.N. leadership as a test case. We will argue that the nationalities of the Secretariat leadership provide a time-varying proxy for the influence of each country at the United Nations.
Second, we explore the possibility of treating the staffing of international institutions as an outcome of the distribution of power across states. According to John Mearsheimer, realists maintain that international institutions are "basically a reflection of the distribution of power in the world" (1994). There is a substantial literature on the measurement of power (Nye 2011b), and a vigorous debate on whether power is even measurable (Guzzini 2013). By identifying the distribution of positions in the Secretariat, we provide an objective measure across all countries of a zero-sum dimension of power, the power to control international institutions (Bachrach and Baratz 1962; Barnett and Duvall 2005), that appears to be of significant concern to governments around the world. While this method of comparing power across countries has limitations, which we discuss below, it addresses some of the critiques in the literature on the empirical measurement of power, especially the critique that a focus on military capability is too narrow in scope.
Using multiple data sources, we researched the nationality of each individual referenced in the United Nations Yearbook  listing of senior officials, which identifies the holders of the approximately 80 most senior positions in the United Nations Secretariat each year (out of some 43,000 total staff members today, United Nations 2012). We first create a descriptive measure of excess representation in the Secretariat, which is the share of senior Secretariat positions held by a country, divided by that country's share of world population. 1 Ranking countries by this measure, we find that the top positions are dominated by rich democracies: the five most overrepresented countries in the Secretariat are Finland, Sweden, Norway, New Zealand, and Ireland. The United States is overrepresented, and China is significantly underrepresented.
We then estimate a multivariate model to identify the factors that give countries more ability to secure scarce Secretariat positions, beginning with the possibility that the official hiring rules of the United Nations may be an important constraint on hiring from certain countries. The Charter of the United Nations mandates the selection of staff on the basis of competence and integrity, which we proxy at the country level with college-educated population and freedom from corruption. The results support the hypothesis that staffing rules limit geographic representation. But these factors alone do not determine who sits in the Secretariat. Drawing on the literature on country influence in international institutions (Cortell and Peterson 2006;Oatley and Yackee 2004;Stone 2011), we include additional measures of total economic output, per capita wealth, military spending, investment in diplomacy, foreign aid spending, and democracy. We find that investment in diplomacy is the most robust correlate of influence in the Secretariat. Democracy, per capita wealth, economic output, military expenditure and foreign aid are also positive but weaker predictors of influence. The results support hypotheses of international institutions as reflections of global power, but also extend them beyond military power: investment in diplomacy and foreign aid has returns, as does similarity in political system with founding members. The results on diplomacy and foreign aid suggest that countries can invest in institutional power in a manner that is at least partially orthogonal to military investment and economic output.
Next, we use our measure of senior position nationality to describe the influence of the United States in the Secretariat over the last 60 years. We present three new facts. First, we observe a secular decline in the share of senior positions held by Americans since the 1960s. Second, the secular decline is apparent even if we control for seats held by allies, as proxied by United Nations General Assembly Voting (Gartzke 1998). Third, these senior positions have not been replaced by growing middle income countries, but by citizens of other rich democracies-largely the allies of the United States at the founding of the United Nations. While the world population share of Western Europe and its offshoots fell from 18% in 1965 to 13% in 2005, their share of senior Secretariat positions marginally rose over time, from 41% in 1965 to 45% in 2005.
We then operationalize state preferences (Moravcsik 1997) in order to make sense of these facts. If American and European ideologies were the same as they were in 1950 (as measured by General Assembly voting), the United States would not have experienced any real loss of ally-weighted influence at the Secretariat, as the positions lost by the United States were on average taken up by its 1950 allies, mainly Western Europe and the Western Offshoots. However, according to our measure, the ideologies of the United States and its 1950 allies have diverged, especially since the 1980s, so the placement of old allies in former U.S. positions has led to a substantial loss in American influence at the United Nations. These findings both reinforce and add nuance to Keohane's (1984) proposition that international institutions can outlast the circumstances determining their creation. The United Nations Secretariat continues to reflect the membership of the U.S.-led alliance that was instrumental in its creation, in spite of that alliance's declining economic and military dominance of the world. But the United Nations Secretariat no longer represents American ideology; American control over the United Nations is now constrained by the allies with which it built the institution.
The methodology of using nationality of senior staff to learn about an unobserved diplomatic struggle is, to our knowledge, novel. This method could be applied to any international institution with sufficient depth of senior staff, such as the European Central Bank, or the European Commission. 2 This methodology could allow for new empirical approaches to the study of how individual states exert influence within international institutions, or how those institutions' bureaucracies function (Cox and Jacobson 1973;Finnemore 1999, 2004).
The literature on delegation to international organizations has focused on the principal-agent relationship between countries and IOs, largely treating IOs as unified entities (Copelovitch 2010;Hawkins et al. 2006;Nielson and Tierney 2003). Parizek (2017) explores the cross-sectional staffing distribution in several other international organizations, focusing on functional and legitimation problems for IOs with unrepresentative staff nationality. Cortell and Peterson (2006) argue that staffing rules are an important determinant of IO autonomy. It is a natural extension to consider that if staff of international organizations serve both the interests of the IO and the interests of their principals, then control over staff nationality may give principals greater control over IO behavior. Formal theoretical exploration of the three-way relationship between countries, IOs, and national staff is beyond the scope of this paper.
Our paper contributes to at least two other areas of research in the political economy of international organizations. One area focuses on the role of major powers in shaping outcomes including loans, agendas, concessions, votes, or peacekeeping scope (Stone 2004;Kuziemko and Werker 2006;Kaja and Werker 2010;Kilby 2011;Kleine 2013;Allen and Yuen 2013;Mikulaschek 2017;McLean 2017). Another takes advantage of the unique institutional environment of the United Nations to answer broader questions, in the spirit of the second half of this paper (Gartzke 1998;Fisman and Miguel 2007).

What does the Secretariat do?
The United Nations is the primary international organization responsible for maintaining peace and facilitating cooperation among states to resolve issues that require collective action. The United Nations' executive arm is the Secretariat. It serves the other bodies of the United Nations, conducts surveys and research, and communicates with non-state actors such as media and non-government organizations. The Secretariat also manages global peacekeeping operations, houses the U.N. Department of Political Affairs, essentially a ministry of foreign affairs with active policy around the world. While the decision-making powers of the United Nations reside within its deliberative bodies (the General Assembly, Economic and Social Council, and Security Council), the Secretariat plays a key role in setting the agenda for those bodies. The content of the resolutions debated in the deliberative bodies originates in the Secretariat, and many of the programs are implemented by organs of the Secretariat. The Secretariat is the main source of economic and political analysis for the General Assembly and Security Council, and operates political field missions which provide knowledge to those bodies. The Secretariat prepares the technical assessments that precede peacekeeping operations and appoints the leaders of peacekeeping operations. While the deliberative bodies maintain high-level oversight of the budget and objectives of a given operation, it is these Heads of Mission, reporting directly to the under-secretary general of the Secretariat, that directly implement peacekeeping operations. Given this range of roles, the Secretariat has more decision-making power than its de jure status suggests. In a 1955 address, Secretary General Dag Hammerskjold described this power as follows: The United Nations is what member states made it, but within the limits set by government action and government cooperation, much depends on what the Secretariat makes it... [it] has creative capacity. It can introduce new ideas. It can, in proper forms, take initiatives. It can put before member governments findings which will influence their actions (Kelen 1968). 3 Traub (2007) describes a number of domains in which Secretariat staff played key roles in matters of global importance. In late 1998, during a crisis of Iraqi resistance to UN arms inspectors, Kofi Annan was the key negotiator on an agreement with dictator Saddam Hussein allowing for continued weapons inspections, so as to keep the Americans from being able to threaten military action (pp. 78-84). In late 2001, following the U.S.-backed invasion against the Afghan Taliban, it was Secretariat staffer (and former Algerian foreign minister) Lakhdar Brahimi who developed and built support with various factions for the blueprint for post-war Afghanistan. Annan in early 2002 asked the UN Security Council to pass a resolution authorizing 3 Many were troubled by this normative view of the Secretariat, and perceived Hammerskjold as too activist a Secretary-General; upon his death, he was replaced by U Thant who was expected to guide the Secretariat to a more subdued role. The case serves to demonstrate that countries place significant importance on the behavior and composition of the U.N. Secretariat. a UN Assistance Mission in . Assistant Secretary-General John Ruggie was considered the intellectual force behind of the Global Compact and the background report We the Peoples, which would eventually form the bedrock for the Millennium Development Goals and convene the Millennium Summit-"the mightiest conclave of heads of state in world history." This summit would define the headline developmental goals for most nations and aid organizations over the ensuing years (pp. 145-150). In all these cases, the actions of Secretariat staff were unquestionably constrained by member nations, but staff were in situations that gave them significant discretion to influence final outcomes.
The staff of the Secretariat are ostensibly international civil servants who serve the goals of the United Nations rather than their home countries. However, the spoken and unspoken struggle between states to place their nationals in senior positions at the United Nations speaks both to the importance of the creative power of the Secretariat, as well as the widespread belief among member nations that Secretariat staff continue to favor the interests of their home countries.

Official procedures
The Secretary-General (SG) heads the United Nations Secretariat, and is selected by the Security Council, with approval from the General Assembly. Under-secretaries general are largely selected by the SG, though need General Assembly approval. The remaining approximately 43,000 staff of the Secretariat are appointed by senior Secretariat officials without direct interaction with the deliberative bodies (Wynes and Zahran 2011).
Appointment of Secretariat officials is guided by two criteria stated in the Charter of the United Nations: The paramount consideration in the employment of the staff and in the determination of the conditions of service shall be the necessity of securing the highest standards of efficiency, competence, and integrity. Due regard shall be paid to the importance of recruiting the staff on as wide a geographical basis as possible (United Nations 1945).
Since 1958, internal documents have discussed a "desirable range" of staff that should come from each country, a range which would be based on membership (i.e., some minimum number of positions per country), population, and assessment of dues, with the largest weight on assessment. 4 While these desirable ranges are intended to apply at all levels of the hierarchy, in practice the focus of internal studies on geographic distribution and proposals for binding quotas have almost entirely focused on the overall distribution of staff (Finger 1975). The nationality distribution of senior positions departs more significantly from the desirable range than the distribution of all positions, a fact which is intermittently discussed in the General Assembly (Meron 1982). 5 Two anecdotes from the early years of the United Nations (both from Ameri 1996) highlight the challenges of achieving fair geographic coverage of top officials. In the early years of the United Nations, citizens of the United States held a disproportionate share of positions in the Secretariat-from 20-25% of all senior positions in the 1950s. A factor contributing to this was that the location of headquarters in New York made it difficult to recruit nationals outside of North America. 6 A second common factor in early staffing decisions was that many countries had a shortage of individuals with sufficient education and experience to fill a senior position at the United Nations. Governments were often invited to recommend their nationals for senior positions, but declined to do so on the grounds that they did not have capable staff to spare.

Unofficial jockeying
It is widely recognized that the top positions are contested in an intensely political process. Historically, the position of Secretary-General has rotated across major regions. Had this process continued in 2016, Secretary-General Ban Ki Moon was expected to be replaced by a national from Eastern Europe. However, in the wake of a series of civil society campaigns, including the "1 for 7 Billion" campaign which was recognized in several national parliaments, a more open process was put in place with debates and informal dialogs between candidates in front of the General Assembly (Borger 2016b). Antonio Guterres from Portugal was ultimately selected with unanimous support from the Security Council but speculation persisted that other senior positions would need to be granted to Russia and China as a quid pro quo (Morello 2016;Borger 2016a).
The struggle for influence over the appointment of the previous Secretary-General, Ban Ki-Moon, was more typical, with China's growing global influence an important factor in the selection of a candidate from Asia (Thant and Scott 2007). Ban's appointed top management team was, according to the Financial Times, "dominated by officials from powerful countries" (2007). Among them was then-U.S. ambassador to Indonesia Lynn Pascoe who was to head the political affairs department despite other countries' objections that he was "a State Department guy" (Turner 2007). Three years later, when Briton and former Blair cabinet member Valerie Amos was appointed under-secretary-general for humanitarian affairs, a source to the Guardian newspaper observed: "This is a massively significant job, one of the top 5 We speculate that the transformation of the goal of achieving geographic representation in senior management into the goal of achieving geographic representation across all staff but not necessarily at the top may be exactly the kind of soft agenda power that the Secretariat leadership is good at exercising. 6 It would be incorrect to view this as a historical accident; the locating of U.N. headquarters in New York occurred because of the United States' position as the dominant world power; the location of main offices of U.N. agencies is one reflection of individual states' importance in the organization. five at the U.N. [...] It would be unthinkable for Britain not to have one of the top five jobs" (Watt 2010). Competition over other high-level positions is often blatant, as governments support the candidacy of their own nationals (The Economist 1989).
Historically, the selection of under-secretaries was an arena of conflict in the Cold War, and frequently discussed in the General Assembly, with Soviets pushing for a transparent division of powers with three under-secretaries representing respectively the Western Block, the Eastern Block and the Non-aligned countries (Reymond 1967). Such a division would make more explicit the allegiance that Secretariat staff often retained to their home countries. The U.S. Senate Select Committee on Intelligence routinely published a report entitled, "Soviet Presence in the U.N. Secretariat," (United States Senate 1985) one edition of which claimed, Soviets in the Secretariat function reasonably well as adjuncts of the Soviet Foreign Ministry and intelligence services [...] The 800 Soviets assigned to the United Nations as international civil servants report directly to the Soviet missions and are part of an organization managed by the Soviet Foreign Ministry, intelligence services, and the Central Committee of the Communist Party. [... They] are involved in shaping conference papers, controlling the flow of news to staff and delegations, influencing delegations seeking Secretariat advice, and aiding Soviet diplomats during conference and other deliberations. [...] Approximately one fourth of the Soviets in the U.N. Secretariat are intelligence officers and many more are co-opted by the KGB.
The Committee Report implies that the United States did not exert the same measure of control over Americans at the United Nations; nevertheless the importance of placing nationals of aligned countries in senior positions is evident.
There have been several high profile examples of Secretariat officials acting in the interests of their home countries over those of the United Nations. During the Iraq Oil-for-Food program, which ran from 1996-2003, Secretariat members had significant discretion over which companies would be able to purchase Iraqi oil. Subsequent investigations revealed significant kickbacks and bribery. The French head of the program, Benon Savan, arranged for disproportionate oil allocations to go to French companies and individuals, including several high level French diplomats. A second example involves the Assistant Secretary-General Tun Myat from Myanmar, who was investigated for assisting a Myanmarese company in selling teak wood into Iraq under the same program. While Myat was ultimately not charged for his actions, he was revealed to have used his connections in the Secretariat to fast-track his compatriots' requests to Program staff, ultimately resulting in contracts being awarded to the Myanmarese company. The Inquiry Committee into the Oil-for-Food Program observed that, "it is commonplace at the United Nations for staff members to be contacted for assistance by private parties from their home country" (United Nations 2005).
Around the same time, the United Nations Special Commission (UNSCOM), the weapons inspection organization in Iraq created to ensure compliance with the 1991 UN Security Council resolution requiring Iraq to destroy its chemical, biological, and nuclear weapons, was operating within the control structures of the UN Secretariat. The staffing of this commission was highly political. It was directed by Swedish diplomat and disarmament expert Carl Ekéus (from 1991-97) and by Australian Richard Butler (from 1997-99), the latter of whom is in our dataset. The assistant director was an American, Charles Duelfer, who had served as a Deputy Assistant Secretary in the U.S. Department of State. A French counselor, Eric Fournier, was a late addition to the team after a meeting between Butler and the French foreign minister. China and Russia both offered experts, who were not appointed (United Nations 1998). There were repeated allegations in multiple outlets that UNSCOM was providing, or had been pressured to provide, cover for espionage by the CIA (Gellman 1999;Billger 2002;Ritter 2002), with Duelfer as the key intermediary. In one instance, CIA operatives embedded UNSCOM equipment to intercept Iraqi military communications (Gellman 1999). In another, according to former weapons inspector Scott Ritter, Duelfer brokered a meeting between UNSCOM and the CIA at Manhattan's Princeton Club, in which the CIA representative gave his blessing for UNSCOM to bring its U2 spy plane footage to the Israelis to interpret, so long as there would be "no American fingerprints" on it (Ritter 2005).
The politicization of the Secretariat is recognized and debated within the United Nations as well. In its own newsletter, the United Nations reported, "The United Nations has increasingly become a political arena where high officials engage in political give-and-take and where 'interest groups' lobby for their country's interests [...] Political appointees are frequently not loyal to the United Nations, but to their respective governments, upon which they depend for further reward or punishment" (Finger and Hanan 1980). Increasing evidence from other domains suggests that the national identities of bureaucrats affect the decision-making of institutions (Kaja and Werker 2010;Johns 2007).

Data: Constructing the database of Secretariat positions
We compiled data on Secretariat staffing from the annual Yearbook of the United Nations, which summarizes the annual activities of all the organs of the United Nations. 7 The Appendix of the U.N. Yearbook lists the names and titles of the most senior staff in the Secretariat and programmes and funds, beginning with the Secretary-General. 8 Using the name and position of each person listed, we researched their nationality, drawing on directories (such as Who's Who in the United Nations), media articles, and other historical documents.
To verify that our list accurately captured the most senior positions in the Secretariat of the United Nations, and to rank the positions in terms of importance, we hired two independent consultants, each of whom had decades of experience working with the United Nations. 9 Both consultants confirmed that our list did not have significant lacunae, and independently assigned each position a weight on a scale of 1 to 6 reflecting the relative importance of that position. The Secretaries General were assigned a ranking of 6, the under-secretaries were assigned 5, and so on. Using the mean importance rank of the two consultants, we created a second measure of Secretariat representation, which is the share of positions held by each country in a given year, weighted by the importance of each position held. We rescaled this measure so that it sums to one in each year.
We use several other country-year variables when we analyze the determinants of Secretariat representation: (i) GDP and population (World Development Indicators, Penn World Tables); (ii) the stock of people with tertiary education (Barro and Lee 2012); (iii) the Worldwide Governance Indicators (of which freedom from corruption is a component) (Kaufmann et al. 2011); (iv) the Combined Polity Score measure of democracy from the Polity IV database (Marshall and Jaggers 2002); (v) state military expenditure and the Composite Index of National Capability from Correlates of War v3.0 (Singer 1987); (vi) an annual count of the number of foreign embassies operated by each country as a proxy for national investment in international diplomacy (the Diplomatic Contacts database from Rhamey et al. 2010); and (vii) net official development assistance by donor, from OECD. Finally, we calculated assessments of dues to the United Nations using data on national GDP, population, U.N. membership, and the formulae described in Section 2.2.1. 10 We define a country's raw representation as the share of Secretariat positions held in a given year by that country's nationals, and we use this measure of representation unless otherwise stated. We consider an alternate measure that weights each position by the expert assessment of its importance, but it does not substantively change any of our findings. Table 1 displays summary statistics of all measures used.
4 Who runs the United Nations?

Representation and excess representation in the Secretariat
We begin by exploring several descriptive measures of representation and excess representation. Any notion of excess representation depends on some reference point.
Here, we define the reference point as a situation where each country is represented in proportion to its population; we thus define a country's excess representation as its share of senior positions in the U.N. Secretariat divided by its global population share. This definition of excess representation can be thought of as a measure of the extra influence in the United Nations enjoyed by each citizen of a member country. Alternate measures are possible; for example, one could begin from a reference point of one-dollar-one-vote or one vote per country. We decided against the former because a one-dollar-one-vote equilibrium can be thought of as an outcome of the exertion of economic power. One vote per country, although it is the voting rule in the General Assembly, seemed like a less useful baseline scenario, as arbitrary divisions between countries would dramatically change their levels of excess representation. Both the multivariate analysis and case study of the United States below are agnostic on this measure and do not take any stand on how excess representation should be defined. Figure 1 shows the excess representation of the United States, Germany, the United Kingdom, Russia, China, and Japan. As discussed above, the United States had a disproportionate share of positions at the inception of the United Nations, a share which fell significantly but then stabilized in the 1980s. Japan and Germany had little representation following the end of the Second World War, but have steadily risen in prominence, surpassing the United States in excess representation by the 1980s. 11 The Soviet Union was almost never overrepresented at the United Nations, in part because of Stalin's significant efforts to undermine the institution in the early years (Finger 1975). The breakup of the Soviet Union led to a further drop in Russia's influence in the Secretariat. China has been and continues to be dramatically underrepresented in the U.N. Secretariat-with an average over all years of only 1% of senior positions in the Secretariat. 12 Figure 2 shows the world population share and share of U.N. Secretariat positions of the Western European powers and the Western Offshoots (Canada, the United States, New Zealand, and Australia). While the Western countries' share of world population and GDP have been steadily declining since the creation of the United Nations, their control over the U.N. Secretariat has not wavered; in 2007 they continued to hold 47% of Secretariat positions, while their world population share fell from 18% to 12% over the sample period. The graph shows that in spite of the widely discussed rise to international prominence of middle income countries like the BRICs, Western Europe and its offshoots have not lost control over this key U.N. body. This evidence supports the claim that China set out to create the BRICS bank and Asian Infrastructure Investment Bank because of underrepresentation in post-war international institutions (The Economist 2014). Figure 3 displays scatterplots of the share of positions in the Secretariat against log population and log assessment. 13 There is a clear upward sloping relationship in both graphs, but many countries are far from the 45-degree line. The R 2 measure for the regression of Secretariat share on population share is 0.11, while for representation on assessment of dues it is 0.70. In the bivariate analysis, payment of dues (which is almost a linear function of GDP) is a much better indicator of Secretariat control than population. Table 2 presents excess representation in the Secretariat, by country, averaged over all years. 14 To save space, we list the top 20 countries, and all other countries with an average population (over all years in the sample) higher than 20 million. 15 Excess representation is defined as the ratio of a country's share of Secretariat positions to its share of world population, and is proportional to a country's vertical distance from the 45-degree line in the top panel of Fig. 3. The Nordic countries dominate the list, occupying 4 of the top 6 positions. Since 1950, Sweden has had on average 0.1% of world population and 0.8% of world GDP, but has held 4.3% of senior positions in the Secretariat, including a Secretary-General from 1953-61. Finland has had a slightly lower share of seats, but with even less population and GDP than Sweden. 16 Western countries are largely overrepresented, with the United States, Canada, and 12 Figure A3 shows a version of the figure with raw representation, defined as the total number of senior positions in the Secretariat, rather than excess representation. 13 We use logs on the x-axis to display small and large countries on the same graph. We must then use a log scale for the y-axis in order for the 45 degree line (which indicates a notion of "equal" representation) to be straight. 14 Online Appendix Table A3 lists countries ranked by their total number of senior Secretariat positions, without taking population into account. 15 Appendix Table A2 shows the complete list with all countries included. 16 Using the importance-weighted measures of positions does not substantively change the list, or any other results below.   The high ranks of small democracies (especially Nordic countries) are consistent with literature on small, open economies, which finds that they take on additional insurance against negative shocks given their dependence on outside factors. These countries tend to have democratic corporatist relationships between business and labor (Katzenstein 1985), larger governments (Rodrik 2014), and strategic investments in security-focused international organizations (Mosser 2000). That small, open economies would seek a greater role at the United Nations is thus not surprising: investing in international institutions may be their best route to global influence.
Second, there may be a global perception that Nordic bureaucrats can be trusted to behave honestly and fairly. The high incidence of Nordic countries thus may not be the result of an exertion of power on their part, but a consequence of their reputation for fairness and reliability. Nevertheless, finding themselves well represented in these senior positions gives the Nordic countries an opportunity to exercise power over international institutions; they may find it to their benefit to continue to cultivate a reputation for impartiality, attaining power through legitimacy (Schimmelfennig 2001;Hurd 2008). Moreover, the ability to shape the preferences of other states is one of the faces of power (Barnett and Duvall 2005;Lukes 1974); the other states may not even realize the power they have bestowed on the Nordics.

Determinants of Secretariat leadership
In this section, we present a multivariate analysis of the factors that predict high representation in the U.N. Secretariat. The measure of representation is the share of senior positions in the Secretariat; it does not depend on any assumptions about the definition of excess representation.
The Charter specifies the primary official criteria for staff selection as efficiency, competence, and integrity; equitable geographic distribution is secondary to these. We begin by taking seriously this mandate, and examine the extent to which efficiency, competence, and integrity can explain the nationalities in the Secretariat. We do not directly observe the caliber of individual bureaucrats in the Secretariat. Instead, we use a set of country-level measures to proxy for the availability of qualified staff from each country. To proxy for efficiency and competence, we use a measure of human capital: the stock of individuals with tertiary education (often a prerequisite for employment at the United Nations). To proxy for integrity, we use the freedom from corruption measure from the Worldwide Governance Indicators (Kaufmann et al. 2011). The latter is motivated by Fisman and Miguel (2007), who found that U.N. diplomats' compliance with law in New York was correlated with the corruption level in their home countries. Given the roles of assessment of dues and population in desirable representation formulae, we include these variables in our baseline model.
Our model of Secretariat representation takes the following form: Y it is a measure of representation, usually the share of senior positions held by country i in year t. P OP ULAT I ON, EDU C, CORRU P T I ON and DU ES are the country-level variables described above, and X it is a vector of additional factors suggested by the literature on state power, which we describe below. An observation is a country-year. Because all these variables are serially correlated, we cluster standard errors by country; this ensures, for instance, that we are not double counting Secretariat positions held for more than one year.
A desirable empirical specification should have two main characteristics. First, all predictive factors should be treated as zero sum, since the share of Secretariat positions is also zero sum. In other words, increasing a country's population by 10% should not affect our prediction of that country's representation if the population of all other countries has also increased by the same 10%. To achieve this, we rescale all observations on Secretariat positions, population, assessments of dues, human capital stock, military spending, diplomatic contacts, and aid spending to convert these to shares of the world total of each value. For instance, instead of gross military spending, we use share of global military spending. For the remaining variables (democracy, freedom from corruption, GDP per capita), there is no notion of a "global total." We normalize these variables to mean zero and standard deviation of one in each year so that each of these describes a country relative to all the other countries. 17 Table 1 shows descriptive statistics for both transformed and untransformed variables. The table shows coefficients from estimation of Eq. 1, with the share of positions in the United Nations Secretariat as the dependent variable. Each observation is a country-year. The dependent variable in column 5 is the share of importance-weighted secretariat representation, with weights equal to expert rankings of the important of each position. Population, human capital, and assessment of dues are represented as world shares. The freedom from corruption index is rescaled to have a mean of 0 and standard deviation of 1 in each year. Standard errors are clustered by country *p < 0.10, **p < 0.05, ***p < 0.01 Table 3 presents estimates with the variables in Eq. 1 added sequentially in Columns 1 through 4. 18 Column 5 repeats the estimation with all variables included, but the dependent variable is the importance-weighted share of positions in the Secretariat, again rescaled so the total number of positions remains one.
The coefficient on population is insignificant in all specifications except the bivariate. Human capital enters positively with statistical significance but is dominated by assessment of dues; we discuss the multicollinearity between these variables below. The freedom from corruption indicator is also positive and significant in all specifications: countries with low corruption and good governance are better represented in the Secretariat. Column 5 indicates that these results are robust to using importance weighted measures of Secretariat representation. The results are consistent with the precedence of the staffing mandate: high competence and integrity are predictors of composition of the U.N. Secretariat. The role of assessment of dues is more difficult to interpret because dues are almost perfectly correlated with GDP. The importance of this variable could therefore be interpreted either as a one-dollar-one-vote relationship between funding and leadership, or it could reflect the importance of economic power as a determinant of control over international institutions. The analysis thus far uses both cross-sectional and time series variation in representation. Table 4 presents estimates from a model with country fixed effects, and thus controls for unobserved country-level characteristics (but also eliminates the interesting variation between countries). The results on population, assessment of dues, and human capital are sustained; the effect of corruption is not visible in the time series alone, largely because it is available for only a third of the years and does not change significantly over time. The persistence of these treatment coefficients in a fixed effect specification lends weight to a causal interpretation: GDP and human capital are not only correlated with influence over the U.N. in the cross section, but the countries that have increased their GDP and human capital over time have gained Secretariat leadership positions as well.
We now examine the role of factors beyond the official staffing mandate of the United Nations. Scholars like Stone (2011) have made a strong case that powerful states exert outsized influence over international organizations. We include variables to proxy for several mechanisms of influence described in the literature. Economic power is already proxied by GDP (which is too colinear with assessment of dues to be separately estimated). We include a measure of wealth (per capita GDP), which may be correlated with control over international institutions to the extent that control can be purchased (Zakaria 1999). 19 To capture coercive power, we include total military expenditure. 20 To proxy for country interest in influencing the United Nations, we The table shows coefficients from estimation of Eq. 1, with the share of positions in the United Nations Secretariat as the dependent variable. Each observation is a country-year. Columns 1 through 3 separately include the highly colinear assessment of dues, share of tertiary educated population, and share of military spending. Column 4 includes the three colinear controls together. Population, human capital, assessment of dues, diplomatic contacts, and military spending are represented as world shares. The freedom from corruption index, Polity score, and per capita wealth are rescaled to have a mean of 0 and standard deviation of 1 in each year. Standard errors are clustered by country *p < 0.10, **p < 0.05, ***p < 0.01 use investment in diplomacy, or the number of foreign embassies operated by each country (Rhamey et al. 2010). Because foreign aid is a widely discussed tool of international policy influence (Morgenthau 1962), we include a country's development assistance budget. Finally, to capture the element of power that results from competition and emulation (as opposed to coercion) (Simmons et al. 2006), we include a country's level of democracy as a proxy for being part of the dominant intellectual ideology or member of the dominant alliance (Lai and Reiter 2000). Table 5 shows results from the expanded estimation. Human capital, military expenditure, GDP, and assessment of dues are all highly colinear, which makes GDP, which we wish to control for separately. Results are robust to the use of CINC in place of military expenditure share and available upon request. it difficult to separate their individual effects. 21 Columns 1 through 3 show that these variables are positive and similar in magnitude when included separately, with marginal statistical significance.
Diplomacy is again the most robust predictor of within-country cross-time variation in Secretariat leadership, suggesting that investment in soft power (Nye 1990) is a key predictor of influence in the U.N. Secretariat. A 1 percentage point increase in a country's share of the world's embassies is associated with a 0.6 to 0.7 percentage point increase in Secretariat representation. Democracy, wealth, foreign aid, and economic output are also correlated with representation, though with weaker statistical significance. The inverse corruption proxy of governance no longer predicts Secretariat positions; it appears to have been proxying wealth, which is highly correlated with a lack of corruption. We also tested for various measures of neutrality and found that it was not a significant predictor of Secretariat influence, undermining this as an explanation for the success of the Nordic countries. 22

The United States and the United Nations
This section demonstrates the utility of our measure of control over senior Secretariat positions as an analytical tool for understanding power over international institutions. We focus on American influence over the United Nations Secretariat as a case study for demonstrative purposes only; the analysis of other countries' influences may be equally fruitful. Throughout this section, we define representation in the U.N. Secretariat as the share of senior positions; no stand is required on a particular measure of "excess" representation. Results are similar if we use an importance-weighted measure or restrict our attention to the most senior positions.
As Fig. 1 depicts, the United States held over 30% of senior Secretariat positions in the early years of the United Nations' existence. Its share of positions declined steadily until around 1980, and recovered slightly in the 1990s. The time series for West Germany and Japan is a striking contrast; for twenty years after the second world war they did not hold a single position between them; but since the 1970s, both countries have risen in prominence to a point where their share of senior positions is more than double their global population share.
To complete this picture, we need to take state preferences into account (Moravscik 1993(Moravscik , 1997. Countries may form coalitions to advance shared interests, and will prefer to have allies in seats of power rather than rivals. Consider the campaign to 21 The correlation coefficients are ρ GDP ,dues = 0.98; ρ dues,tertiaryed = 0.92; ρ dues,militaryshare = 0.89; ρ militaryshare,tertiaryed = 0.84. Note that the per capita measures are less highly correlated, but as described above, the world shares are more appropriate to our theory. When assessment, human capital stock, and military spending are included together (column 4), the education variable dominates, but we are reluctant to infer too much from this since it is based on a small amount of variation between these three measures. 22 Finland and Sweden (respectively #1 and #2 on our measures of excess representation), are not members of NATO; Sweden maintained its neutrality through the second world war and Finland repeatedly switched allegiances. prevent global warming. A state like Palau might be equally satisfied between holding a Secretariat position itself or having Maldives hold a position, since both low-lying island states share the same goal of preventing climate change. The United States might be content to have Canadians and New Zealanders staff the peacekeeping department, if their outlook, strategy, and judgments would be similar to Americans'. Our measure of American decline in the U.N. Secretariat could be overstated if the positions lost by Americans were taken up by their allies.
A good preference-weighted measure of influence in the Secretariat will (i) be increasing both in own positions and in positions of allies; and (ii) put the highest weight on the allies with the greatest similarity of preferences. We can define, at time t, the similarity of preferences between the Secretariat and a given country j by the following expression: where P OSI T I ONS i,t is the share of positions held by country i at time t and φ i,j,t is a measure of the similarity of preferences between country i and country j in year t, with φ i,i,t = 1. In words, for country j , we first calculate its similarity of preference with every other nation in the world-this is φ i,j,t for each country i. We then weight these preference similarities by the share of each country's positions in the Secretariat, to get the similarity between the Secretariat and country j . Thus if a country i has very similar preferences to country j (i.e., φ i,j,t is large), and a large number of seats in the Secretariat (i.e., P OSI T I ONS i,t is large), then country j 's influence in the Secretariat would be increased by country i's positions. Conversely, positions held by opponents (φ i,j,t < 0) decrease a country's Secretariat influence. If a country held every position in the Secretariat, then φ j,SECRET ARI AT ,t would be equal to one. Under this measure, positions for perfect allies (i.e., countries with identical preferences) are as valuable as positions for a country itself. Pairs of countries with weaker preference similarity receive correspondingly less benefit from each others' positions. Many cross-country measures of preference similarities have been proposed; none are perfect. We proxy similarity of preferences with similarity in voting at the U.N. General Assembly, as we think this measure comes the closest to measuring the relevant dimension of preferences for the management of the Secretariat. Following Gartzke (2006), we define similarity of preferences as: where d is the number of times that i votes against j , and d max is the number of General Assembly votes. φ i,j and φ i,SECRET ARI AT are thus both bounded between −1 and 1. The assumption behind our formulation is that officials from two countries with identical preferences over General Assembly resolutions will behave in the same manner if given senior positions in the Secretariat. While there are documented weaknesses in General Assembly voting data as a proxy for preferences, it remains the methodology of choice in many papers in political science and economics   (Voeten 2013). The weighting method described in Eq. 2 would nevertheless work equally well with a different measure of preference similarity. 23 Defining the United States as country j , the solid lines in Fig. 4 show American preference-similarity with the Secretariat over time (in black), along with the unadjusted U.S. Secretariat representation measure (in gray) from Fig. 1. The measures track each other until the mid-1960s. From 1965mid-1960s. From to 1980, the number of positions held by Americans declines, but America's alliance-weighted representation remains constant, indicating that lost American positions are being filled by American allies. Around 1980, U.S. alliance-weighted representation falls in the Secretariat and drops below zero, indicating that the average senior official in the Secretariat is from a state that is an opponent of the U.S. rather than an ally. 24 From 1981 to 2007, the Secretariat is on average staffed by officials from states that are opposed to the United States, with a slight upturn during the Clinton administration and a monotonic fall during the subsequent Bush years.
These changes in alliance-weighted control of the Secretariat could be driven by either a change in composition of the Secretariat or a change in preference similarity across countries. In other words, a country loses influence when (i) it or its allies lose Secretariat positions; or (ii) countries that already have positions become more opposed to the country in question. We can test between these two alternatives by holding preferences constant from the 1950s. This scenario is represented by the dashed line in Fig. 4. When we hold alliances constant, the U.S. decline almost completely disappears, indicating that all of the lost American positions were taken by its 1950s allies, but that these allies no longer share ideology with the United States. America's declining influence is therefore due to a divergence of preference between the United States and its allies, perhaps driven by trends in domestic politics (Moravcsik 1993). These results suggest that, with the exception of declining U.S. influence, the post-war balance of control at the United Nations has been largely static over a 60-year period, in spite of significant changes in both the influence of specific countries at the United Nations and the balance of global economic power over this period. This finding is consistent with the idea that the characteristics of international organizations persist beyond the conditions of their origination (Keohane 1984;Ikenberry 2001). The United Nations was founded by Western states and the Secretariat has been led by them from the beginning to the present.

Measuring institutional bias
We have shown a secular decline in U.S. control over the U.N. Secretariat from 1950 to the present. The average senior official in the Secretariat is from a state that is weakly opposed to the U.S., but it remains possible that the Secretariat has a pro-U.S. bias relative to the world. In other words, a United Nations that is opposed to the United States might be even more opposed to the United States if the Secretariat represented all countries equally. In this section we describe a method to measure the international bias of an institution.
First, we generate a measure of the similarity of preferences between a country and the rest of the world. This is analogous to φ i,SECRET ARI AT above: it is defined as the population-weighted mean of each country's preference similarity with the base country. We label this measure φ i,W ORLD : where P OP ULAT I ON i,t is country i's share of world population at time t. 25 If the nationality distribution of senior officials in the Secretariat was representative of the global population, then φ USA,W ORLD would be equal to φ U SA,SECRET ARI AT , defined above. We define the bias of an institution toward a country as the difference between the country's influence in the institution and that country's preference similarity with the rest of the world: BI AS U SA,SECRET ARI AT ,t = φ U SA,SECRET ARI AT ,t − φ USA,W ORLD,t . (5) 25 As above, this measure could be used with any measure of preference similarity. We plot Secretariat BI AS toward the United States over time in Fig. 5. The figure shows that relative to the world as a whole, the U.N. Secretariat has been consistently biased toward the United States. If the staffing of the Secretariat were proportional to national populations, the United States would have even less influence in the institution. In other words, despite increased U.S. isolation, the leadership of the United Nations Secretariat is more closely aligned with the United States than the world as a whole.
This methodology can be used to examine the bias of any international organization, given data on the nationalities of key officials. As an example, we analyze the bureaucratic leadership of the United Nations Educational, Scientific, and Cultural Organization (UNESCO) during the 1970s and early 1980s. This was the second United Nations agency from which the United States withdrew, in 1984 (Joyner and Lawson 1986). 26 UNESCO's General Conference regularly adopted anti-Israel resolutions, sponsored disarmament activities that the Americans thought were biased in favor of Soviet positions, and promoted restrictions on the freedom of the press through the controversial New World Information and Communication Order (NWICO) (Jacobson 1984;Puchala 1990). UNESCO's Secretariat played a decisive role in the organization's stance and direction (Joyner and Lawson 1986). In Fig. 6 we compute UNESCO's pro-U.S. bias and compare it with the U.N. Secretariat through

Leadership nationality as measure of power in international institutions
The search for objective measures of power that are comparable across countries goes back at least to the eighteenth century (Gulick 1955, cited in Baldwin 2013. Capability-based measures became ever more complex, as reflected in Morgenthau (1948), culminating in Cline's formula (1975), which was a nonlinear combination of population, territory, income, energy, minerals, manufacturing, food, trade, as well as strategy and will. Cline's measure was used by the U.S. army among others to estimate long run trends in national capabilities (Tellis et al. 2000). The capability-based approach has been critiqued by scholars across a number of disciplines, who argue that power is situationally specific and relational and therefore "not objectively measurable" (Guzzini 2009). Contrary to the analysts who count national manpower and resources, Guzzini argues that power is not fungible; what generates power in one context may not generate power in another. For example, U.S. military resources may not be usable against friends as they would be against enemies.
A key development in this literature over the last half century has been the description and categorization of different types of power. Dahl's (1957) seminal study defined a notion of compulsory power as the ability of A to get B to do what B otherwise would not do. Bachrach and Baratz (1962) argued that there was a second face of power: the ability to set the agenda. Lukes (1974) added a third face: the ability to influence others' initial preferences. More recently, Barnett and Duvall's (2005) typology suggests four faces of power. While the traditional capability-based approach is well captured by the idea of compulsory power, control over the staffing of international institutions is closer to Barnett and Duvall's Institutional and Productive Powers: first, the ability to control actors through their diffuse interactions in institutions, and second, the ability to influence "systems of knowledge and discursive practices of broad and general social scope" (Barnett and Duvall 2005), a reasonable description of the Secretariat's global agenda-setting role.
The outcomes of wars provide information about the military capabilities of warring countries. Our proposition is that the outcomes of diplomatic struggles over the leadership of international institutions analogously provides information about the underlying capabilities of the competing states to influence those institutions. Steinberg (2002) noted that state powers "extrinsic" to the rules of an international organization have the effect of "invisibly weighting the decision-making process," generating unequal outcomes. We seek to make visible those weights. We are thus proposing a new, objective measure of national power in international institutions: the share of conationals in key decision-making roles.
This approach brings together two traditions of scholarship that have diverged in recent decades: the measurement of power, traditionally a realist undertaking, and the treatment of power as multidimensional. Our scope is narrow. We focus on the realization of a single dimension of state power: power in international institutions. We cannot predict whether one state can exert power over another in a general sense (e.g., in a war), and our measure is only relevant to the extent that control over international institutions is perceived by states as being important.
With that qualification, our method has several desirable characteristics. First, we are measuring a global outcome that involves nearly all countries in the world, measured with equal accuracy for all countries. Second, it is a continuous measure, available each year that we observe the senior staff positions in the institution of choice, allowing us to observe changes over time. These features give researchers a new opportunity to study the expression of an arguably increasingly important dimension of power in a panel data setting.
The main weakness of United Nations Secretariat leadership as a measure of power is that countries do not equally value the United Nations, so some will exert less effort to secure senior positions in the organization. While the Secretariat is one of the most representative and central institutions of the international system, it is clearly not the case that every country puts in the same level of resources to get staff into key positions. The measure is therefore a combination of the desire and the ability to influence international outcomes. Nevertheless, countries that can obtain these positions at a lower cost can be considered more powerful (Harsanyi 1962, cited in Baldwin 2013, and are likely to obtain more positions, all other things equal. 27 A second possible deficiency in the alliance-weighted measure as an indicator of power is that it would accord the same amount of power to a small country and a large country with very similar preferences. Surely Sweden exerts more control over the United Nations than Iceland, even if their voting records are very similar. Yet from the perspective of their individual governments, there may not be a major difference in terms of whether one country or the other is exercising agency at the United Nations. In Moravcsik (1993), outcomes in intergovernmental negotiations are determined by preferences and bargaining power, and how they interact. Having similar preferences to larger countries would allow smaller countries to focus on those areas in which their preference intensity might be higher, implying this might not be a deficiency after all. Nevertheless, one possible remedy would be to increase the weight on own seats as compared to allies' seats, which would restore the expected rankings.
Finally, we consider whether our measure of influence in the United Nations Secretariat can shed light on one of the most public debates in the field of power and international relations: whether the United States' power is in decline. As Joseph Nye wrote in, "The Decline and Fall of America's Decline and Fall" (2011a), America's decline has been frequently anticipated: first the Soviets in the 1950s and 1960s, then the Japanese in the 1980s, and now the Chinese have all been predicted to "get the better of America." Taking the long view, Paul Kennedy had predicted decline (1989) but even by 2012 many prominent writers were unconvinced (Kagan 2012;Lieber 2012). Related, Keohane (1984) argued that the decline of U.S. hegemony occurred through the 1970s. A parallel, and sometimes overlapping debate has been occurring on whether we are in an era of "American Empire" (Nexon and Wright 2007). The breadth of opinions leaves open the question of whether America's power is in decline, fluctuating, or even in ascent.
The United States structured the international system after World War II to help enforce the so-called Pax Americana. With the United Nations and the Bretton Woods organizations, the United States, wrote John Ikenberry, "spun a web of institutions that connected other states to an emerging American-dominated economic and security order" (Ikenberry 2001). The plan may never have been to dominate the individual institutions once they matured but rather to share them with other states who subscribed to the American worldview. While this view seems plausible for the Cold War years, the notion of a "Western Europe and offshoot" alliance in recent international affairs is less obvious. The United States often has significant disagreement with European countries in matters of international organization, with Americans becoming increasingly unilateralist and Europeans multilateralists (Rubenfeld 2004). 27 Independent of the preferences of governments, nationals of different countries may have different preferences regarding the desirability of U.N. Secretariat positions. While this could be a driver of staff composition, it would not affect the fact that these countries would then have an outsize influence at the United Nations as a result of the preferences of their nationals. In fact, we would expect preferences of citizens toward influencing the United Nations to be correlated with the preferences of their elected governments. Finally, this effect would go in the opposite direction of our findings on wealth, because citizens of poor countries would likely value a secure and well-remunerated U.N. position more highly than citizens of rich countries.
Our measure of Secretariat representation offers an objective and consistent methodology for contributing to this debate. Our evidence shows that the United States has fewer senior positions in the U.N. Secretariat than it used to, and less influence, even after controlling for the positions held by its allies, such that the Secretariat is no longer a projection of U.S. power. The Secretariat is as pro-Western as ever, but this is not to say that the balance of power is entirely static: it is the former allies of the United States that now control the institution rather than America itself. In assisting in the creation of an organization that would be dominated by the West rather than dominated by America (Ikenberry 2001), the United States in the 1940s may have constrained its own future administrations to be more Wilsonian than they would otherwise be.

Conclusion
In 1946, Norwegian Trygve Lie was made the first Secretary-General of the United Nations, in part because of the strong Soviet opinion that the position should go to someone who was neither British, French, nor American (Thant and Scott 2007). The Nordic countries have since continued to dominate the senior ranks of the United Nations. It may be no accident that the bureaucratic arms of the United Nations tend toward an ideology that is not dissilimar to that of the Nordic states. To the extent that international institutions constrain the actions of states, this may put the Nordic countries in a far more influential international position than their economic or military strength would suggest.
We have argued that countries with greater ability to influence international organizations will be more successful in placing their nationals into senior positions, and that this is an important aspect of state power. Since these positions are scarce and central to the operation of the United Nations, the resulting allocation of senior positions gives us information on countries' capabilities in the competition for influence in the international system. In spite of the modesty of its setting, this measure of power has some advantages over traditional capability-based measures.
We find that democracies and countries that invest in bilateral diplomacy and, to a lesser extent, foreign aid, are the most effective at placing staff in the Secretariateven after controlling for monetary contributions to the U.N. and the staffing mandate of competence and integrity. This suggests that exercising influence via a multilateral institution may be a complement to exercising it through bilateral soft power. Examining our measure over time, we find that Western Europe and its offshoots have retained control over a disproportionate share of positions in the Secretariat, even while their share of global GDP and population has fallen.
Going further, we put forward a measure of representation that takes shared preferences between countries into account. We examine the alliance-weighted representation of the United States and find that American influence has been in decline since the formation of the United Nations, especially since 1980, and that the growing ideological distance between the United States and its 1950 allies is the key factor in this decline. However, the Secretariat remains weakly biased toward the interests of the United States, when compared with the interests of the world as a whole.
Our contribution to the great debates in international relations is primarily methodological and demonstrative. We present an empirical approach that can be used to test various hypotheses coming from different schools of international relations. For those studying power in the international system, this paper may lead to new tests being designed for hypotheses other than that which we explored here. For instance, what are the determinants of power in different international spheres? How do major events change the distribution of power? These questions may be explored through datasets other than the U.N. Secretariat but with a similar approach of identifying the "invisible weights" of global governance. And for those analyzing state interactions and institutional outcomes, the panel-data format of the data we bring to bear permits a complementary approach to the case studies that have sourced the large part of the evidence in these debates thus far.
This paper is not a conclusive operationalization of power, or a comprehensive answer to the question of which countries exert the most power in international institutions. Rather, it attempts to plumb the rich information hidden in plain sight, which is the national composition of the senior staff of the world's most global institution. This information offers a quantifiable approach to help understand how states interact in a globalized context.