A priori voting power distribution under contemporary Security Council reform proposals Alexandru Volacu National School of Political and Administrative Studies, Expozitiei 30A Boulevard, Bucharest, Romania. E-mail:
[email protected]
This paper analyses the distribution of a priori voting power of states and regional groups within the UN Security Council, under the most salient reform proposals of the past decade. The results obtained show that moderate proposals, which do not seek to expand veto rights, generally yield a higher voting power share for non-veto states in the Council (both individually and collectively), at the expense of veto states, but they do not alter the geographical distribution of voting power significantly. By contrast, the more radical proposals, which seek to expand veto rights, yield a more balanced regional representation, but have the adverse effect of concentrating voting power almost entirely in the hands of the small number of states with veto rights, with non-veto states holding cumulatively less than 2 per cent of all voting power shares in the Security Council under the Normalized Banzhaf index and less than 0.01 per cent under the Shapley-Shubik index. Journal of International Relations and Development advance online publication, 14 August 2015; doi:10.1057/jird.2015.32 Keywords: reform proposals; security council; voting power; voting power indices; voting rules
Introduction The topic of reforming the UN Security Council (UNSC) has been present in the debates on the institutional configuration of the UN system ever since the adoption of the UN Charter.1 Both in the academic space and in the United Nations (UN), discussions about the reformation of the UNSC usually revolve around six key dimensions: categories of membership, veto rights, regional representation, the size of the Council, working methods, and the relationship between the UN Security Council and the UN General Assembly (Swart 2013: 24). In this paper, I seek to provide an analysis of a priori voting power distribution under the most recent draft resolutions proposed2 by UN member states with regard to the reform of the Security Council.3 Starting from this type of analysis, I will try to address the following research questions: Which reform proposals would be least/most favourable to an increase in the voting power of non-veto states in the Security Council? How would the reform proposals diminish or increase the individual voting power of each state Journal of International Relations and Development, 2015, (1–28) © 2015 Macmillan Publishers Ltd. 1408-6980/15
www.palgrave-journals.com/jird/
Journal of International Relations and Development 2015
2
with veto rights? How would the balance of total voting power shares of veto states and non-veto states be affected? Which reform proposals would be most and least favourable to a more equitable regional distribution of voting power? A priori voting power indices ‘calculate voting power under a given electoral system without reference to past or anticipated voting patterns in any particular elections’ (Gelman et al. 2004: 658), instead being derived exclusively from the voting rules and the size of the committee in question.4 By using a priori voting power indices,5 I seek to analyse the influence that each member state or regional group exerts on the decision-making process of the Security Council, in a decidedly abstract fashion, since a priori voting power only ‘tells us about the power implications of the legislature itself in a pure sense’ (Leech 2002a: 11) and it ‘provides an insight into the power implications of the decision rule’ (Leech 2002b: 6). I will therefore bracket other dimensions of power in general and, even more narrowly, actual or a posteriori voting power,6 which would need to take into account historical and prospective patterns of coalition formation in the Council. While such dimensions are surely important in their own right, the relevance of this paper lies primarily in identifying and assessing shifts in voting power caused exclusively by structural modifications of the Security Council and not by informal aspects of voting behaviour. In order to properly evaluate the impact of changing the ‘rules of the game’ on the voting power of its players, we need to hold other dimensions constant and only focus on what these changes directly entail from an institutional perspective. The present work is embedded in two different strands of literature on the UNSC. The first one concerns the reform of the Security Council, a rapidly expanding research direction in the past years, with recent important contributions made, inter alia, by Russett et al. (1997), who offer a moderate proposal of reform targeting the dimension of the UNSC, the required majority for substantive decision making, limitations on the possibility of immediate re-election into non-permanent seats and the scope of the veto; Fassbender (1998), who places veto rights at the heart of his legal and normative analysis of a UN reform; Weiss (2003), who discusses the obstacles preventing a UNSC reform; Bourantonis (2005), who offers a historical overview of the most salient dimensions of the UN reform; Blum (2005) and Weiss and Young (2005), who provide in-depth analyses of the 2004 Report of the HighLevel Panel on Threats, Challenges and Change, concluding that a reform which would go beyond mere changes to the working methods of the Council would be unlikely in the near future; Schrijver (2007), who analyses the 2004–2005 UNSC reform proposals focusing on the size and composition of the Council; Cox (2009), who analyses the 2004–2005 UNSC reform proposals focusing on how faithfully they address the primary functions of the Council, that is conflict prevention and resolution, opposing the claims that the reform should rather be targeted at increasing legitimacy or representation; Davis (2010) and Schlichtmann (2011), who argue against an enlargement of the UNSC; Butler (2012), who proposes a UNSC reform
Alexandru Volacu A priori voting power distribution
3
that dispenses with veto rights; or Gould and Rablen (2013), who analyse the problem of equity in the Security Council and study the implications of their findings for potential reform proposals. This paper represents a contribution to this type of literature since it develops an alternative framework for assessing various UNSC reform proposals, by focusing on the a priori voting power of individual states and regional groups. Thus, the paper integrates issues such as the expansion of veto rights, the size of a reformed UNSC and majority quotas, discussed in the literature previously mentioned, into a unified dimension of analysis and evaluates reform proposals on this basis. The second strand of literature upon which this paper is built concerns the a priori voting power analysis of the Security Council. Salient contributions to this literature include Shapley and Shubik (1954), who apply the index subsequently named after them to several committees, dedicating a paragraph to the Security Council; Schwodiauer (1968), who extends this type of analysis by seeking to answer various research questions such as: (1) how did the distribution of voting power look like in the UN at that time, (2) did the 1965 expansion of the Council influence significantly the distribution of voting power in the UN, and (3) what effect would the abolition of veto rights have on the distribution of voting power in the UN; Coleman (1971), who illustrates the two indices developed by him with an application to the Security Council; Brams (1975), who compares the Shapley-Shubik and Banzhaf indices as applied to the Security Council; Junn and Park (1977), who use the Shapley-Shubik index to study the voting shares of nonpermanent UNSC members after the 1965 reform; O’Neill (1996), who studies the effects on voting power of adding new seats with or without veto rights and with or without increasing the quota required for taking substantive decisions; and Strand and Rapkin (2011), who analyse voting power in various reform proposals of the UNSC which prescribe the use of weighted voting systems instead of the current system of voting rules. This paper contributes to this literature by expanding its scope through the analysis of potential reconfigurations of the Council, thereby moving the focus from an analysis of the existing institutional arrangement to arrangements that have gained traction with a significant amount of the UN member states in the past years. Finally, Hosli et al. (2011: 171–81) have recently developed an analysis at the intersection of these two approaches, seeking to assess a number of recently proposed UNSC reforms from the perspective of their impact on the decisional capacity of the Council and on the voting power distribution within the Council. This work is particularly important for this paper, which in some sense can be viewed as an expansion of Hosli et al. (2011) in four distinct ways: first, it provides a more indepth exploration of the consequences which the draft resolutions and the Report of the Secretary-General would have on the voting power of member states; second, it takes account of an omission made by Hosli et al. (2011) in calculating voting power indices for the G4 proposal and reinterprets the results obtained in light of this rectification; third, it extends the scope of an a priori voting power analysis of the UNSC reform proposals to the more recently proposed L69/CARICOM
Journal of International Relations and Development 2015
4
Draft resolution; and fourth, it extends the analysis to regional groups as well as individual member states. The paper is structured as follows: in the next section, I describe the Security Council reform proposals discussed in this paper, with a focus on the elements relevant to voting power, that is the number of veto and non-veto states, the majority required for substantive decision making and the distribution of seats among regional groups. In the subsequent section, I explain the methodological framework describing the procedure through which the voting rules of the Security Council can be reinterpreted as a weighted voting system, thus enabling its analysis through the means of voting power indices. In the following section, I analyse the a priori voting power of individual states in reform proposals by comparison with both the status quo and with other proposals and I discuss the issue of cumulative voting power shares of veto and non-veto member states. In the penultimate section, I analyse the a priori cumulative voting power shares of regional groups in reform proposals, once again by comparison with both the status quo and with other proposals. In the final section, I draw the conclusions of the analysis, synthesising the main findings, and I discuss some potential limits of the paper.
The UNSC reform As Weiss (2003) notes, when ‘the UN Charter was drafted, delegates who were dissatisfied with a revival of a kind of nineteenth-century Concert of Europe […] but also did not wish to impede the effective creation of the new world body expected that a review conference for all UN member states would be convened relatively quickly to discuss changes in the charter and organizational structures’ (148). But in the almost 70 years since the draft of the UN Charter, the Security Council experienced a single modification of its membership configuration, namely the 1965 expansion from 11 to 15 members (and, consequently, from six non-permanent members to ten) and the alteration of the majority requirements for substantive votes from seven affirmative votes (including P5) to nine (including P5).7 The regional distribution of the non-permanent members, which are not eligible for immediate re-election is: three seats to African states, two seats to Asian states, two seats to the Latin-America and the Caribbean group, two seats to the Western Europe and Others group, and one seat to Eastern Europe (Cox 2009: 92). The main area of focus for the reform has always been an extended representation, better suited to both regional arrangements and the global power structure. However, other topics such as the abolition or a scope reduction of veto rights, enhancing the effectiveness and procedural workings as well as an increase in the openness of the Council have also been among the main sources of debate and disagreement among the members of the UN General Assembly (Cox 2009: 102). Two main problems with passing reform measures are synthesised by Weiss (2003:
Alexandru Volacu A priori voting power distribution
5
150–52). First of all, there is the fact that major reforms, such as those concerned with membership categories or substantive voting rules, can only be introduced by amending the UN Charter. According to Article 108 of the UN Charter, any amendments to it require ratification by all permanent members of the Security Council,8 which effectively amounts to granting each of them veto rights over amending the UN Charter. Second, the issue of membership extension of both permanent and non-permanent members has usually led to what Weiss (2003) terms ‘political paralysis’ (151), since smaller regional powers seek to block the access of bigger regional powers to permanent seats, while regional blocs disagree amongst themselves on the distribution of newly created non-permanent seats. The first of these two issues, that is the potential of each P5 member to block unilaterally any reform proposal, has had the consequence that while most UN members consider the veto right ‘anachronistic’ (Cox 2009: 119), none of the reform proposals would take away this right from the current P5 members.9 The second of these issues has led to the creation of various coalitions of either geographically heterogeneous (such as the sponsors of the G4 resolution or the United for Consensus group) or geographically homogenous countries (such as the members of the Ezulwini Consensus), supporting a common position in the reconfiguration of the Council. In the following sub-sections, I outline briefly the five reform proposals analysed in this paper. Since I am only concerned with voting power, I intentionally omit various operative clauses or points in the draft resolutions/reports, which target other issues such as increased transparency and accountability, restraining (but not eliminating) the use of veto rights, consultation with other bodies, and so on,10 and I only focus on, first, the number of newly created permanent, semi-permanent and non-permanent seats; second, the question of extending veto rights to newly created permanent seats; third, the regional distribution of newly created seats; and fourth, the majority required for substantive decision making. Thus, I do not take into account two types of reform proposals. The first of these bracketed proposals are those which can only be found in the academic literature, such as Russett et al. (1997), Schlichtmann (2011) or Butler (2012). While these reform proposals are undoubtedly important in their own right, I have chosen to only focus on those proposals that have been drawn from individual states or groups of states, and circulated in the UN General Assembly, since they have a much greater practical weight for the process of reformation. Second, I only take into account those proposals that have clear provisions concerning the first, the second and the third issue above,11 since they are the only ones which alter the distribution of a priori voting power.12 Therefore, the five proposals taken into consideration are: the In larger freedom: towards development, security and human rights for all Report of the Secretary-General,13 the G4 proposal, the Uniting for Consensus proposal, the Ezulwini Consensus proposal, and the L69/CARICOM proposal. The essential characteristic that divides these five proposals is their answer to the question of
Journal of International Relations and Development 2015
6
expanding veto rights, with the first three being opposed to this idea and the other two being in favour of an expansion.14 In larger freedom: towards development, security and human rights for all The report In larger freedom: towards development, security and human rights for all, delivered by Kofi Annan in his capacity as Secretary-General of the UN to the General Assembly on 21 March, 2005,15 was the first major UNSC reform proposal of the new millennium,16 although this was only a small (albeit highly significant) part of a wider project which aimed at overhauling the entire UN system.17 Concerning the issues of membership expansion and voting rules, the report contained two different plans, termed Plan A and Plan B. Plan A, which mirrors to a certain extent the proposal sponsored by the G4 states, envisioned the creation of six new permanent seats (distributed as follows: two for Africa, two for Asia and the Pacific, one for Europe, and one for the Americas) and three new non-permanent seats (with no possibility of immediate renewal), thus bringing the total number of Security Council mandates to 24. Plan B proposed the restructuring of the Council with the same number of states as Plan A, and furthermore, geographically distributed in the same manner, that is six African states, six states from Asia and the Pacific, six European states, and six American states, but the new categories of membership differed, in deference to the concerns of some of the countries supporting the Uniting for Consensus group. Unlike Plan A, this proposal would thus create no new permanent seats, but eight new 4-year renewable seats18 and one new 2-year non-permanent and non-renewable seat. The eligibility criteria for the newly established seats tried to balance diversity in representation and wider accountability with the current power structure, as reflected in the first principle of the operationalisation of membership expansion, which stated that the reforms should lead to an ‘increase [in] the involvement in decision-making of those who contribute most to the United Nations financially, militarily and diplomatically, specifically in terms of contributions to United Nations assessed budgets, participation in mandated peace operations, contributions to voluntary activities of the United Nations in the areas of security and development, and diplomatic activities in support of United Nations objectives and mandates’.19 A final important aspect is that both plans denied the expansion of veto rights to other members, be it the new permanent members (as in the case of Plan A), the semipermanent ones (Plan B) or the non-permanent ones, and, at the same time, they denied the abolition of veto right for the current P5 members. The G4 proposal The so-called G4 proposal,20 embodied in Draft Resolution A/59/L.64, was named after the group containing Brazil, Germany, Japan, and India. One of the primary
Alexandru Volacu A priori voting power distribution
7
reasons for the proposal was for the G4 to obtain permanent seats in the Security Council, in proportionality with their status in the international order and with their financial contribution to the UN. The draft resolution was co-sponsored by 23 other states including many European ones (France, Belgium, Denmark, Greece, the Czech Republic, Poland, Portugal, and so on) and countries from the Pacific region (Fiji, Kiribati, Nauru, Palau, the Solomon Islands, and Tuvalu), and tabled by the General Assembly on 6 July, 2005. Identically to Plan A, discussed above, the G4 proposal called for the addition of six new permanent members, but with the further addition of four new nonpermanent members instead of three. The distribution of new permanent seats was also similar to Plan A, but more in tune with the current distributional arrangements in place for the Security Council, prescribing two seats for African states, two for Asian states, one for Latin-America and the Caribbean states, and one for the Western Europe and Others group. The four non-permanent seats were to be distributed equitably between Africa, Asia, Eastern Europe and the Latin-American and Caribbean states. Unlike the Annan proposals, the G4 draft resolution also provided specifically an amendment of the UN Charter, increasing the majority requirements for substantive decision making from 9 to 14 in the newly restructured Council. The question of veto rights was a particularly significant one, since while the G4 members claimed that their bid for a seat of equal importance as the current P5 members was perfectly legitimate, they could not, for reasons of practicality, go so far as to propose either a remove of veto rights for the current P5 members or an outright extension of veto rights to new permanent seats, without alienating both some of the current P5 members and many other countries who sought to block the proposal even in its lighter form. The solution devised in the draft resolution was to strike a balance between the opposing views, by maintaining the status quo for a period of at least 15 years after the entry into force of the amendments of the UN Charter, followed by a review of the situation, aiming to establish at that time whether veto rights would be extended to the new permanent members or not. The Uniting for Consensus proposal The Uniting for Consensus proposal,21 which became Draft Resolution A /59/L.68, was an initiative spearheaded by Italy and joined by Argentina, Canada, Colombia, Costa Rica, Malta, Mexico, Pakistan, Republic of Korea, San Marino, Spain, and Turkey, tabled on 21 July, 2005. It constituted a ‘reaction against the alleged efforts of certain G4 plan proponents to become permanent members of the Council’ (Cox 2009: 107). Aside from other provisions concerning the internal workings of the Security Council, such as the introduction of procedures for increasing transparency in the decision-making process or a wider engagement of states which are not members of the Council and other UN bodies, the draft resolution proposed an
Journal of International Relations and Development 2015
8
overall increase of seats to 25, with no new permanent or semi-permanent members. According to the Uniting for Consensus proposal, the 20 non-permanent seats would be distributed in accordance with the following regional arrangement: six African states, four Latin-American and Caribbean states, five Asian states, three states from the Western Europe and Others group, and two states from Eastern Europe. All regional groups would decide internally on the re-election and rotation of its members, a further important provision being that any state would be able to run for immediate re-election, following the end of its 2-year term. Finally, the proposed majority required for taking substantive decisions according to the resolution was to be of 15 members, obviously including the P5 members, signalling the intention of the group to promote a wider engagement of actors22 (in this case Council members) in the decision-making process. The Ezulwini Consensus The Ezulwini Consensus, alternatively termed the Common African Position on the Proposed Reform of the UN, is a position adopted by the African Union at its 7th extraordinary session, on 7–8 March, 2005.23 The text adopted was subsequently proposed as a draft resolution in the Security Council, namely Draft Resolution A/59/ L.67, on 14 July, 2005.24 The primary target of the sponsor states, which all came from Africa, was to reconfigure the Security Council in such a way as to shift ‘from an enforcement body concerned with defeating aggression to a guarantor of general […] freedoms’ (Cox 2009: 108), redrawing the membership arrangements and voting rules of the Council with African states at its core. Thus, the proposal was to extend the Council to 26 seats, more than the number provisioned by any other UNSC reform proposal up to that point. Within the 11 newly created seats, six additional permanent seats were to be created: two for Africa, two for Asia, one for Latin-America and the Caribbean, and one for the Western Europe and Others group. Out of the five remaining non-permanent seats, two were supposed to be distributed to Africa, one to Asia, one to Latin-America and the Caribbean, and one to Eastern Europe. Even more interesting than the wide expansion of African states representation in the Council was the proposal made through article (b) of the draft resolution, whereby the newly established permanent seats of the Security Council were to be granted the same privileges as those of the current P5 members, specifically the right to veto. The L69/CARICOM proposal The L69 group is a geographically heterogeneous group, named after the number given to a limitedly distributed Draft resolution sponsored in 2007 by some of its current members, that is ‘A/61/L.69’.25 The 2007 draft resolution contained only six points, which called for an expansion of both permanent and non-permanent
Alexandru Volacu A priori voting power distribution
9
members, greater representation of developing countries, a more equitable geographical distribution and improvement of the working methods, but all of these points were formulated rather vaguely,26 without any specific suggestions for any of the issues. While only 25 UN member states sponsored the original draft resolution, membership has increased to more than 40 states in 2012,27 when the L69 drafted a new proposal containing much more concrete provisions. One year later, the Caribbean Community (CARICOM) circulated another draft resolution which is identical to the L69 proposal on all aspects relevant to a priori voting power (Swart and Perry 2013: 87–88). This resemblance makes it justifiable to consider both proposals together in the analysis. The first important provision of the L69/ CARICOM draft resolutions is an expansion to no less than 27 members,28 with two additional permanent and two non-permanent seats being granted to Africa, two permanent and one non-permanent seat to Asia, one permanent and one nonpermanent seat to the Latin-America and the Caribbean group, one permanent seat to the Western Europe and Others group, one non-permanent seat to Eastern Europe, and one non-permanent seat to ‘small island developing states across all regions’ (Swart and Perry 2013: 88). A second essential provision concerns the expansion of veto rights to all new permanent members. As Gowan and Gordon (2014: 29) remark, this move can be seen as an attempt to gather the support of African states, as it basically satisfies all of their major requirements (differing only in the number of new non-permanent members). A synthesis of UN reform proposals The five reform proposals discussed above are marked by similarities and differences. All of them recognise the need to expand the Security Council in order to ensure a wider representation of the UN member states, in accordance with contemporary evolutions in the field of international political and economic relations. Some of them also agree on not expanding veto rights (at least for the moment, in the case of the G4 proposal) to other new members, irrespective of whether they are nonpermanent, semi-permanent or permanent, while others demand such an expansion. Some of them consider that membership criteria should continue to take into account primarily the contribution of states to maintaining international peace and security, alongside an equitable geographical representation, while others do not.29 All proposals, with the exception of the Report of the Secretary-General, also maintain the same composition for regional blocs as that used by the UNSC in 2005, namely the five-tier division between Africa, Asia, Latin-America and the Caribbean, Western Europe and Others, and Eastern Europe.30 The main points of multilateral divergence (between at least three or even all five proposals) concern primarily the number of new positions added to the Security Council, the nature of these positions (permanent, non-permanent, semi-permanent), the distribution of new seats among regional groups and the majority required for substantive decision making.
Journal of International Relations and Development 2015
10
An illustration of the main points of agreement and disagreement relevant to voting power can be found in Table 1.
Methodological framework The basic procedure for analysing a priori voting power within a committee generally consists of two steps: first, a reinterpretation of the voting rules in the committee as a weighted voting system, and second, an application of voting power indices using components of the weighted voting system. In order to undertake the first step, let us consider that a number v of states in the Security Council have veto rights while a number n do not have such rights, but have the right to vote. Then N = v+n represents the total amount of seats in the Security Council. Let us also consider that there is a number q that represents the majority required for a decision to pass. Under these conditions, we can adjust the number of votes assigned to each v-type state so that the quota in place is impossible to reach without its approval. We can express this in mathematical terms as the system of linear equation (1):31 vw + ðq - vÞ ¼ c ðv - 1Þw + n ¼ c - 1
(1)
In (1) v denotes the number of states with veto rights, w the amount of votes assigned to each veto state in the reinterpreted voting system, q the quota required for decision making, c the quota required for decision making in the reinterpreted voting system, and n the number of states without veto rights. As an illustration, let us apply this system of equations to the current composition of the Security Council. We know that the Security Council has five members with veto rights, therefore v = 5, that the Security Council has ten members without veto rights, therefore n = 10 and the majority required to take decisions, that is q = 9. Then by replacing the terms in the equation above we obtain: 5w+4 = c and 4w+10 = c−1. If we add the negative expression of the second equation to the first, we then obtain w = 7. By replacing w in the first equation, we further obtain c = 39. The voting rules of the Security Council can, therefore, be reinterpreted as a weighted voting system, whereby the P5 states have seven votes each and non-permanent member states each have one vote each, with 39 votes required for any resolution to pass. By appealing to the above mechanism, we can apply a number of voting power indices to both the current Security Council and the configurations envisioned in the reform proposals.32 For the Report of Secretary-General, the Ezulwini Consensus and the L69/CARICOM draft resolutions, the proposals do not address specifically the question of the majority required for adopting decisions. In these three cases, I followed a practice compatible with the historical developments of the Security Council and considered that q = [n/2]+2, thus obtaining q = 14 for the Report and
Alexandru Volacu A priori voting power distribution
11 Table 1 A synthesis of UNSC reform provisions which influence voting power Proposal
Total number of members
Number of states with veto rights
Status quo
15
5
In Larger Freedom
24
5
G4
25
5
Uniting for Consensus
25
5
Ezulwini Consensus
26
11
L69/ CARICOM
27
11
Regional membership Africa: 3 Asia: 3 Latin-America and the Caribbean: 2 Western Europe and Others: 5 Eastern Europe: 2 Africa: 6 Asia and the Pacific: 6 Europe: 6 Americas: 6 Africa: 6 Asia: 6 Latin-America and the Caribbean: 4 Western Europe and Others: 6 Eastern Europe: 3 Africa: 6 Asia: 6 Latin-America and the Caribbean: 4 Western Europe and Others: 6 Eastern Europe: 3 Africa: 7 Asia: 6 Latin-America and the Caribbean: 4 Western Europe and Others: 6 Eastern Europe: 3 Africa: 7 Asia: 6 Latin-America and the Caribbean: 4 Western Europe and Others: 6 Eastern Europe: 3 Small island developing states: 1
Majority requirement for substantive decisions 9
not specified
14
15
not specified
not specified
Journal of International Relations and Development 2015
12 Table 2 A reinterpretation of proposed UN voting mechanisms as weighted voting systems Proposal
v
n
q
w
c
Status quo In Larger Freedom G4 Uniting for Consensus Ezulwini Consensus L69/CARICOM
5 5 5 5 11 11
10 19 20 20 15 16
9 14 14 15 15 15
7 11 12 11 12 13
39 64 69 65 136 147
q = 15 for the Ezulwini Consensus and the L69/CARICOM proposal.33 The results of the calculations are synthesised in Table 2. In general, voting power indices are based on the idea that ‘a member’s power rests on how often he or she can add his or her votes to those of a losing coalition so that it wins’ (Leech 2002a: 5), although both the theoretical and the methodological frameworks used differ between indices. In this paper, I will use four power indices:34 (1) the Normalised Banzhaf index (Banzhaf 1965), (2) the Coleman Power to Prevent Action index (Coleman 1971), (3) the Coleman Power to Initiate Action index (Coleman 1971), and (4) the Shapley-Shubik index (Shapley and Shubik 1954), which I consider sufficient in order to show, first, that the same patterns of voting power distribution generally emerge under various power indices and, second, that there can be relative differences in voting power between states in the Security Council and between the same state under the status quo and under proposed reforms, when we use different indices. In the following lines, I will describe briefly the indices used here, with no reference to their formal structure.35 The Normalised Banzhaf index draws largely on the work of Penrose (1946) and is alternatively called simply the Banzhaf index, in contrast to the Penrose index which is sometimes termed the Absolute Banzhaf index (see Miroiu 2007: 92–94). The Penrose power index seeks to reflect the probability that a player is critical for the coalition to which it belongs in the committee to be a winning one, when all coalitions are equiprobable. The index is calculated on a scale from 0 to 1 for each player, with a player who is essential for all successful coalitions having the power 1 and one who is not essential for any having the power 0. The Banzhaf (1965) index normalises the Penrose index by calculating the relative voting power of states, with the total sum of voting powers equalling 1. Since both indices give results which are proportional to one another and indicate the same direction (although the actual values differ), I will only incorporate the latter into the analysis. The Coleman (1971) indices emerge as a reaction against the game-theoretical approach of Shapley and Shubik (described below) and in convergence with the probability approach of Penrose and Banzhaf. As Leech (2002b) notes, Coleman’s proposal ‘shifted the focus of the analysis from
Alexandru Volacu A priori voting power distribution
13
the powers of the members in relation to each other to the relationship between the powers of individual members and that of the collectivity’ (3). A very brief description of the two indices can be made in the following manner: the Power to Prevent Action represents ‘the ability of member i to prevent action by withholding his vote from a group which would win to one which loses’ (Leech 2002b: 7), while the Power to Initiate Action represents ‘the ability of member i to swing a vote that would fail to produce a majority for action without him, to one which wins with him’ (Leech 2002b: 7). Finally, the Shapley-Shubik index represents the probability that a player is pivotal for a sequential coalition36 to become a winning one, when all orderings of players are equally likely (Leech 2002a: 7). The ShapleyShubik index is also normalised (i.e. it reveals the voting shares of players with the sum of all voting power values being equal to 1), which is something it has in common with the Banzhaf index, but it differs from the two Coleman indices which do not take values that must necessarily sum up to 1.
Voting power of individual states under the UNSC reform proposals By applying the indices described in the third section to the five reform proposals outlined in the previous section, we obtain a number of interesting results.37 A summary of voting power values resulting from this application can be found in Table 3.38 The first interesting, albeit intuitive, result which is similar for the Banzhaf, Coleman PIA and Shapley-Shubik indices is that, under all reform proposals, the voting power of states with veto rights would decrease. While in the Shapley-Shubik index and the Coleman PIA index the decrease is almost unobservable for the first three proposals, ranging between 0.1139 and 0.38 per cent for the Shapley-Shubik
Table 3 Voting power of member states under UNSC reform proposals Proposal
Normalised Banzhaf V
Status quo In Larger Freedom G4 Uniting for Consensus Ezulwini Consensus L69/CARICOM
N-V
Coleman PPA V
N-V
Coleman PIA V
N-V
Shapley-Shubik
V
N-V
0.166929 0.016535 1.000000 0.099057 0.026566 0.002632 0.196270 0.001865 0.136145 0.016804 1.000000 0.123428 0.021587 0.002664 0.193944 0.001594 0.144372 0.013907 1.000000 0.096329 0.023944 0.002306 0.195155 0.001211 0.125062 0.018735 1.000000 0.149802 0.018722 0.002805 0.192464 0.001884 0.089529 0.001012 1.000000 0.011307 0.000480 0.000005 0.090905 0.000003 0.089991 0.000632 1.000000 0.007017 0.000483 0.000003 0.090907 0.000002
Journal of International Relations and Development 2015
14
and between 0.26 and 0.78 per cent for the Coleman PIA indices (thus the shifts are lower than 1 per cent), in the Banzhaf index, the changes are somewhat more relevant, with the lowest decrease being that of the G4 proposal, of only 2.25 per cent, the second lowest being that of the Report of the Secretary-General, of 3.07 per cent, the third lowest being that of the Uniting for Consensus group, of 4.18 per cent, and the highest being that of the Ezulwini Consensus and the L69/CARICOM proposals, of a significant 7.74 and 7.69 per cent, respectively. The Shapley-Shubik and the Coleman PIA also capture a massive decrease in voting power for states with veto rights under the Ezulwini Consensus and L69/CARICOM, their percentage of voting power being reduced by 10.53 per cent for the Shapley-Shubik and 2.608 per cent for the Coleman PIA, to a value under 0.1 per cent.40 The Coleman PPA obviously shows no decrease in the voting power of states with veto rights under any proposal, since each of them still retains the power to block unilaterally the passing of any draft resolution. However, while all indices point to a decrease in the voting power of individual states with veto rights, Table 4 shows that, if we add the total of voting shares for veto and non-veto members respectively, the Ezulwini Consensus proposal actually increases the voting power of all veto states by 15.02 per cent and the L69/CARICOM has a similar, but slightly higher, effect, with a 15.52 per cent increase under the Banzhaf index.41 The other three proposals, however, show a significant decrease in the cumulative voting power of veto states, with the most radical being the Uniting for Consensus proposal, which would yield only 62.53 per cent of voting power to veto states in the Council, maintaining a difference of 20.93 per cent to the current status quo. In contrast, the Shapley-Shubik index is not very sensitive to these variations, as the cumulative voting power of veto states is almost absolute, with 95 per cent of voting shares being held by them in each case. However, the Ezulwini Consensus and the L69/CARICOM proposals are indeed by far the most favourable to veto states according to the Shapley-Shubik index, if taken together, with 99.99 per cent of the voting shares belonging to them.
Table 4 Cumulative voting shares of states with veto rights and states without veto rights under UNSC reform proposals Proposal
Status quo In Larger Freedom G4 Uniting for Consensus Ezulwini Consensus L69/CARICOM
Banzhaf
Shapley-Shubik
V (t)
N-V (t)
V (t)
N-V (t)
0.834645 0.680725 0.72186 0.62531 0.984819 0.989901
0.16535 0.319275 0.27814 0.3747 0.01518 0.010112
0.98135 0.96972 0.975775 0.96232 0.999955 0.999977
0.01865 0.030286 0.02422 0.03768 0.000045 0.000032
Alexandru Volacu A priori voting power distribution
15
If the four indices used point in the same direction in the case of the individual voting power of veto states under all reform proposals, albeit with different values, the same convergence is not present in the case of non-veto states. First of all, the Banzhaf index, the Coleman PPA and the Coleman PIA all point to a decrease in the voting power of non-veto members under the G4 proposal, the Ezulwini Consensus and the L69/CARICOM, while the In Larger Freedom Report and the Uniting for Consensus proposals point to an increase. While the increase in the SecretaryGeneral Report is not very high (0.026 per cent for the Banzhaf, 2.43 per cent for the Coleman PPA, and 0.003 per cent for the Coleman PIA), the increase in the Uniting for Consensus proposal is somewhat higher (0.22 per cent in the Banzhaf, 5.075 per cent in the Coleman PPA, and 0.02 per cent in the Coleman PIA42). The ShapleyShubik index also shows an increase in voting power for non-veto states under the Uniting for Consensus proposal, although the increase is almost insignificantly small, but it does not follow the pattern of the other three indices in the case of the Secretary-General Report. Somewhat surprisingly, the G4 proposal would lead to a decrease in the voting power of non-veto states, most significantly captured by the Banzhaf index, which shows that the difference amounts to 0.26 per cent. All other indices point to a similar decrease of under 1 per cent. Finally, the Ezulwini Consensus would lead to a major decrease of voting power for all individual non-veto states, from 1.65 per cent under the status quo to 0.1 per cent, according to the Banzhaf index, from 9.9 per cent under the status quo to 1.1 per cent according to the Coleman PPA, from 0.26 per cent under the status quo to 0.0005 per cent according to the Coleman PIA, and from 0.18 per cent under the status quo to 0.0003 per cent according to the Shapley-Shubik index. The L69/CARICOM proposal once again points in the same direction as the Ezulwini Consensus,43 with a decrease in voting power from 1.65 to 0.06 per cent for each non-veto state, according to the Banzhaf index, from 9.9 to 0.7 per cent under the Coleman PPA, from 0.26 to 0.0003 per cent under the Coleman PIA, and from 0.18 to 0.0002 per cent under the Shapley-Shubik. The massive decreases in the Coleman Power PPA index for both proposals are particularly illustrative of the heavily diminished role which individual states would play in a Security Council reformed in accordance with these radical proposals. However, if we are interested in the total voting shares of non-veto member states, both the Banzhaf and the Shapley-Shubik indices point to an increase in their voting power, the lowest one under the Banzhaf being for the G4 proposal, of 11.28 per cent, and the highest one being for the Uniting for Consensus proposal, of 20.94 per cent. These results vary significantly between the Banzhaf and the ShapleyShubik indices, since in the latter, the increase is almost insignificant for all three reform proposals, with the highest one (Uniting for Consensus) amounting to only 1.9 per cent. In contrast to the first three proposals, the Ezulwini Consensus and the L69/CARICOM propose a massive decrease in the voting power of non-veto states both individually and collectively, with the Banzhaf index showing a decrease
Journal of International Relations and Development 2015
16
of 15.02, to less than 2 per cent of the total voting power shares and the ShapleyShubik index showing a decrease to 0.0045 per cent for the first proposal and a decrease of 15.52 to almost 1 per cent according to the Banzhaf index and a decrease to 0.0032 per cent according to Shapley-Shubik for the L69/CARICOM. Synthesising these results, we can draw the following conclusions. First, the Uniting for Consensus proposal is the most favourable to non-veto member states, as it leads to an increase in their voting power under all indices used, both individually and cumulatively, and to a decrease in the voting power of veto states, under all indices used except for the Coleman PPA.44 Further, the Banzhaf index shows that, under the Uniting for Consensus proposal, the voting power of all non-veto states would approach significantly that of veto states, with a ratio of approximately 37 per cent for the former and 63 per cent for the latter. Second, the In Larger Freedom Report would also expand individual voting power for non-veto member states, both individually45 and cumulatively, and would lead to a decrease in voting power for veto states, both individually and cumulatively, but at a lower rate than the Uniting for Consensus proposal. Third, the G4 proposal is much more moderate than both Uniting for Consensus and In Larger Freedom, with individual and cumulative voting power still decreasing for veto states (although at a much lower rate), but with individual voting power also decreasing for non-veto states under all indices used. This fact confirms O’Neill’s (1996: 224) claims concerning the essential role played by majority quotas, since the only difference between the G4 proposal and the Uniting for Consensus proposal, which is relevant to voting power, is an increase of 1 in the majority quota for the latter.46 Fourth, the Ezulwini Consensus and the L69/CARICOM proposals are devastating47 to both the individual and the collective power of non-veto states, reducing them to almost 0 under all indices. These proposals would also lead to a decrease in the voting power of each veto state, but to a massive increase of the cumulative voting power of these states. Under both of these proposals, the veto states would effectively hold all the power in the Security Council, while the role of non-veto states would be virtually non-existent. Thus, it is unclear how the proposals would address the concerns raised in the pre-ambulatory clauses of the draft resolutions,48 namely ‘acknowledging the need for the Security Council to reflect present world realities and be more responsive to the aspirations of all States Members of the United Nations’ or being ‘mindful of the need to ensure Africa, like all the other regions of the world, effective representation in the Security Council’,49 since the proposal would simply amount to a quasi-cartelisation of the Security Council by 11 UN member states, while all other 181 states50 would not be ensured any effective representation. Thus, it is important to note that Slaughter’s (2005) claim that ‘a larger and more representative Security Council is likely to diminish the impact of the veto in a number of ways, from isolating a veto-wielding permanent member seeking to cast its veto against the weight of global public opinion, to creating a larger potential majority in favour of having an action proceed in the face of a veto’ (631) can be
Alexandru Volacu A priori voting power distribution
17
extended to a priori voting power, but only in cases in which the expansion of the Security Council does not apply to veto rights as well.
Voting power of regional groups under the UNSC reform proposals The need for a more inclusive and geographically representative Security Council is strongly highlighted in all of its reform proposals. While the meaning of ‘representation’ is widely underspecified in both legal documents and the IR literature on the UN system (and in particular on the Security Council),51 one plausible interpretation of an increase in representation is that regional groups, either in their present configuration or in a modified one,52 should receive a more balanced distribution of seats in the Council. One way in which we can make sense of this idea is to consider that regional representation corresponds, at least in some small part, to the voting power shares held by each regional group in the Security Council, calculated as a sum of individual voting shares of veto and non-veto states which compose that group.53 Since the point of interest is represented by voting shares, I will only take into account normalised indices, that is the Banzhaf and the Shapley-Shubik ones. The results of the analysis are displayed in Table 5. The first aspect which we can notice by looking at Table 5 is that the Banzhaf and the Shapley-Shubik indices give similar results in terms of the direction of increase/ decrease of voting power shares for regional groups, but in many cases differ widely on the size of these shifts. Thus, while under the Banzhaf index the voting power shares for the African group would grow to 10.08 per cent in the Secretary-General Report proposal, to 8.34 per cent in the G4 proposal, and to 11.24 per cent in the Uniting for Consensus proposal, under the Shapley-Shubik index the increase in voting power would be practically insignificant, to 0.9 per cent in the SecretaryGeneral Report, 0.72 per cent in the G4 proposal, and 1.13 per cent in the Uniting for Consensus proposal. The same major differences apply for the increase in the voting power of the Latin-America and the Caribbean group and for the decrease in the voting power of the Western Europe and Others group. In the case of Asia and the Pacific and Eastern Europe, the indices are relatively similar.54 According to the Banzhaf index, all three proposals would have moderately significant effects on representation, with the voting shares of the African group almost doubling in all proposals, the Asian group increasing by between 1 and 2 per cent and the Latin-America and the Caribbean group almost doubling in the G4 and the Uniting for Consensus proposals. Also, the introduction of the United States in the Americas group under the In Larger Freedom Report would have the effect of raising its voting share from 3.3 to 22.01 per cent. Finally, the voting shares of the Western Europe and Others group or Europe in the Secretary-General Report proposal would decrease by between approximately 6 and 10 per cent.55
18
Proposal
Status Quo In Larger Freedom G4 Uniting for Consensus Ezulwini Consensus L69/CARICOM
Indices Africa
Asia and the Pacific
Latin America and the Caribbean
Western Europe and Others
Eastern Europe
0.0330 0.0037
0.5337 0.5923
0.1834 0.1981
B S-S B
0.0496 0.0055 0.1008
0.1999 0.1999 0.2201
S-S B S-S B
0.0090 0.0834 0.0072 0.1124
0.2018 0.2139 0.2011 0.2187
0.0556 0.0048 0.0749
0.4748 0.5890 0.4313
0.1721 0.1975 0.1625
S-S B
0.0113 0.1840
0.2018 0.2688
0.0075 0.0925
0.5829 0.3601
0.1962 0.0915
S-S B S-S
0.1819 0.1831 0.1818
0.2727 0.2718 0.2727
0.0909 0.0918 0.0909
0.3636 0.3612 0.3636
0.0909 0.0912 0.0909
Europe Americas
0.4588
0.2201
0.5862
0.2018
Small islands and developing states
0.0006 0.000002
Journal of International Relations and Development 2015
Table 5 Cumulative voting power shares of regional groups under UNSC reform proposals
Alexandru Volacu A priori voting power distribution
19
Under the Shapley-Shubik index, however, the three moderate proposals for reconfiguration would be almost void of any real impact. In the most optimistic case (the Uniting for Consensus proposal), Africa would gain merely 0.58 per cent voting shares, Asia and the Pacific would get 0.19 per cent, Latin-America and the Caribbean would get 0.38 per cent, Eastern Europe would lose 0.19 per cent, and the Western Europe and Others group would lose 0.93 per cent. Under the SecretaryGeneral Report, the reconfiguration of groups would indeed be the cause of yielding a higher share (of 20.18 per cent) to the Americas, but Europe itself would be left untouched, retaining 58.62 per cent of the total of voting shares. The cases of the Ezulwini Consensus and the L69/CARICOM proposals, however, differ radically from the previous three. First of all, the Banzhaf and the ShapleyShubik indices give approximately the same values (the highest difference margin being 0.39 per cent, for Asia and the Pacific in the case of the Ezulwini Consensus and 0.24 per cent for the Western Europe and Others group in the case of the L69/ CARICOM) for all regional groups, showing a prima facie balance in the distribution of veto rights. Second, the results depict an image of a much more equitable distribution of voting power shares among regional groups,56 with values ranging between approximately 9 per cent for Eastern Europe and Latin-America and the Caribbean and 36 per cent for Western Europe and Others, with the African group at 18 per cent and the Asia and the Pacific group at 27 per cent in the middle. A third interesting fact is that, while the Ezulwini Consensus and the L69/CARICOM proposals are the only ones genuinely endorsed by sponsor states on the basis of regional geographical allegiance, they still only provide the sponsor group with a relatively low voting power compared with other groups. For instance, under the Ezulwini Consensus, the African group would retain only about a half of the voting shares of the Western Europe and Others group, while under the L69/CARICOM proposal, the Latin-America and the Caribbean group would only get about a quarter of the voting shares of the Western Europe and Others group. Thus, the proposals are significantly less radical than thought when we only look at the analysis in the fifth section. Synthesising these results, we can state that the Ezulwini Consensus and the L69/ CARICOM proposals are by far the most favourable to expanding regional representation, while the other three proposals offer mixed results on the question of their effectiveness. While the Ezulwini Consensus and the L69/CARICOM would increase unambiguously the voting power shares of Africa, the Latin-America and the Caribbean group, and Asia and the Pacific, leaving the difference between the ‘weakest’ and ‘strongest’ regional groups57 at only approximately 26 per cent, the other three proposals would increase these voting power shares only moderately under the Banzhaf index analysis, leaving the difference at a higher 35.8 per cent for the Secretary-General Report, 41.92 per cent for the G4 proposal, and 35.64 per cent for the Uniting for Consensus proposal, and they would increase almost unobservably the voting power shares under the Shapley-Shubik index analysis, leaving the
Journal of International Relations and Development 2015
20
difference at much higher figures of 57.72 per cent for the Secretary-General Report, 58.42 per cent for the G4 proposal, and 57.54 per cent for the Uniting for Consensus proposal.58
Discussion The conclusions of the paper can be summarised as follows: out of the three moderate proposals for reforming the Security Council, which did not incorporate an expansion of veto rights,59 the Uniting for Consensus proposal is the most favourable to an increase in the a priori voting power of non-veto states in the Council, both when taken individually and cumulatively, to a decrease in the a priori voting power of veto states, both when taken individually and cumulatively, to an increase in the a priori voting power of the most underrepresented regional groups, that is Africa, Latin-America and the Caribbean, and Asia and the Pacific, and to a decrease in the voting power of the most overrepresented regional group, that is Western Europe and Others. While the G4 proposal and the In Larger Freedom Report of the UN Secretary-General both point generally in the same direction, they call for a weaker, less representation-enhancing reform than that proposed by the Uniting for Consensus group.60 The two major reform proposals that seek an expansion of veto rights to six more states, that is the Ezulwini Consensus and L69/CARICOM, can be considered both the least able to provide better representation and the most able to do it, depending on the philosophical vantage point of the observer. If we atomise the unit of analysis at the level of states, and include these units in the explanans for a theory of representation, then the Ezulwini Consensus and L69/CARICOM are severely detrimental to it, since they basically channel all the a priori voting power in the Security Council into the hands of 11 states, while leaving all other states of the UN with almost no power to influence the passing of resolutions. As shown in the fourth section, under these reform proposals, all non-veto states would hold collectively between 1 and 2 per cent of the total voting power shares in the Security Council. If, however, we atomise the unit of analysis at the level of regional groups, by claiming that states in a certain regional group act as genuine representatives of that group, then the Ezulwini Consensus and the L69/CARICOM are the most representation-enhancing proposals, as they increase significantly the a priori voting power shares of the three most underrepresented regional groups and they decrease significantly those of the Western Europe and Others group. However, as long as the current status quo concerning regional group configuration is in place, atomising at the level of regional groups does not appear to make sense across-the-board, since states in many regional groups have highly diverging positions on a wide number of policies.61 There are various inherent limits to the type of analysis developed here. First of all, it is important to notice that a priori voting power measurements capture a single
Alexandru Volacu A priori voting power distribution
21
dimension of power in a committee and this dimension is complemented in practice by other considerations. O’Neill (1996: 221), for instance, mentions prestige and access to information as other components of power conveyed by a seat in the Security Council as well as the possibility to induce others to support a certain member by using resource incentives. Kuziemko and Werker (2006) further show that membership in the Security Council can also provide other benefits, such as the capacity to accumulate resources by attracting international aid from the UN and the United States, and Dreher et al. (2009) show that the same conclusion can be drawn in relation to funding from the World Bank.62 These aspects raise a problem for measuring voting power via indices, since they are not able to capture the difference between permanent, semi-permanent and non-permanent seats, if the seats entail no veto rights, but the continuous presence of a state in the Security Council may be sufficient for extending its power on dimensions unrelated to a priori voting power.63 Second, it might be claimed that a priori voting power itself is only one out of a number of dimensions relevant to understanding the issue of effective representation and legitimacy of the Security Council. In this vein, it is argued that our focus might go alternatively to the study of other topics connected to voting rules, such as the ‘deliberative deficit’ (Johnstone 2008), imposing strict deadlines on decision making in combination with a limitation on the number of vetoes that a state would be able to cast on the same issue (Winter 1996), ensuring that the Council is free from external influence and that it ‘uses force under Chapter VII only as a last resort’ (Davis 2010: 37), or other such issues. However, using only a priori voting power analysis, we can extract a number of conclusions which might be useful for designing reform proposals that enhance the degree of representation for regional groups and individual members that do not have veto rights. One of the most important aspects of a UNSC reform which seeks to provide a higher degree of representation would be to reconfigure regional groups64 in such a way that both veto and non-veto states, which represent the respective group in the Security Council, would actually adhere, at least broadly, to the policies and positions of those states. A second important aspect would be to ensure a balanced representation of veto rights among these groups since, as the ShapleyShubik index in particular shows, no veto rights for a regional group means a lack of any effective representation. A third aspect, irrespective of any change that might occur in the reconfiguration of regional groups, is that the massive power of veto rights can be balanced (even if only partially) by an increase in the quota required for substantive decision making,65 thus a reform proposal which seeks to increase both the individual and the cumulative power of non-veto states should attempt to raise the quota as high as possible.66 While an increase in the number of non-veto seats in the Council can be an important factor in enhancing representation, it can only be used effectively in combination with a significant increase in the quota required for substantive decision making.
Journal of International Relations and Development 2015
22
In the last few decades, the most effervescent period for the UNSC reform debates was the year 2005, when the Report of the Secretary-General, the G4 proposal, the Uniting for Consensus proposal and the Ezulwini Consensus proposal were circulated and discussed among members of the UN General Assembly. After the discussions stalled and it became clear that no consensual result would be obtained, the debate experienced a significant cool down. While none of these four proposals were ever enacted, studying them is still relevant today. First, they are still generally reflective of the preferences for reformed institutional arrangements which their supporters defended at the time.67 Second, consequently, they are critical in understanding the historical development of the L69/CARICOM proposal and the reasons why enacting this proposal is likely to be met with resistance. And third, they are instrumental in illuminating the way in which voting power distribution changes with incremental shifts in one of its constitutive elements and, thus, they are helpful as tools for constructing future recommendations for reform. As previously mentioned, the L69/CARICOM proposal seems to be gaining some traction68 and with its provisions attempting explicitly to cater to the interests of African, Latin-American and Caribbean states as well as those of developing states and small islands, it is possible that this proposal will become a point of reference for reform discussions in the near future. If this is indeed the case, it is important to understand both the positive and the negative effects which a Security Council reconfigured in the manner proposed by the L69/CARICOM would have on voting power and representation. While such a reform would genuinely yield a more balanced representation at the level of regional groups, it would concentrate all voting power in the hands of states with veto rights, obliterating almost entirely the voting power of non-veto states in the process. Is such a trade-off politically acceptable for a wide majority of the UN member states, and in particular for the permanent members of the Security Council? Should we allow this radical concentration of power in the hands of 11 states, for the sake of a more balanced geographical representation? Or should we rather focus on extending both the individual and the cumulative voting power shares of non-veto member states in the Security Council, as the Uniting for Consensus proposal would do most effectively? I leave these questions unanswered for the time being, as the purpose of this paper was primarily to open such lines of inquiry, thereby contributing descriptively to the present debates surrounding the UNSC reform, rather than to provide answers for them.
Acknowledgements The author would like to thank the audience at the 23rd IPSA World Congress of Political Science, 19–24 July, 2014, Montreal, Canada, where an earlier draft of this paper was presented, and three anonymous reviewers for comments. He is also grateful to Robert Leech for providing him with the programmes used in computing voting power indices.
Alexandru Volacu A priori voting power distribution
23
Notes 1 For a detailed history of the debates surrounding the Security Council reform before 2008, see von Freiesleben (2013). For an in-depth account of contemporary developments, see Swart (2013). 2 By ‘proposed’ I do not necessarily mean that the documents have been submitted on the formal agenda of the UNGA or UNSC, but only distributed among its members. 3 Alongside the 2005 proposal of Kofi Annan, at that time the UN Secretary-General, termed In larger freedom: towards development, security and human rights for all, which was in turn based on a Report of the High-Level Panel on Threats, Challenges and Change. 4 A somewhat more technical definition is provided by Felsenthal and Machover (2004: 3) as follows: ‘by the a priori voting power of a voter under a given decision rule, we mean, roughly speaking, the amount of potential influence over the possible outcomes of divisions which the voter possesses by virtue of the rule. This last qualification — “by virtue of the rule” — is of vital importance’. 5 Namely the Normalised Banzhaf index, the Coleman Power to Prevent Action index, the Coleman Power to Initiate Action index and the Shapley-Shubik index. 6 Which is ‘not reducible to the formalities of decisional rules’ and ‘depends on a complex interaction of real-world factors’ (Felsenthal and Machover 2004: 13). 7 Through Resolution 1991 of the UN General Assembly, adopted on 17 December, 1963. See: http:// www.un.org/documents/ga/res/18/ares18.htm (last accessed on 2 May, 2013). 8 Alongside a requirement of ratification by two thirds of all UN member states. 9 Various arguments can be constructed, however, in favour of retaining veto rights, not as a political compromise, but as a principled position. Voeten (2007: 296), for instance, uses rational choice institutionalism to show that granting veto rights to states, which have the potential to undermine the decision taken by a committee is a Pareto optimal outcome. 10 For comprehensive analyses of the 2005 UNSC reform proposals, which also approach these topics, see Blum (2005) and Cox (2009). 11 With regard to the fourth issue, some proposals have explicit provisions, while others do not. In the latter case, I use a procedure described in the third section in order to estimate the most plausible figure, following current and past practice in the Security Council. My procedure differs from that of Hosli et al. (2011), and while the results are identical for all cases analysed by both papers, in the case of the L69/CARICOM draft proposal, they are slightly different. 12 Thus, the L69 Draft resolution from 2007, the Liechtenstein Draft resolution from 2010 and the S5 Draft resolution from 2012 are not considered in the present analysis. 13 Hereafter the In Larger Freedom Report. 14 Thus, throughout the paper, I refer to the first group as belonging to the category of moderate proposals and the second one as belonging to the category of radical proposals, thereby reflecting the extent of their departure from the status quo. 15 http://daccess-dds-ny.un.org/doc/UNDOC/GEN/N05/270/78/PDF/N0527078.pdf?OpenElement (last accessed on 4 May, 2013). 16 The most recent significant proposal before this was the Razali plan, introduced on 20 March, 1997, which was never subjected to vote. See https://www.globalpolicy.org/security-council/securitycouncil-as-an-institution/41310-razali-reform-paper.html (last accessed on 4 June, 2013). 17 The report contained no less than 222 paragraphs, out of which only 4 (from 167 to 170) targeted specifically the reconfiguration of the Security Council. 18 Equitably distributed among the four regional groups. 19 http://daccess-dds-ny.un.org/doc/UNDOC/GEN/N05/270/78/PDF/N0527078.pdf?OpenElement Article 169 (a), p. 42 (last accessed on 4 May, 2013). 20 http://daccess-dds-ny.un.org/doc/UNDOC/LTD/N05/410/80/PDF/N0541080.pdf?OpenElement (last accessed on 4 May, 2013).
Journal of International Relations and Development 2015
24 21 http://daccess-dds-ny.un.org/doc/UNDOC/LTD/N05/434/76/PDF/N0543476.pdf?OpenElement (last accessed on 5 May, 2013). 22 That is, wider than in the case of the G4 proposal. 23 http://www.safpi.org/sites/default/files/publications/au_executive_council_ezulwini_consensus.pdf (last accessed on 6 May, 2013). 24 http://daccess-dds-ny.un.org/doc/UNDOC/LTD/N05/421/67/PDF/N0542167.pdf?OpenElement (last accessed on 6 May, 2014). 25 http://daccess-dds-ny.un.org/doc/UNDOC/LTD/N07/506/04/PDF/N0750604.pdf?OpenElement (last accessed on 23 December, 2014). 26 The draft resolution itself was ultimately not subjected to vote in the General Assembly; see von Freiesleben (2013: 15–16) for an account of the circumstances surrounding the withdrawal. 27 From Africa, Latin-America and the Caribbean, Asia and the Pacific. 28 The largest expansion of the Security Council proposed to this day. 29 Out of the five, the Ezulwini Consensus and the L69/CARICOM proposals are the only ones that do not make any references to the importance of contributions to the UN, but rather emphasise the paramount importance of effective regional representation in the first case, and a more equitable representation of developing states in the second case. 30 In 2011, the Asia group changed its name to the Asia-Pacific group, in response to states in the Pacific who were previously a part of the group as well. For the full list of regional group membership, see http://www.un.int/wcm/webdav/site/gmun/shared/documents/GA_regionalgrps_Web.pdf (last accessed on 6 June, 2013). 31 This method is used implicitly in most papers on voting power in the Security Council (for instance, Schwodiauer 1968: 4) and made explicit in others (for instance, Miroiu 2007: 90). The system of linear equations is designed to capture the lowest number required for a decision to pass under the weighted voting system (first equation) and the highest number for which a decision cannot pass (second equation). 32 I will not include the detailed calculations for the five reform proposals in the paper. Since the mathematical apparatus is extremely simple, the reader can easily verify their correctness. 33 Hosli et al. (2011: 176) use an alternative procedure, of choosing the quota closest to 60 per cent of the total number of members. Since there is no officially mandated Security Council procedure, the historical pattern of majority requirements in the Council can be legitimately interpreted in both ways, but as the number of members increases, the two estimations will yield increasingly divergent results. In this case, estimations made on the Secretary-General Report and the Ezulwini Consensus yield identical results, but estimations made on the L69/CARICOM proposal using the Hosli et al. (2011) method would yield a quota of 16, instead of 15. The results, however, do not differ significantly between the two quotas. 34 The main reason why I use these particular four indices, instead of other more recently developed ones (for instance, Deegan and Packel 1978; Johnston 1978; Holler 1982; Hoede and Bakker 1982) is that they are the the most well-known and widely used indices in the literature on voting power and, at the same time, they embody the two major directions of development, that is the I-power (power as influence) approach and the P-power (power as prize) approach (Felsenthal and Machover 1998). For critiques of some of these indices or the whole a priori voting power approach itself, see Garrett and Tsebelis (1999), Felsenthal and Machover (2002), Albert (2003), Gelman et al. (2004) or Kaniovski (2008). 35 See Leech (2002a: 6–10) for the formal framework. 36 Namely, a coalition composed of all players, ordered in a specific sequence. 37 The indices were computed through the ipgenf and ssgenf programs developed by Dennis Leech and Robert Leech. 38 A major difference between the results described here and those of Hosli et al. (2011: 179–80) concerns the Uniting for Consensus and the G4 proposals, under the Normalised Banzhaf index (this is
Alexandru Volacu A priori voting power distribution
25
39 40 41 42 43 44 45 46 47 48 49 50 51
52
53
54
the only comparable index between our studies). While here the index yields different values, in Hosli et al. (2011) the Uniting for Consensus and the G4 proposals yield exactly the same results for the Banzhaf index. Since the index used and all other conditions are identical, what is the source of this divergence between studies? The problem lies with the fact that Hosli et al. (2011) have failed to take into account the explicit specifications of the Uniting for Consensus and the G4 proposals regarding the majority quota required for decision making. Thus, they simply apply the procedure of approximating the closest value to 60 per cent of the total number of votes, as mentioned in a previous endnote, in order to obtain the quota. While this is consistent with the actual majority requirement in the Uniting for Consensus proposal, the G4 proposal states explicitly that the Charter will be amended so as to ‘require the affirmative vote of 14 of 25 members of the Security Council for a decision’, which is lower than the 15 member quota used by Hosli et al. (2011). The percentages are relative to the highest values which the indices can take, namely 1, and are not calculated as a percentage of the shift from the status quo. The impact for both proposals is identical in these respects for the first five digits after the decimal point. To a total value of 98.48 and 98.99 per cent, respectively. If we take into account the fact that even veto states do not exceed the value of 3 per cent in either the status quo or the reform proposals, we notice that the 0.02 per cent increase is not a very small value. But with a higher intensity. Which cannot be altered as long as veto rights remain in place. For all indices except Shapley-Shubik. At the same time, it shows that Hosli et al.’s (2011) neglection of the G4 proposed majority quota is highly relevant to the results. With the L69/CARICOM proposal being slightly more radical than the Ezulwini Consensus in all cases. While we cite only from the Ezulwini Consensus, the same clauses (with minor stylistic adjustments) are also present in the L69 and the CARICOM proposals. http://daccess-dds-ny.un.org/doc/UNDOC/LTD/N05/421/67/PDF/N0542167.pdf?OpenElement (last accessed on 28 April, 2013). At the time when this article was written. For instance, neither of the four salient definitions of Pitkin (1967) do not seem to overlap with the term representation, as used in the proposals described in the second section, although the descriptive definition of representation appears to resemble it to some extent. See for example the In Larger Freedom Report (p. 43) which seeks to replace the five regional groups currently in existence with four, by replacing the Eastern Europe, Western Europe and Others and Latin-America and the Caribbean groups with Europe and Americas, or the L69/CARICOM proposal which adds a new group for developing states. An alternative course of action would have been to consider that the regional groups vote en bloc, namely that all states vote the same way all the time. Such an analysis yields interesting results as well, with the African and the Latin-American and Caribbean groups shown to have 0 voting power under all indices for both the status quo and any reform proposal except for the Ezulwini Consensus and the L69/CARICOM proposal, where they would have voting power equal to all other regional groups. The assumption of bloc voting, however, is severely restrictive and unreflective of reality; therefore, I will not adopt it here. Using cumulative voting power shares requires no such unrealistic assumption as that of unitary bloc voting. This is a natural consequence of the fact that veto rights weigh much more under the Shapley-Shubik index than under the Banzhaf index. When we have no states with veto rights in a group, the ShapleyShubik index tends to grant it a lot less power than the Banzhaf index and, conversely, when there are a lot of veto states in a group, the Shapley-Shubik index estimates its voting power at a much higher level than the Banzhaf index.
Journal of International Relations and Development 2015
26 55 Under the G4 proposal and the Uniting for Consensus proposal, Eastern Europe would also have a slightly lower voting share, of approximately 1–2 per cent. 56 A new group, the Small islands and developing states, was also introduced in the L69/CARICOM proposal, which would hold 0.06 per cent of the voting power under the Banzhaf index and 0.0002 per cent under the Shapley-Shubik index. 57 Excepting the group created for developing states under the L69/CARICOM proposal. 58 The difference in the present configuration of the Security Council being of 50.07 per cent for the Banzhaf index and 58.86 per cent for the Shapley-Shubik index. 59 I also include the G4 proposal in this category, since their proposal simply called for a reopening of the discussion on the matter at a later time in the future. 60 With the former being lighter still than the latter. 61 See, for instance, the Eastern Europe group, which, on the one hand, includes the Russian Federation and, on the other hand, many states which are currently much more closely aligned to the positions of the Western Europe and Others group. 62 De Mesquita and Smith (2010) argue, however, that the aid increases have a detrimental effect on the welfare of citizens in recipient states, both from an economic standpoint and from the perspective of civil liberties in non-democratic regimes. 63 This is, in fact, one of the most controversial issues between the proposals, representing the main difference between Plan A and Plan B in the Secretary-General Report and one of the main differences between the G4 proposal and the Uniting for Consensus proposal. 64 The proposal of the In Larger Freedom Report would not fare much better though, since it would, for instance, insert the Russian Federation in the Europe group, appearing to give it, as a regional group, more power than it would actually hold. 65 As iterated before already, this is the primary reason why the Uniting for Consensus group yielded a more equitable distribution of voting power than the G4 proposal and the Secretary-General Report. 66 Subject to constraints of efficiency and seeking to avoid objections from current P5 members for reasons concerned with political acceptability. 67 This only applies to the draft resolutions. 68 Although whether or not the L69 proposal is indeed effectively supported by as much as 80 states, as claimed in 2012, is debatable (Swart 2013: 50).
References Albert, Max (2003) ‘The Voting Power Approach: Measurement Without Theory’, European Union Politics 4(3): 351–66. Banzhaf, John (1965) ‘Weighted Voting doesn’t Work: A Mathematical Analysis’, Rutgers Law Review 19(3): 317–43. Blum, Yehuda (2005) ‘Proposals for UN Security Council Reform’, American Journal of International Law 99(3): 632–49. Bourantonis, Dimitris (2005) The History and Politics of UN Security Council Reform, Milton Park: Routledge. Brams, Steven (1975) Game Theory and Politics, New York: Free Press. Butler, Richard (2012) ‘Reform of the United Nations Security Council’, Penn State Journal of Law & International Affairs 1(1): 23–39. Coleman, James (1971) ‘Control of Collectivities and the Power of a Collectivity to Act’, in Bernhardt Lieberman, ed., Social Choice, 269–300, New York: Gordon and Breach. Cox, Brian (2009) ‘United Nations Security Council Reform: Collected Proposals and Possible Consequences’, South Carolina Journal of International Law and Business 6(1): 89–128.
Alexandru Volacu A priori voting power distribution
27 Davis, Ryan (2010) ‘An Unrealistic Proposal: An Argument Against the Enlargement of the United Nations Security Council’, International Affairs Review 19(1): 19–41. De Mesquita, Bruce Bueno and Alastair Smith (2010) ‘The Pernicious Consequences of UN Security Council Membership’, Journal of Conflict Resolution 54(5): 667–86. Deegan, John and Edward Packel (1978) ‘A New Index of Power for Simple n-person Games’, International Journal of Game Theory 7(2): 113–23. Dreher, Axel, Jan-Egbert Sturm and James Vreeland (2009) ‘Development Aid and International Politics: Does Membership on the UN Security Council Influence World Bank Decisions?’ Journal of Development Economics 88(1): 1–18. Fassbender, Bardo (1998) UN Security Council Reform and the Right to Veto: A Constitutional Perspective, The Hague: Kluwer Law International. Felsenthal, Dan and Moshe Machover (1998) The Measurement of Voting Power: Theory and Practice, Problems and Paradoxes, Cheltenham: Edward Elgar. Felsenthal, Dan and Moshe Machover (2002) ‘Models and Reality: The Curious Case of the Absent Abstention’, Homo Oeconomicus 19(3): 297–310. Felsenthal, Dan and Moshe Machover (2004) ‘A Priori Voting Power: What is It All About?’ Political Studies Review 2(1): 1–23. Garrett, Geoffrey and George Tsebelis (1999) ‘Why Resist the Temptation to Apply Power Indices to the European Union?’ Journal of Theoretical Politics 11(3): 291–308. Gelman, Andrew, Jonathan Katz and Joseph Bafumi (2004) ‘Standard Voting Power Indexes Do Not Work: An Empirical Analysis’, British Journal of Political Science 34(4): 657–74. Gould, Matthew and Matthew Rablen (2013) ‘Equitable Representation in the Councils of the United Nations: Theory and Application’, CESifo Working Paper No. 4519. Munich. Gowan, Richard and Nora Gordon (2014) Pathways to Security Council Reform, New York: New York University Center on International Cooperation. Hoede, Cornelis and Rene Ronald Bakker (1982) ‘A Theory of Decisional Power’, Journal of Mathematical Sociology 8(2): 309–22. Holler, Manfred (1982) ‘Forming Coalitions and Measuring Voting Power’, Political Studies 30(2): 262–71. Hosli, Madeleine, Rebecca Moody, Bryan O’Donovan, Serguei Kaniovski and Anna Little (2011) ‘Squaring the Circle? Collective and Distributive Effects of United Nations Security Council Reform’, Review of International Organizations 6(2): 163–87. Johnston, Ron (1978) ‘On the Measurement of Power: Some Reactions to Laver’, Environment and Planning A 10(8): 907–14. Johnstone, Ian (2008) ‘Legislation and Adjudication in the UN Security Council: Bringing down the Deliberative Deficit’, American Journal of International Law 102(2): 275–308. Junn, Robert and Tong-Whan Park (1977) ‘Calculus of Voting Power in the UN Security Council’, Social Science Quarterly 58(1): 104–10. Kaniovski, Serguei (2008) ‘The Exact Bias of the Banzhaf measure of Power when Votes are Neither Equiprobable Nor Independent’, Social Choice and Welfare 31(2): 281–300. Kuziemko, Ilyana and Eric Werker (2006) ‘How Much is a Seat on the Security Council Worth? Foreign Aid and Bribery at the United Nations’, Journal of Political Economy 114(5): 905–30. Leech, Dennis (2002a) ‘Computation of Power Indices’, Warwick Economics Research Paper Series No. 644. Coventry. Leech, Dennis (2002b) ‘The Use of Coleman’s Power Indices to Inform the Choice of Voting Rule with Reference to the IMF Governing Body and the EU Council of Ministers’, Warwick Economics Research Paper Series No. 644. Coventry. Miroiu, Adrian (2007) The Foundations of Politics: Rationality and Collective Action, Iasi: Polirom. O’Neill, Barry (1996) ‘Power and Satisfaction in the United Nations Security Council’, Journal of Conflict Resolution 40(2): 219–37.
Journal of International Relations and Development 2015
28 Penrose, Lionel S (1946) ‘The Elementary Statistics of Majority Voting’, Journal of the Royal Statistical Society 109(1): 53–57. Pitkin, Hanna (1967) The Concept of Representation, Berkeley: University of California Press. Russett, Bruce, Barry O’Neill and James Sutterlin (1997) ‘Breaking the Restructuring Logjam’, in Bruce Russett, ed., The Once and Future Security Council, 153–72, New York: St Martin’s Press. Schlichtmann, Klaus (2011) ‘An Enduring Concept for Security Council Reform’, Beijing Law Review 2(2): 97–110. Schrijver, Niko (2007) ‘Reforming the UN Security Council in Pursuance of Collective Security’, Journal of Conflict & Security Law 12(1): 127–37. Schwodiauer, Gerhard (1968) ‘Calculation of a Priori Power Distributions for the United Nations’, in, Research Memorandum, 24. Vienna: Institut fur Hohere Studien. Shapley, Lloyd and Martin Shubik (1954) ‘A Method for Evaluating the Distribution of Power in a Committee System’, American Political Science Review 48(3): 787–92. Slaughter, Anne-Marie (2005) ‘Security, Solidarity, and Sovereignty: The Grand Themes of UN Reform’, American Journal of International Law 99(3): 619–31. Strand, Jonathan and David Rapkin (2011) ‘Weighted Voting in the United Nations Security Council: A Simulation’, Simulation Gaming 42(6): 772–802. Swart, Lydia (2013) ‘Reform of the Security Council: 2007–2013’, in Lydia Swart and Estelle Perry, eds, Governing and Managing Change at the United Nations: Security Council Reform from 1945 to September 2013, 23–59, New York: Center for UN Reform Education. Swart, Lydia and Estelle Perry (2013) Governing and Managing Change at the United Nations: Security Council Reform from 1945 to September 2013, New York: Center for UN Reform Education. Von Freiesleben, Jonas (2013) ‘Reform of the Security Council: 1945–2008’, in Lydia Swart and Estelle Perry, eds, Governing and Managing Change at the United Nations: Security Council Reform from 1945 to September 2013, 1–22, New York: Center for UN Reform Education. Voeten, Erik (2007) ‘Why no UN Security Council Reform? Lessons for and from Institutionalist Theory’, in Dimitris Bourantonis, Kostas Infantis and Panayotis Tsakonas, eds, Multilateralism and Security Institutions in the Era of Globalization, 288–305, New York: Routledge. Weiss, Thomas (2003) ‘The Illusion of UN Security Council Reform’, Washington Quarterly 26(4): 147–61. Weiss, Thomas and Karen Young (2005) ‘Compromise and Credibility: Security Council Reform?’ Security Dialogue 36(2): 131–54. Winter, Eyal (1996) ‘Voting and Vetoing’, American Political Science Review 90(4): 813–23.
About the author Alexandru Volacu is a Ph.D. Candidate and Teaching Assistant at the National School of Political and Administrative Studies (SNSPA) in Bucharest, Romania. He is currently an editor of the Romanian Journal of Society and Politics. His research interests include voting power analysis, the UN System, political philosophy, ethics, game theory, social choice theory, and public choice theory.