The Proliferation of Developing Country Classifications

Abstract

We study the external and internal proliferation of country classifications in development policy. The number of classifications increased from four (1985) to 17 (2013) when the average in our sample of 111 developing countries exceeded three classifications per country. Based on historical overview and comparative case study for land-locked development countries and small-island development states (geographically defined classifications without overlap) we find that internal proliferation is associated with lacking a clear rationale, no definition of country characteristics, and possibly the direct involvement of developing countries in designing the category. External proliferation may reflect antinomic delegation, geopolitical and bureaucratic motives.

1. Introduction

One of the remarkable phenomena in the arena of international development policy is the proliferation of classifications and categories of developing countries introduced by International Organisations since the early 1990s (Table 1). In the 1960s, 1970s and 1980s, basically only four classifications were used. Since 1990, in contrast, 13 new classifications have been introduced (or 26 if one also counts the sub-categories). Let it be clear from the start that classification per se is useful, both analytically and for operational purposes (Nielsen, 2013; Tezanos Vázquez & Sumner, 2013). Analytically, classification helps to understand and disentangle problems in an increasingly complex and heterogeneous world: how and why do countries differ in their developmental achievements and processes? Operationally, classifications are important for evidence based differential treatment of groups of countries: which kind of country gets what conditions and what resources? While we see a clear need for classifications, we also note that the current debate on the existing classifications is controversial and vivid (see, for example, Koch, 2015 on the EU debate) and reflects both on the changing geography of poverty and the need to consider development from a multidimensional perspective rather than the predominantly used per capita income indicators.

Table 1. The proliferation of categories of developing countries

Created	Category	Sub-categories
1957	Landlocked Developing Countries (LLDC)
1971	Least Developed Countries (LDC)
1980s	World Bank’s income level categories	low-income countries (LIC) lower-middle-income countries (LMIC) upper-middle-income countries (UMIC) high-income countries (HIC)
1985	World Bank’s small island exception list
1990	Human Development Index	low human development (LHD) medium human development (MHD) high human development (HHD)
1991	Alliance of Small Island States (AOSIS)^4
1994	UNCTAD’s unofficial list of Small Island Developing States (SIDS)
1996	Heavily Indebted Poor Countries (HIPC)	completion point HIPCs decision point HIPCs pre-decision point HIPCs
1997	UNDESA’s (SIDSnet) list of SIDS
2001	UN-OHRLLS’ list of SIDS
2001	WTO’s Small, Vulnerable Economies (SVEs)
2001/2002	Low Income Countries Under Stress (LICUS)	severe LICUS core LICUS	LICUS with prolonged political crisis LICUS in fragile transition LICUS with weak governance and slow progress LICUS with deteriorating governance
after 2002	Fragile States (FS)	core fragile states marginal fragile states	post-conflict countries re-engaging and turnaround countries deteriorating situations prolonged impasse
2003	Transit Developing Countries (TDC)
2004	Graduating LDCs
2007	Structurally Weak, Vulnerable and Small Economies (SWVSE)
2008	UNESCO’s list of SIDS

Sources: http://www.unohrlls.org/en/home/, http://data.worldbank.org/about/country-classifications, http://aosis.org/about-aosis/, http://www.sidsnet.org/, http://web.worldbank.org/WBSITE/EXTERNAL/TOPICS/EXTDEBTDEPT/0,contentMDK:20260411~menuPK:64166739~pagePK:64166689~piPK:64166646~theSitePK:469043,00.html, http://www.wto.org/english/thewto_e/minist_e/min11_e/brief_svc_e.htm, http://ieg.worldbankgroup.org/Data/reports/licus_approach_paper.pdf, http://unctad.org/en/Docs/tdb54crp4_en.pdf (19/1/13) http://www.un.org/en/development/desa/policy/cdp/ARES67221_en.pdf

A number of observers, moreover, have expressed concerns about the extent of proliferation. According to Alonso et al. (2014, p. 31) ‘the recent proliferation of categories has created a lot of confusion and fragmentation, and many categories generate substantial problems’. Tezanos Vázquez and Sumner (2015, p. 19) note that a very large number of classifications carries an obvious cost in terms of complexity. The proliferation reflects a lack of both coherence and sound analytical underpinnings; it also implies that the allocation of differential treatment becomes increasingly vulnerable for political influence. Existing taxonomies suffer from lack of clarity with regard to how they distinguish among country groupings.¹ Alonso et al. (2014, p. 5) therefore conclude that ‘Great caution should be exercised in devising new country categories’. Indeed, the intention of country classifications is to provide analytical clarity and rules that govern special treatment of groups of developing countries. One of the most important reasons why multilateral and bilateral donors support certain country groupings is because these classifications are used as tools to allocate development resources on the basis of technical criteria, even though aid allocation processes are known to be primarily political,² as already observed by Drezner (2009, p. 67): ‘Paradoxically, after a certain point institutional and legal proliferation can shift global governance structures from a Lockean world of binding rules to a Hobbesian world of plastic rules’.³ Rather than creating predictability, rationality and transparency about rules and principles and protecting developing countries against the vagaries of large countries, the proliferation of classifications injects the global governance system with discretion. Also from this perspective, it is important to understand what drives this process in order to be able to contain further proliferation.

The changes in the number of countries in a category are sometimes puzzling. For example, the earliest UN list of Small Island Developing States (SIDS) that we were able to identify (UNCTAD’s [United Nations Conference on Trade and Development] 1994 unofficial list of SIDS) contained 29 island states, in contrast to the 2001 UN-OHRLLS (United Nations Office of the High Representative for the Least Developed Countries, Landlocked Developing Countries and Small Island States) list of SIDS that contains 52 states some of which are not islands and some of which are not small.

with N = the number of other categories and X is the number of countries in the category (j) and that overlap with another category (i).

Secondly, the overlap of the categories (Table 2) illustrates the complexity of the current developing country differentiation landscape. The average overlap reported in the bottom line of Table 2 differs by category. Comparatively low levels of overlap occur for medium human development (MHD) and lower-middle-income countries (LMIC). For heavily indebted poor countries (HIPC), least developed countries (LDC) and low-income countries (LIC) we find an overlap of about 50 per cent. To a large extent overlap obviously reflects the fact that these classifications are based on different criteria, but the lack of a methodology or consensus on how to classify countries also matters (see Madrueño-Aguilar, 2015, pp. 1–18 for an overview of the shortcomings and drawbacks regarding the composition of the development taxonomy and the notion of middle-income countries; and Nielsen, 2013, pp. 1090–1092 for a brief overview of the different development taxonomies in the Bretton Woods institutions).

Table 2. Matrix of selected developing countries’ categories

	LDC	LLDC	LIC	LMIC	LHD	MHD	HIPC	FS
LDC	48	17	30	18	39	6	31	24
LLDC		31	15	11	14	12	13	5
LIC			35	0	28	3	26	16
LMIC				57	15	31	12	13
LHD					46	0	32	22
MHD						47	6	5
HIPC							39	18
FS								36
Average overlap^a (%)	49	40	48	25	47	19	51	41

Source: Calculations based on Fialho (2015)

Note: ^a Average overlap for category j is defined as

In this article we will study both the external proliferation (the number of categories) and the internal proliferation (the dynamic development of the number of countries in a category). From the perspective of this paper, overlap in the context of an increasing number of categories is mainly problematic for giving analytical and operational order.⁵ The internal proliferation is not always problematic, for example if new countries emerge or existing countries, due to adverse developments, meet the criteria for inclusion.⁶ This is, however, different when lists of countries in a classification grow without changes in country conditions or classification criteria.

Given the stylised facts that we discussed above, we are particularly interested in finding out under what conditions both forms of proliferation can be reduced. Our methodology is based on analysis of documents and background interviews, providing a comparative case study for two classifications (the landlocked developing countries and the small island developing states) that would both a priori appear to be based on geographical characteristics and thus less prone to proliferation. Importantly, given the geographical nature of the categories, (changes in) overlap cannot blur our analysis.

The remainder of this article is structured as follows. Section 2 provides a general picture of the developing country differentiation landscape and the special and differential treatment related to the multilateral classifications. In order to understand the process of proliferation, Section 3 discusses the drivers of the main players: international organisations, developed countries and developing countries. Section 4 provides a case study. We investigate why the classifications of land-locked developing countries and small island development states show completely different dynamics in terms of the speed and extent of proliferation of the countries in the classification. Section 5 discusses implications and draws conclusions.

2. An Overview of Multilateral Classifications

Starting in 1957 17 multilateral efforts have been implemented to (re)cluster and differentiate developing countries according to common characteristics thought to delay or impede their development and economic growth. Attached to these clusters/groups are different sets of special and differential treatment/benefits (Table 3). Country differentiation results in development-promoting benefits being assigned, in principle, to different categories of developing countries.

Table 3. General overview of selected developing countries categories and their respective benefits

Category	Benefits/Special and differential treatment
LDC	ODA: bilateral donors’ commitment to allocate 0.15% of GNP as ODA to LDCs (according to the Brussels Declaration and Programme of Action) Several UN organisations target technical cooperation programmes to LDCs or earmark a proportion of their budgets for LDCs Preferential market access: Generalised System of Preferences (GSP – non-reciprocal) and Global System of Trade Preferences (GSTP – among developing countries, is a reciprocal scheme available for signatories) Special treatment regarding WTO obligations (for LDCs that are WTO member states) and WTO accession Trade-related capacity building: through the Integrated Framework for Trade-related technical assistance to LDCs; a multi-agency, multi-donor programme to assist LDCs in developing trade-related capacities Financial support: provided by the UN (and its organisations) for the participation of LDC representatives in annual sessions of the General Assembly and in other UN meetings Entitlement to 90% discount in LDC contributions to UN peacekeeping operations LDC contributions to the UN regular budget are capped at 0.01% of the total UN budget
LLDC	Almaty Programme of Action: establishes general guidelines/recommendations for the special and differential treatment of LLDCs in the areas of transit policy, infrastructure development, international trade and ODA EU-ACP Agreement: previews special treatment for ACP countries that are LLDCs Debt relief under HIPC initiative: previewed for LLDCs that are also HIPC
SIDS	World Bank’s small island exception: provision of IDA resources to small islands whose per capita income is above the IDA eligibility cut-off but have no or very limited creditworthiness, which impedes their access to IBRD borrowing EU-ACP Agreement: previews special treatment for ACP countries that are also SIDS
LIC	Access to the World Bank’s IDA concessional lending Several bilateral donors use the LIC classification to determine ODA allocation Regional and multilateral financial institutions often allocate concessionary financing to developing countries based on the LIC classification
LMIC	Some are considered blend countries and are eligible for both IDA concessional loans (due to their low per capita incomes) and non-concessional IBRD loans (because they are financially creditworthy)
HIPC	The Multilateral Debt Relief Initiative (MDRI) allows for 100% relief on eligible debts by three multilateral institutions – the IMF, the World Bank, and the African Development Fund (AfDF) – for countries completing the HIPC Initiative process.
FS	World Bank takes a differentiated approach to fragile states by adapting its interventions to specific challenges faced by these countries in particular by enabling more risk taking and a quicker operational process and implementation support

Note: IDA, International Development Association.

Source: compilation based on UN (2003, 2006), UNDESA (2004), OECD-DAC (http://www.oecd.org/dacfragilestates/), http://www.imf.org/external/np/exr/facts/hipc.htm and http://web.worldbank.org/WBSITE/EXTERNAL/NEWS/0, contentMDK:20042303~menuPK:34480~pagePK:64257043~piPK:437376~theSitePK:4607,00.html

The UN is the mother of four important categories. In November 1971 the General Assembly first approved the list of what it considered ‘hardcore’⁷ Least Developed Countries (LDC) and through this category donors should provide special treatment to these countries, in terms of aid, trade and technical assistance (Fialho, 2012). From the initial 25 LDCs identified in 1971, the category grew to a total of 51 countries as more countries became independent in the 1970s and the poor performance of other developing countries made them join the group in the 1980s and 1990s. Present membership is 48 LDCs following four graduation cases (Botswana [1994], Cape Verde [2008], Maldives [2011] and Samoa [2014]), and the inclusion of the newly created Republic of South Sudan, in 2012 (UNCDP, 2012, p. 1). Two categorisations, apparently exclusively based on geographical factors and widely used, are the landlocked developing countries (LLDC) and the small island developing states (SIDS) that we discuss in detail in Section 4. A fourth category – very much inspired by the work of Mahbub ul Haq and Amartya Sen (see Ul Haq, 2003) – was launched in 1990 by the United Nations Development Program (UNDP), ranking countries according to their human development index: (i) low (LHD), (ii) medium (MHD), (iii) high (HHD) and (iv) very high human development (VHHD).⁸ According to the UNDP, this method ‘introduced a new way of measuring development by combining indicators of life expectancy, educational attainment and income into a composite human development index, the HDI’,⁹ introducing a new paradigm in terms of framing and definition of development, at the heart of which laid the notion of human capabilities (Jolly, Emmerij, Ghai, & Lapeyre, 2004, p. 179). Unlike the other categories, it does not exactly recommend or advocate for the special and differential treatment of certain countries.

Within the Bretton Woods institutions, developing country differentiation has meant the creation of different groups of countries over the years. In the 1980s, the World Bank established a categorisation based exclusively on income level, as measured by per capita Gross National Income (GNI), which today comprises: (i) low-income countries (LIC), (ii) lower-middle-income countries (LMIC) (iii) upper-middle-income countries (UMIC), (iv) high-income countries (HIC), and (v) high-income OECD members.¹⁰ (For the purpose of this article the focus will be exclusively on the bottom two categories: which congregate the world’s 90 poorest countries). Essentially, these are ‘analytical income categories … based on the Bank’s operational lending categories (civil works preferences, IDA eligibility, and so forth)’.¹¹ Accordingly:

These operational guidelines were established based on the view that since poorer countries deserve better conditions from the Bank, comparative estimates of economic capacity needed to be established. GNI, a broad measure, was considered to be the best single indicator of economic capacity and progress; at the same time it was recognised that GNI does not, by itself, constitute or measure welfare or success in development. GNI per capita is therefore the Bank’s main criterion of classifying countries.¹²

The heavily indebted poor country (HIPC) initiative, established in 1996, is a joint approach promoted by the International Monetary Fund (IMF) and the World Bank, aiming at ‘ensuring that no poor country faces a debt burden it cannot manage’.¹³ The majority of HIPCs are overly indebted low-income countries (according to the World Bank’s income level categorisation) that can benefit from IMF and World Bank assistance to service and reduce their debt, ideally bringing it to a sustainable level. The HIPC initiative

is open to the world’s poorest countries … that: (i) are eligible only for highly concessional assistance such as from the World Bank’s International Development Association (IDA) and the IMF’s Poverty Reduction and Growth Facility …; (ii) face an unsustainable debt situation …; and (iii) have a proven track record in implementing strategies focused on reducing poverty and building the foundation for sustainable economic growth.¹⁴

By the start of 2013, 39 such countries¹⁵ (most of them in sub-Saharan Africa) had been found eligible to receive special treatment (more specifically, debt relief) under the HIPC initiative.

Another category of countries – this one still evolving towards a clearer definition – is the fragile states group,¹⁶ which replaced the World Bank’s concept of low-income countries under stress (LICUS).¹⁷ State fragility is a categorising concept applied by the World Bank and the OECD-DAC (Organisation for Economic Cooperation and Development – Development Assistance Committee). Both organisations define this concept by referring to ‘low income and to the World Bank CPIA¹⁸ rating, and they differ only by the CPIA threshold’ (Guillaumont, 2009, p. 14). More specifically, fragile states are

defined as having either: a) a composite World Bank, African Development Bank and Asian Development Bank Country Policy and Institutional Assessment rating of 3.2 or less; or b) the presence of a United Nations and/or regional peace-keeping or peace-building mission …, with the exclusion of border monitoring operations, during the past three years.¹⁹

In addition, ‘[c]ountries are considered core fragile states if their CPIA is below 3.0 … [and] marginal fragile states if their CPIA score is between 3.0 and 3.2’.²⁰ These are countries that ‘face severe development challenges such as insecurity, weak governance, limited administrative capacity, chronic humanitarian crises, persistent social tensions, violence or the legacy of civil war’.²¹ Among the 36 countries categorised as fragile states²² by the start of 2013, the World Bank differentiates between four situations: (i) post-conflict countries, (ii) re-engaging and turnaround countries, (iii) deteriorating situations, and (iv) prolonged impasse.²³

The latest addition to the differentiation landscape is the implicit category of graduating LDCs or LDCs in transition (UN, 2013). The new (implicit) category ensures extended – eventually to be phased out – access to differential treatment, continued monitoring by the UN and availability of core UN bureaucratic functions.

3. Drivers of Proliferation

One of the most obvious drivers of the external proliferation (the number of categories) is (the recognition of) substantial and increased heterogeneity of the group of developing countries as noted by van Bergeijk and Van Marrewijk (2013, p. 1):

UNCTAD, the IMF and the World Bank seem to be involved in an intellectual competition to find ever-new acronyms to re-classify and re-group the developing world. The existence of CITs, CAFS, LDCs, LICs, LICUS, HIPCs, SIDs and SWVSEs testifies of the substantial amount of macro and political heterogeneity that is characteristic of what once was perceived to be a more or less coherent group of Third World countries.²⁴

Alonso et al. (2014, figures 1 and 2) empirically illustrate this increased heterogeneity in terms of average income levels for the broad group of developing countries, in particular since the 1990s when the coefficient of variation of per capita GDP in the developing world became larger than the world average. Heterogeneity of countries, however, does not explain the internal dynamics and the increased overlap that we have observed above. In order to more fully understand the process of proliferation, it is important to consider what is in it for the main players: the international organisations, the developed countries and the developing countries.

3.1. International Organisations

The international organisations (IOs) are the main suppliers of the categories; only one of the existing lists can be considered as a co-creation. The management of country lists is an important activity of the bureaucrats employed by the IOs. These tasks involve all aspects of their authority: rule-making, definition and creation of categories and of the discourse in ‘their’ policy arena, creation and reshaping of incentives and/or interests and (political) organisation (Barnett & Finnemore, 1999, p. 699). Indeed, the bureaucrats propose lists of countries to be offered differential treatment, determine the countries to be actually included on those lists and negotiate the benefits of differential treatment. In this regard, Haftel and Thompson (2006, p. 261) mention that ‘some secretariats and commissions can initiate and recommend policies and thereby promote the goals of the organisation, prerogatives that greatly enhance bureaucratic authority’.

The differentiation of countries and classifications often are meant to serve a useful purpose. In general, for the IOs, the categories of developing countries work primarily (but not exclusively) as internal policy instruments. Firstly, classifications are used by international organisations in order to increase their efficiency in policymaking and policy implementation. This is to be achieved by customisation of policy-prescriptions according to particular development issues/conditions (for example, landlockness, indebtedness, islandness, fragility, and so forth) and helped by in-depth knowledge of, and specialisation in, development issues faced by different groups of developing countries. New problems create new tasks and in a number of cases that requires new instruments and may thus lead to new categorisations. Examples are the World Bank that systematises different income categories to facilitate its loan/grant allocation and the IMF/World Bank that institutionalised the HIPC initiative to assist the implementation of their debt relief programmes. Secondly, new categories may provide valuable communication tools. The UNDP, for example, created the human development ranking to help it advocate a more comprehensive type of development that went beyond per capita production and that allowed for a stronger focus on health and education. Another example is the World Bank (and the OECD-DAC) category of fragile states that is meant to create awareness of policymakers when considering development prescriptions for these countries. Thirdly, categorisations may be a gateway to increased donor attention and funding by the public sector, by NGOs and by commercial parties. An example of the latter is the HIPC initiative that was complemented by the Inter-American Development Bank and where, according to the IMF and World Bank, moral suasion was used to ensure voluntary creditor participation in the initiative.²⁵

It can therefore be argued that the creation of a new category may reflect an effort to normalise/depoliticise development support with the aim of stimulating a needs-based/recipient-focused type of assistance allocation. A category can be helpful in advocating the prioritisation of assistance to certain countries and subjecting international organisations and donor countries to closer scrutiny in terms of international pledges/commitments made towards specific groups of developing countries.

However, the proliferation of categories may also be a reflection of wasteful competition between and even within international organisations (Frey, 2008). IOs, as any bureaucracy, try to maximise their independence and one way to do this is by using discretionary rather than rule-based approaches (Vaubel, 1996). Many of the categories reflect an ever increasing discretionary approach that undercuts or modifies criteria. An example is the LDC category for which the criteria have been revised several times in order to ‘prevent’ countries from graduating (Fialho, 2012, 2015). Setting and adjustment of the rules of the classification game potentially influences the stability of classifications within categories. Indeed, an

elementary feature of bureaucracies is that they classify and organize information and knowledge. This classification process is bound up with power … The ability to classify objects, to shift their very definition and identity, is one of bureaucracy’s greatest sources of power … Categorization and classification are a ubiquitous feature of bureaucratization that has potentially important implications for those being classified. To classify is to engage in an act of power. (Barnett & Finnemore, 1999, pp. 710–711)

The proliferation of overlapping categories can thus also be seen as a reflection of self-interested actors maximising the likelihood of the organisation’s survival, amongst others by increasing its depth and scope (Schneider & Tobin, 2011), and may reflect continued existence and decision-making power of certain bureaucratic structures that over the years: (i) have been purposely created just to manage these categories and, consequently, have gained in-depth knowledge of the development hurdles faced by these countries; and (ii) have based their mandates on diagnosing development/economic problems faced by these specific groups of countries and prescribing interventions for them.

3.2. Developed Countries

Ultimately, international organisations are extensions (and, hence, serve the interests and values) of their most powerful member states (Abbott & Snidal, 1998; Barnett & Finnemore, 1999; Drezner, 2009; Kochler, 2006). On the one hand, developed countries have an important role to play in curbing the tendency of international bureaucracies to further complicate the country differentiation landscape. After all, more differentiation means more committees, more meetings, more travel, more time and thus more costs that often have to be covered by developed countries. Indeed, the checks against bureaucratic waste mainly have to come from the major principals (the dominant countries or the major contributors). Clear and unambiguous mandates (that is, the delegation by the principals) are key. Gutner (2012, pp. 350–352) points out that the delegation of conflicting tasks (so-called antinomic delegation) will stimulate mission creep, that is: the growth of tasks, goals and often the mandate beyond levels that were originally envisaged. Blurring development and other goals (environment, peace-keeping, and so forth) may thus stimulate bureaucratic proliferation processes, both regarding the proliferation of categories and the instability of classifications within categories.

On the other hand, however, developed countries may also stimulate proliferation. According to Drezner (2009, p. 68) there are ‘powerful reasons to believe that regime complexity will enhance rather than limit the great powers’. Indeed, the impact of geopolitical interests on developmental relationships is a well-established stylised fact at all levels of international development cooperation for the UN, the World Bank, the Asian Development Bank and in bilateral development assistance (see for example, Kilby, 2006; Boschini & Olofsgård, 2007; Dreher, Sturm, & Vreeland, 2009). There is no a priori reason why this would not be the case for the proliferation of categories. Actually, it is likely that IOs sponsor reforms and policies aimed, ultimately, at replicating values supported by the powerful onto less powerful societies (Tabb, 2004).

3.3. Developing Countries

Developing countries may oppose new categorisations if they do not fit the criteria of the new category so that the differential treatment is not available for them. If donors reallocate ODA (official development assistance), this will put those countries at a disadvantage. Indeed opposition has been observed, for example, regarding the creation of the LDC category (Fialho, 2012, 2015). A motive for developing countries to push for new categories would seem to consist of the efficiency and efficacy of lobbying due to greater ease of coordination with like-minded countries in order to better advance interests and/or claim benefits. A category creates an international setting where countries can learn about each other’s shared interests and shared concerns. The existence of specific instances could then help to directly claim special and differential treatment. However, as noted by Miranda (2015), the costs of participation in international discussions may become too much for small and remote developing economies, possibly providing resistance to further proliferation from these countries as well.

The growing economic strength of developing countries since the 1990s adds other motivations. The sea change in their economic conditions over the past decade may provide a check to geopolitics and bureaucratic pathologies (see Humphrey & Michaelowa, 2013 for the case of multilateral lending). Moreover, developing countries have become more self-confident and want to play a role in the system (Elsig, 2011).²⁶ It is completely rational that politicians and delegates from smaller countries may prefer to participate in (committees of) international organisations where they can exert influence rather than in those that are dominated by the world powers (Frey, 2008, pp. 340–341). Of course it is also possible that developing countries might seek this proliferation of categories, lured by the possibility of having specific instances where to more directly claim preferential treatment.

Additionally, other important issues worth considering when analysing the external proliferation of developing countries’ categories are:

1. the inevitable tension between any developing country classification with stable criteria and the continuously changing international reality, which could make these classifications potentially less useful and representative overtime (for example, Tezanos Vázquez and Sumner [2015] demonstrate the weakness of existing classifications given an evolving developing world); and

2. the potential disconnect between the (static) analytical foundations of these country groupings and those used in the processes of resource allocation, including development aid.

4. Comparative Case Study: Landlocked Developing Countries (LLDC) Versus Small Island Developing States (SIDS)

Our case study focuses on two classifications that would a priori appear to be based on purely geographical grounds and logically therefore exclude overlap, but still show different internal and external dynamics in terms of proliferation.

The categorisation of LLDCs has been straightforward and unambiguous from the start: countries that ‘lack territorial access to the sea’ (UN, 2011, p. 1). A clear consensus exists on its underlying economic rationale that high transportation and transit costs related to being landlocked hinder trade and investment and, thus, impose economic burdens on these countries. The problems were recognised as early as 1957 in a UN General Assembly resolution calling for the ‘full recognition to the needs of land-locked Member States in the matter of transit and trade and … to accord them adequate facilities in terms of international law and practice’ (UN, 1957, p. 13). In fact, among developing countries, LLDCs present some of the lowest growth rates and ‘are heavily dependent on a very limited number of commodities for their exports’ (G77, 2004). To assist in overcoming these handicaps, the Almaty Programme of Action, established in 2003, represents the response of the international community to address the special needs of LLDCs, suggesting/recommending special and differential treatment in their favour. To our knowledge, the only glimpse of a potential dispute with regard to the definition of the LLDC list is a reference found at UNCTAD’s website, mentioning that the list is ‘informally accepted by UN member States’²⁷ (emphasis added). More importantly, the list has been stable from the start.²⁸

In contrast, ‘there is no clear definition for what constitutes an island-nation’ (Schmidt, 2005, p. A 607). This has generated much confusion. According to Carolina (2015, p. 4):

The existence of the ‘Small Island Development States’ (the SIDS) was recognized in 1992 by the United Nations (the U.N.) and this group was defined as ‘low-lying coastal countries that share similar sustainable development challenges, including population, limited resources, susceptibility to natural disasters, vulnerability to external shocks, and extensive dependence on international trade.’ There exists an inconsistency between the definition of the SIDS and its acronym. As a consequence, non-islands economies as Belize, Suriname and Guyana, are awkwardly classified under the SIDS.

In fact, there is report of the issue of islandness (and the development-hampering conditions it entails) being discussed in international fora as early as 40 years ago:

The third session of UNCTAD, in 1972, decided that a panel of experts should identify and study the problems of island developing countries. UNCTAD IV, in 1976, encouraged the international community to envisage special measures in favour of these countries. In 1977, the UNCTAD secretariat established a Special Programme for Least Developed Countries, and Land-locked and Island Developing Countries, the first such institutional unit within the United Nations … The main characteristics and problems of island developing countries were discussed in UNCTAD reports and raised in United Nations General Assembly resolutions, at regular intervals, between the late 1970s and the mid-1990s. … The notion of ‘island developing countries’ was abandoned by the United Nations in 1994, and gave way to a more focused denomination, that of small island developing States (SIDS) (Hein, 2004, pp. 4–5)

So, according to Hein (2004, p. 8), ‘it was only in 1994 that it became politically possible to exclude larger States from the range of island developing countries that were deemed in need of special attention’. However, to this day, this has still not been achieved. Actually, big islands (and non-islands) are still part of many different lists of SIDS.

The small islands’ categorisation is an initiative strongly backed by developing countries themselves. The Alliance of Small Island States (AOSIS) – an ad hoc lobbying and negotiating group²⁹ that represents the interests of SIDS within the UN system – has a membership of 44 countries and territories (including non-self-governing islands). It is important to note that ‘AOSIS members include Belize, Guinea-Bissau, Guyana, and Suriname, which are all coastal – although not technically island – nations’ (Schmidt, 2005, p. A 607). In addition, Cuba, with a population of 11.3 million, is also a member of AOSIS (Schmidt, 2005, p. A 607). While, contrary to the examples of Belize, Guinea-Bissau, Guyana and Suriname, Cuba is undoubtedly an island, with a population of more than 11 million people it can hardly be considered a small one.

In addition to the AOSIS list of SIDS, which Encontre (2004) considers to be an essentially political list; there is also reference to an economic list of SIDS, an institutional list of SIDS, and the UNCTAD non-official list of SIDS.³⁰ The World Bank’s list of SIDS is the least inclusive, while UNCTAD’s ‘unofficial’ list of SIDS excludes non-independent territories as well as bigger islands and coastal/continental countries considered by the UN-OHRLLS, UNDESA, UNESCO and AOSIS. Table 4 reports six different lists of SIDS; ranging from 13 to 52 such countries. The more inclusive of these SIDS lists is the one advocated by the UN-Office of the High Representative for LDCs, LLDCs and SIDS (UN-OHRLLS), with 52 countries, including indisputably not-so-small islands and non-independent territories. Importantly, the six lists include four different lists of SIDS being considered within one single organisation: the UN.

Table 4. Six different lists of SIDS

	World Bank’s small island exception^a	UNCTAD’s unofficial list of SIDS	UNDESA’s (SIDSnet) list of SIDS	AOSIS’ members and observers	UNESCO’s list of SIDS	UN-OHRLLS’ list of SIDS
Since:	1985	1994 (?)	1997	1991	2008	2001
Number of countries in 2013	13	29	39	44	45	52

Sources: compilation based on Encontre (2004), UNDESA (http://www.sidsnet.org/country-profiles), AOSIS (http://aosis.info/members-and-observers/), UN Office of the High Representative for the LDC, LLDC and SIDS (UN-OHRLLS) (http://www.unohrlls.org/UserFiles/File/UN_SIDS_booklet_5x6-5_062811_web.pdf), World Bank (http://www-Nwds.worldbank.org/external/default/WDSContentServer/WDSP/IB/2012/10/24/000333038_20121024232535/Rendered/PDF/733630BR0IDA0R0Official0Use0Only090.pdf), UNESCO (http://www.unesco.org/new/en/natural-sciences/priority-areas/sids/about-unesco-and-sids/sids-list/) (accessed 5 May 2013.)

Notes: ^a This list is periodically reviewed. The list on this table reflects the one considered in the October 2012 IDA16 Mid-term Review

Paradoxically, the inability to reach consensus (or informal agreement) on a single internationally (or even UN) agreed list of SIDS has not stopped the multiplication of SIDS-specific bureaucratic structures within the UN. These are aimed, essentially, at managing these lists. In effect, besides the Special Programme for LDCs, LLDCs and Island Developing Countries established at UNCTAD in 1977, AOSIS was established in 1991 and the UNESCO SIDS Platform was established in 2008,

A High Representative for the Least Developed Countries, Land-locked Developing Countries and Small Island Developing States was appointed in 2001, and the Department of Economic and Social Affairs has maintained a Small Island Developing States Unit. (Hein, 2004, pp. 8–9)

The UN has not been able to agree on SIDS-specific special and differential treatment to be unanimously advocated among its main funds and agencies, let alone among its member states. Indeed, not much has been done within the UN to convert ‘the recognition of SIDS-specific issues into […] SIDS-specific concessions’ (Encontre, 2004, p. 92). SIDS have been supported outside the UN essentially through ‘North-South arrangements such as those maintained by the European Union to benefit ACP countries, or by the United States in favour of specific regions involving island States (for example, through the Caribbean Basin Initiative)’ (Encontre, 2004, p. 92), or through the World Bank’s small island exception.³¹

Table 5 summarises the main differences between the LLDC and the SIDS categories. In contrast with the LLDC non-proliferation case, the debate on the required characteristics of SIDS has continued for many years, with no consensual list of SIDS defined. SIDS-specific bureaucratic structures were created within the UN, while special and differential treatment in favour of SIDS only comes from other organisations (for example, World Bank and EU). Looking specifically at the size of the six different lists of SIDS considered in this paper, one important explanation for the considerably smaller World Bank category is that the five different SIDS lists at the UN have been created in a situation of conflicting tasks related to (i) development (UNCTAD, UNDESA, UNOHRLLS), (ii) environment (AOSIS), and (iii) education/science/culture (UNESCO), possibly giving rise to mission creep; while the World Bank small islands exemption is straight forwardly related to only one aspect, that is: development finance. This clear one to one link with development finance is also reflected in the amount of bilateral aid disbursement as illustrated by the last line of Table 5. A classification may thus provide a misleading assurance of easier access to special and differential treatment, inducing costly lobbying activities.

Table 5. LLDC versus SIDS

	Landlocked developing countries	Small island developing states
First discussed	1957	1972
Acceptance	Informal	No
Consensus	Yes	No
Initiative	UN	UN and developing countries
Number of lists	1	at least 6
Number of countries	32	ranging from 13 to 52
Number of countries not meeting the technical requirements	0	19 non-independent territories or coastal/non-island countries included in the six different SIDS lists considered (this number goes up if size is considered as well).
Main policy documents	Almaty Programme of Action (2003) Almaty Declaration (2003) Roadmap for the implementation of the Almaty Programme of Action (2004)	Barbados Programme of Action/Programme of Action for the Sustainable Development of SIDS (1994) Mauritius Strategy for the further implementation of the Programme of Action for the Sustainable Development of SIDS (2005) Mauritius Declaration (2005)
UN apparatus/Organisational structures	UN Office of the High Representative for the LDC, LLDC and SIDS UNCTAD’s Special Programme for LDCs, LLDCs and Island Developing Countries	UN Office of the High Representative for the LDC, LLDC and SIDS UNCTAD’s Special Programme for LDCs, LLDCs and Island Developing Countries UNDESA’s SIDS Unit UNESCO’s SIDS Platform/Section for Small Islands and Indigenous Knowledge AOSIS (no budget, nor secretariat; operates out of chairman’s Mission to the UN)
Benefits/Special treatment/Support measures	EU-ACP Agreement: previews special treatment for ACP countries that are LLDCs Debt relief under HIPC initiative: previewed for LLDCs that are also HIPC	World Bank’s small island exception: IDA resources to small islands whose per capita income is above the IDA eligibility cut-off but have no (or very limited) creditworthiness, which impedes access to IBRD borrowing EU-ACP Agreement: previews special treatment for ACP countries that are also SIDS US Caribbean Basin Initiative provides duty-free access to the US market for most goods from 17 SIDS^a in the region
2000–2014 average annual total bilateral aid disbursement	$ 22,146 million (constant 2013 prices)	US$4,730 million (constant 2013 prices)
Idem, share in world total %	12.8	2.7

Sources: compilation based on UN (1957), Hein (2004), www.aosis.org, http://www.unesco.org/new/en/natural-sciences/priority-areas/sids/about-unesco-and-sids/unesco-sids-platform/, http://www.ustr.gov/trade-topics/trade-development/preference-programs/caribbean-basin-initiative-cbi, stats.oecd.org

Notes: ^aAntigua and Barbuda, Aruba, The Bahamas, Barbados, Belize, British Virgin Islands, Dominica, Grenada, Guyana, Haiti, Jamaica, Montserrat, Panama, St. Kitts and Nevis, St. Lucia, St. Vincent and the Grenadines, Trinidad and Tobago.

Our comparative case study thus confirms a number of the drivers of proliferation. Internal proliferation (increase in the number of countries in a category that is illustrated by the increasing number of SIDS) appears to be associated with a lack of a clear consensus on the economic underpinnings (the rationale) for the category, no clear definition of the country characteristics that entitle countries to differential treatment, and possibly the direct involvement of developing countries in drawing up the category. External proliferation (increases in the number of lists as illustrated by the increase in the number of SIDs list) may reflect antinomic delegation, geopolitical as well as bureaucratic motives. Unless a clear definition is reached, lists proliferate, as demonstrated, and there continues to be reason for more international debate, more studies, more counter-studies, more meetings, more expert advice, more travel and, ultimately, more costs (often largely borne by developed countries³²). The proliferation of different SIDS lists within one single organisation can, in fact, be a (i) reflection of self-interested actors maximising the likelihood of particular departments’ (and their own) survival and (ii) a concrete indication that bureaucratic motives are driving the process. Additionally, fuelling an ever-more complex external proliferation might also be in the interest of developed countries not so interested in according more special and differential treatment to yet another category of developing countries considered to be dealing with yet another ‘special situation’. Hein (2004, p. 12) states that

Skepticism remains about the legitimacy of SIDS as a category requiring special attention, and there has been reluctance in providing these countries with concrete forms of special treatment, although this is generally not said openly in international fora dealing with these questions.

Hence, supporting a complex situation or not making a definite effort to clarify a confusing state of affairs can work as a tool to foster the interests of those in power positions, in detriment of smaller and weaker states. Hein’s (2004, p. 13) qualification of ‘a politely supportive, yet almost dismissive attitude’ would seem to be on the spot.

The different proliferation outcomes of the LLDC and SIDS categories can also stem from the objectivity/subjectivity of their respective classification criteria. While both the landlocked nature and the islandness of a country are objective geographical criteria (that is, countries are either landlocked or not), the smallness of a country is arguably subjective. In effect, trying to decide on the exact population and/or territorial size threshold for being considered a small or a non-small country could give rise to several understandings/interpretations. Again this points to the need to develop a consensus on criteria if proliferation is to be reduced.

5. Can Complexity Be Reduced?

The differentiation created within the UN among developing countries has taken root and has resulted in unprecedented segmentation of the developing world, considering the large number of categories that has emerged, some of which with matching purposes. This paper investigated the substantial overlap that results from the significant proliferation of categories of developing countries, as well as the coexistence of several lists of SIDS within the UN, resulting mainly from lack of consensus among different organisational structures. Therefore, there ought to be an optimal point of proliferation of developing countries categories, where negative implications can be curtailed or minimised.

Despite the positive aspects of developing country differentiation (that is, it reduces the political nature of international assistance allocation and repositions it in a rules-based sphere), there are contradictions between UN stated intentions and policy implementation. Significant and oftentimes unnecessary differentiation among developing countries within the UN happens because the organisation also responds to bureaucratic interests, irrespective of whether or not they concur with the interests of certain groups of member states. In the end, it is the organisation and its representatives (and not necessarily the organisation’s most vulnerable member states) who benefit the most from the proliferation of developing countries’ categories. Therefore, keeping intact or empowering even further the bureaucratic structures put in place to support and manage these categories becomes the main purpose. This is a clear subversion of the most important purpose of developing country differentiation, which is the provision of special and differential treatment to countries truly in need of it.

The question that imposes itself and that will help to move this debate further – from observation and diagnosis of the situation to providing some direction to policymakers is: How to make developing country differentiation less complex and, consequently, more efficient (or at least less imperfect)? Given the misalignment of interests discussed in the preceding analysis, changes in the current framework of developing country differentiation should be deep but gradual, so that they can be truly meaningful and long-lasting. Hence, one needs to be pragmatic and accept the fact that these changes will have to be promoted from within an imperfect system.

The experience with the LLDC category during almost six decades suggests that proliferation can be contained: provide a clear rationale, agree on the country characteristics before starting the category and make sure that the IO has one clear mandate with respect to the category. This is a hopeful finding because the complexification of allocation/provision of special and differential treatment to developing countries makes it more difficult and costly to reap the benefits of international development finance.

In addition, the following policy change proposals could be the first concrete steps in contributing to a less dysfunctional framework of developing country differentiation:

1. Discourage the proliferation of even more categories/classifications, to prevent further fragmentation of special and differential treatment;

2. Streamline/rationalise/reorganise existing categories of developing countries within the same organisation in order to reduce proliferation;

3. Insist on inter-organisational harmonisation/coordination efforts;

4. Work towards less institutional fragmentation of efforts and more intra- and inter-organisational coordination;

5. Strive for less bureaucracy-motivated representatives (and, hence, adopt a tighter control of their performance), in order to minimise principal/agent mismatches;

6. Introduce more accountability measures targeted at IOs and at their representatives.

These policy changes could represent fewer costs, more efficiency, less bureaucratisation, and more transparency in creating/managing categories of developing countries and in providing them with special and differential treatment, which would ultimately represent more rationality and transparency in global governance.

Acknowledgements

Preliminary versions of this paper were presented at the Development Dialogue Conference (The Hague, October 2013) and the 8th annual conference on the Political Economy of International Organizations (Berlin, February 2015). Comments by participants and by John Cameron Bruno Frey, Susan Newman, Bernhard Reinsberg and André Corsino Tolentino on earlier drafts of this article are gratefully acknowledged, as well as the grant support provided by Fundação para a Ciência e a Tecnologia (FCT) and by ISS Economics of Development and Emerging Markets (EDEM) research programme.

Disclosure statement

No potential conflict of interest was reported by the authors.

Notes

1. See for example also Nielsen (2013, p. 1103): ‘The World Bank does not explain why the threshold between developed and developing countries is a per capita income level of US$6,000 in 1987-prices and the UNDP does not provide any rationale for why the ratio of developed and developing countries is one to three. As for the IMF’s classification system, it is not clear what threshold is used.’

2. While bilateral aid continues to be mostly politically motivated (Warmerdam & de Haan, 2011, pp. 11–14), multilateral aid is known to be less explicitly so (Rajan & Subramanian, 2008).

3. Other authors have challenged the need for the proliferation of categories on other grounds. Payne (2005, p. 40), for example, advocates for a new critical political economy of development that ‘rejects the “exceptionalism” of a special category of countries deemed to be in particular need of development and endeavours [and] … recast[s] the whole question of development as a universal question, a “transnational problematic”’. In line with this view, Hettne (1995, p. 263) defines development as ‘societal problem solving … [implying that] a society develops as it succeeds in dealing with predicaments of a structural nature, many of them emerging from the global context’. So, the burden of development should be placed on all, and not on specific groups of countries, making it ‘a global and universal problem’ (Hettne, 1995, p. 266), rather than a problem for selected groups of countries.

4. Contrary to the other examples in this table, AOSIS is the only group whose initiative is from the countries themselves. It ‘functions primarily as an ad hoc lobby and negotiating voice for small island developing States (SIDS) within the United Nations system’ (http://aosis.org/about-aosis/).

5. We owe this point to one of the referees.

6. In the same vein a criteria-based reduction of the number of countries within a category is not problematic (compare Fialho, 2015; van Bergeijk, 2015).

7. UN, Resolution on Identification of the least developed among the developing countries, 1971, at http://www.unitar.org/resource/sites/unitar.org.resource/files/document-pdf/GA-2767-XXVI.pdf, accessed 1 September 2010, p. 52.

8. http://hdr.undp.org/en/statistics/

9. http://hdr.undp.org/en/statistics/indices/hdi/

10. http://data.worldbank.org/about/country-classifications/country-and-lending-groups

11. http://data.worldbank.org/about/country-classifications/a-short-history

12. http://data.worldbank.org/about/country-classifications/a-short-history

13. http://www.imf.org/external/np/exr/facts/hipc.htm

14. http://web.worldbank.org/WBSITE/EXTERNAL/TOPICS/EXTDEBTDEPT/0, contentMDK:20259564~pagePK:64166689~piPK:64166646~theSitePK:469043,00.html#04

15. http://www.imf.org/external/np/exr/facts/hipc.htm and http://web.worldbank.org/WBSITE/EXTERNAL/TOPICS/EXTDEBTDEPT/0, contentMDK:20260049~menuPK:64166739~pagePK:64166689~piPK:64166646~theSitePK:469043,00.html

16. http://siteresources.worldbank.org/EXTLICUS/Resources/511777-1269623894864/FCSHarmonisedListFY13.pdf

17. Two criteria defined LICUS (considered to be countries with weak policies, institutions and governance): per capita income within IDA threshold and performance of 3.0 or less on both the overall Country Policy and Institutional Assessment (CPIA) rating and the CPIA rating for Public Sector Management and Institutions. Depending on these criteria, a LICUS country was classified in one of three subgroups: severe, core, or marginal. Marginal LICUS scored on the edge of what was considered LICUS and were identified only for monitoring purposes (http://www.worldbank.org/ieg/licus/licus06_map.html).

18. The Country Policy and Institutional Assessment (CPIA) rates countries according to 16 criteria, grouped in four clusters: (i) economic management; (ii) structural policies; (iii) policies for social inclusion and equity; and (iv) public sector management and institutions (www.worldbank.org).

19. http://web.worldbank.org/WBSITE/EXTERNAL/PROJECTS/STRATEGIES/EXTLICUS/0, menuPK:511784~pagePK:64171540~piPK:64171528~theSitePK:511778,00.html

20. http://web.worldbank.org/WBSITE/EXTERNAL/PROJECTS/STRATEGIES/EXTLICUS/0, contentMDK:22310165~menuPK:6432437~pagePK:64171531~piPK:64171507~theSitePK:511778,00.html

21. http://www.oecd.org/dacfragilestates/

22. http://siteresources.worldbank.org/EXTLICUS/Resources/511777-1269623894864/FCSHarmonisedListFY13.pdf

23. http://web.worldbank.org/WBSITE/EXTERNAL/PROJECTS/STRATEGIES/EXTLICUS/0, contentMDK:22310165~menuPK:6432437~pagePK:64171531~piPK:64171507~theSitePK:511778,00.html

24. The acronyms stand for Countries In Transition, Conflict Affected and Fragile States, Least Developed Countries, Low Income countries, Low Income Countries Under Stress, Heavily Indebted Poor Countries, Small Island Developing States and Structurally Weak, Vulnerable and Small Economies

25. http://www.imf.org/external/np/exr/facts/hipc.htm

26. Elsig specifically studies the WTO, but his analysis is also relevant for other international organisations.

27. http://unctad.org/en/Pages/ALDC/Landlocked%20Developing%20Countries/List-of-land-locked-developing-countries.aspx

28. To our knowledge, the creation and consequent inclusion of South Sudan in this list in 2011 is the only noteworthy change.

29. Created in 1990 at the 2nd World Climate Conference.

30. Additionally, in the analytical categorisation used in the 2012 LDC Report, UNCTAD differentiates island LDCs from other LDCs, grouping together eight African, Asian and Pacific island LDCs, while leaving out Haiti and Madagascar because both are regarded as large islands (UNCTAD, 2012, p. xii). This exclusion is consistent with UNCTAD’s unofficial list of SIDS. However, strangely enough, in the report, Haiti, a Caribbean island, ends up grouped together with African LDCs, revealing a considerable degree of discretion and randomness in the framing of the analysis and thus in perceived and reported needs of island economies.

31. This exception, in effect since 1985, reflects the recognition, by the World Bank, that SIDS typically have to deal with higher transportation costs, fewer opportunities to pursue economies of scale and severe human capital constraints because of their small size and small populations. Thus, the small island economy exception permits the provision of IDA resources to small island economies, with per capita income above the operational cut-off for IDA eligibility.

32. See, however, (Miranda, 2015) who also points out that these costs may become relatively large, if not prohibitive, for small and remote developing economies.