Universal health coverage as a global public health goal : the work of the International Labour Organisation , c . 1925-2018

We examine the efforts of the International Labour Organisation (ILO) to extend medical care under social security, through international conventions, advocacy and technical assistance. We consider the challenges faced by the ILO in advancing global health coverage through its labourist, social security model. The narrative begins in the interwar period, with the early conventions on sickness insurance, then discusses the rightsbased universalistic vision expressed in the Philadelphia Declaration (1944). We characterize the ILO’s postwar research and technical assistance as “progressive gradualism” then show how from the late-1970s the ILO became increasingly marginalized, though it retained an advisory role within the now dominant “co-operative pluralistic” model.


Abstract
We examine the efforts of the International Labour Organisation (ILO) to extend medical care under social security, through international conventions, advocacy and technical assistance. We consider the challenges faced by the ILO in advancing global health coverage through its labourist, social security model. The narrative begins in the interwar period, with the early conventions on sickness insurance, then discusses the rightsbased universalistic vision expressed in the Philadelphia Declaration (1944). We characterize the ILO's postwar research and technical assistance as "progressive gradualism" then show how from the late-1970s the ILO became increasingly marginalized, though it retained an advisory role within the now dominant "co-operative pluralistic" model.
Keywords: universal health coverage; International Labour Organisation (ILO); labour; health insurance rights.
História, Ciências, Saúde -Manguinhos, Rio de Janeiro I n September 2019, universal health coverage (UHC) was high on the international policy agenda, when a high-level meeting of the United Nations (UN) gave priority to this aspect of its Sustainable Development Goals (SDGs) (UN, 23 Sept. 2019). Although today's conception of UHC is relatively recent, an international concern for extending access to health care has a longer history, taking different forms across time. The first global statement of intent had come a hundred years ago, with the founding in 1919 of the International Labour Organisation (ILO), whose Constitution called for "the protection of the worker against sickness, disease and injury arising out of his employment" (ILO, 2009, p.249). Amongst the international organisations of the twentieth century, the ILO is distinctive for its longevity and its formal "tripartite representation" of workers, employers and national governments. Its contributions to the promotion of universal rights to health are the subject of this article.
Born from the Treaty of Versailles, the ILO's creation was impelled by the conviction that peace must be founded on social justice. This rationale framed its early attempts to advance international standards of medical cover under social security. In the mid-century, articulation of health coverage as an international goal changed, now adopting a language of rights. The ILO's Philadelphia Declaration, conceived in 1944 just as the tide of the Second World War was turning, affirmed the rights which working people might legitimately claim when peace was won. The Declaration boldly asserted that "labour is not a commodity," and that "all human beings … have the right to pursue ... their material well-being … in conditions of freedom and dignity, of economic security and equal opportunity." One such right would be "the extension of social security measures to provide … comprehensive medical care" (ILO, 1944a, p.4-6).
This earlier articulation of UHC is somewhat removed from today's global imperative. For the World Health Organisation (WHO), the rhetoric of universalism is humanitarian, asking that "all 'people and communities' receive the health services they need without suffering financial hardship" (WHO, n.d.; emphasis added). The World Bank, the other leading international proponent, is more equivocal, advocating "that 'people' have access to the health care they need without suffering financial hardship," and adding a utilitarian appeal that "countries … make the most of their strongest asset: human capital" (Jamison et al., 2018;emphasis added). What is striking about this language is the near absence of a rights discourse, and its rather defensive and aspirational tone. It is also a far cry from the welfare states of the West, where entitlement to health security originally emerged from work and citizenship.
In the following text, we ask how this early conceptualisation, grounded in universal rights and the dignity of work, arose, and why it gave way to the framing of today, in which "need" trumps rights, and labour is once more a commodity: "human capital." The methodology is principally documentary research in the archives of the ILO, supplemented with a number of oral history interviews. The exposition will proceed chronologically, following a narrative arc of "rise, decline and fall." It should be stressed that the argument is not that the ILO was particularly important in furthering the spread of UHC, at least in the postwar period (Sirrs, 2019;Landy, 1970;Strang, Chang, 1993). Rather, we use this case to explore the reasons why the advance of access to health care as a supra-national goal has been so problematic over time. Before this, we introduce the ILO and describe its constitution and activities prior to the Philadelphia Declaration.

Preliminaries: establishment of the ILO
Launched in 1919, the ILO grew from the peace settlement that followed the First World War and three converging motivations. First was the impetus for social reform arising from the labour movement and from progressive liberalism and "solidarisme." Article 247 of the Treaty of Versailles enshrined these in a Labour Charter, which promised inter alia adequate wages, equal pay, and fair conditions. Legitimation of capitalism also mattered in the era of Bolshevik Revolution. Workers' conditions would be addressed consensually, through a tripartite representation of governments, employers (through leading business-people or confederate bodies) and labour (through trade unions and their national organisations) in ratios of 12:6:6 on the ILO's Governing Body and 2:1:1 at its annual International Labour Conference (ILC). Capital also sought a level playing field between nations, so that firms implementing progressive labour reforms were not competitively disadvantaged.
How would the ILO function? It was empowered to agree on international conventions, which member states would enact, and to promulgate non-binding "recommendations" setting out optimal standards. A standing administration, the International Labour Office, was established (and headquartered in Geneva); it issued a regular Bulletin, with ad hoc committees to address emergent issues. The first director-general, French trade unionist Albert Thomas, set the tone. Following the Preamble to the ILO's Constitution, he argued that only a programme of social justice would ensure lasting peace with the ILO in the expression of the worker's interest to employers and governments (Haas, 1965, p.144, 501). The Office would liaise with national trade union movements, and provide technical expertise on social security and improved working conditions (Haas, 1965, p.145-148). The ILO's Constitution (1919) specified that this should include "(p)rotection of the worker against sickness, disease and injury arising out of his employment" (ILO, 2009).
In what sense was the ILO a "global" agency? At its foundation, there were 42 member states, predominantly from Western and Eastern Europe (Ghebali, 1989, p.116). The first four directors-general were French, British, American and Irish. Latin American nations were also represented (reflecting industrialisation and unionisation in mining, transport and manufacturing), along with the Anglosphere of British dominions. As for the great powers, the United States acceded only in 1934; the Soviet Union joined in 1934 but was excluded in 1939; Germany left the ILO in 1935 following Hitler's accession, as did Austria after the Anschluss (1938), Italy (1939), Japan (1940) and Spain (1941. China and India were members, the former in political turmoil, the latter still under colonial rule. The imperial nations, mainly Britain and France, spoke for their colonies; the only African members were Liberia and Ethiopia (Ghebali, 1989, p.117-118). So despite its democratising features, the ILO was compromised by its dominant balance of power. For example, "progressive" colonial states were less amenable when a convention banning forced labour was under discussion (Maul, 2012, p.17-27).
Thus, when the right to health under social security was broached, it was in the context of the more advanced industrial economies. The dominant model was Bismarck's Germany, which in 1883-1889 implemented mandatory sickness, accident and disability insurance for workers (Hennock, 2007). The new European welfare states were a radical break, collectivising individual risk, redistributing responsibility across employers and employees, and using actuarial science to gauge liabilities and premiums (Ewald, 1986). By the interwar period, such coverage was partial in Western nations and absent in the United States. Municipal and philanthropic medical institutions completed the patchwork, leaving many at the mercy of the market. Elsewhere in the world, biomedical services were largely limited to colonial enclaves, or concerned with maintaining healthy workers for primary production.

1925-1942: the rise of transnational health security
In 1925, sickness insurance appeared on the ILC agenda, following the Organisation's initial focus on occupational health (Weindling, 1995;Cayet, Rosental, Thébaud-Sorger, 2009). Two conventions, one for industry and the other for agriculture, were adopted in 1927; the former committed members to extending compulsory contributory insurance to "manual and non-manual workers," with costs split between employers, employees and (subject to "national laws") the state. Benefits included cash payments and medical care, with administration by self-governing funds in which the insured "shall participate in the management" (ILO, 1927a). Office documents signalled universalist intent, with the suggested scope "(p)ractically all persons under a contract of service..." (ILO, 1925 p.808). The agriculture convention was virtually identical, although it omitted the same rights of maternity protection for women (ILO, 1927a(ILO, , art.8, 1927c. The purpose of distinguishing the two was to allow countries where rural labour enjoyed paternalistic protections, or payment in kind, to ratify a convention protecting only its industrial workforce (ILO, 1927b, p.288, 292-293). Both conventions were comfortably passed by the ILC, by votes of 72-4 and 72-0, respectively (ILO, 1927b, p.322-327).
An uneasy compromise between workers and employers lay behind these initiatives. The ILC's Social Security Committee was overwhelmingly Western: in 1925 only 6% were non-European, and in 1927, 24% (ILO, 1925, p.lxxxii-lxxxiii). These governments sought a "level playing field" through standardisation − "to reduce the obstacles resulting from competition" for nations adopting national insurance (ILO, 1925, p.4). For labour, the concern was "social justice" and "effective protection ... against risks," while for capital it was "a healthy and vigorous labour supply" to develop "productive capacity" (ILO, 1925, p.813). Workers and employers were divided over whether to emphasise compulsion or to permit state-subsidised voluntary insurance systems, such as in Switzerland and Denmark (ILO, 1927b, p.294, 298, 635-636). Compulsion was favoured by workers and nations like Britain, which saw the decline of voluntary arrangements as a historical inevitability (ILO, 1927b, p.293, 298, 300-302, 587). Labour also called unsuccessfully for insured workers to have majority representation on fund boards (ILO, 1927b, p.292, 314-315). Non-European voices were marginal. Brazil supported partial exemptions for countries with large, thinly populated areas (ILO, 1927b, p.409, 414). The Indian representative (from the All-India Trades Union Congress) embarrassed colonial governments by pointing out the lack of social insurance for the "abject and miserable workers of India" (ILO, 1925, p.494).
Behind all this lay the guiding influence of the ILO's Social Security Section, led by Adrien Tixier, a disabled French war veteran (Anonymous, 1946). This was a nexus for the classic "epistemic community" which international organisations foster, that is a network of experts who renounce national interests in the spirit of transnational goals (Haas, 1965(Haas, , 1992Barona, 2019). Particularly important were German social security experts, whose technical authority derived from long experience with social insurance and the superior data of the Reich statistical office (Kott, 2008). Bismarckian influence also came from the Chair of the ILC Social Security Committee, Andreas Grieser, Ministerialdirektor of Germany's Ministry of Labour (Kott, 2018). Such officials argued sickness cover deserved priority as the "most fundamental form of insurance" on which to build social security systems (ILO, 1925, p.811-812). Thus began what would subsequently be considered the European social model of welfare (Kott, 2010).
The Conventions entered into force in 1928, the eve of the Depression, and take-up was very limited. Just 14 states ratified by 1939, including only four non-European countries (Chile, Uruguay, Colombia, Nicaragua), although routine national reporting allowed the Office to monitor developments going forward (ILO, 1936(ILO, -1937. The late 1920s saw the ILO's first research and advisory work on health system structure, after Czechoslovakia initiated a joint enquiry with the League of Nations Health Organisation (LNHO) on how best to integrate social health insurance with public health administration. Chaired by a cautious centrist, Britain's George Newman, this body advised in favour of the existing loose pluralism.
Economic depression in the 1930s propelled the ILO towards a more activist role, championing interventionist employment policies, and a turn from setting standards to leading a response to the crisis (Haas, 1965, p.149-153). The Social Security Section stepped up its technical advice, disseminating knowledge of social insurance and its organisation and advocating more boldly. A joint LNHO report on health effects of the slump asserted: "In the present economic and social conditions … compulsory sickness insurance must be regarded as the most appropriate and rational method of organising the protection of the working classes" (ILO, 1933, p.18). A changing balance of power at the ILO helps explain this. After Germany's withdrawal it opened to the Americas, particularly following the United States's accession and the domestic politics of President Roosevelt's New Deal which included improved labour rights and welfare protections (though not health insurance) (Rodgers, 1998;Jensen, 2013). Meanwhile, Labour in New Zealand introduced a Social Security Act (1938) establishing the first "NHS" (National Health System) in a liberal democracy, with a tax-funded, universal and comprehensive health service (Hanson, 1980).
With this political momentum, ILO officials became more proactive, particularly in Latin America, after Chile's President Alessandri, invited them to convene a regional conference in Santiago in 1935(Alcock, 1971. Chile was a pioneer of social security, having initiated Bismarckian workers' insurance in 1924, responding to labour mobilisation in the nitrate industry. In the 1930s, coverage expanded under an economic recovery programme also founded on debt relief and industrial stimulus, so the conference probably had both legitimation and "level playing field" impulses (Raczynski, 1994;Collier, Slater, 1996, p.226-232). Tixier and his deputy, Osvald Stein, a Czech insurance expert, subsequently led several delegations to Latin America, for example to advise Venezuela on drafting insurance legislation and a labour code (Anonymous, 1946(Anonymous, , 1944Alcock, 1971, p.135-136, 146-148). In 1942 this resulted in a continental social security code, accepted in the Declaration of Santiago de Chile (Cohen, Oct. 1942).
Writing about these extensions of social insurance, Tixier (1935, p.779) adopted a language of rights and universalism, explicitly criticising welfare pluralism: "It is generally recognised nowadays that individual saving, public assistance, and voluntary insurance are inadequate, and that compulsory social insurance against the various occupational and social risks is the most scientific and the most effective means of providing the employed population as a whole with the protection to which it is entitled." As the international crisis deepened at that time, the ILO championed a workers' welfare state using a discourse of science, rationality and modernity.

1942-1952: the universalist moment
In 1944 the Philadelphia Declaration included a clear articulation of universal health coverage as part of labour's right to social security. It called for: "(f) the extension of social security measures to provide a basic income to all in need of such protection and comprehensive medical care; (g) adequate protection for the life and health of workers in all occupations; (h) provision for child welfare and maternity protection" (ILO, 1944a. p.5-6).
Accompanying this was a Recommendation advising member states on how this could be achieved. Here universalism without discrimination was made explicit: 8. The medical care service should cover all members of the community, whether or not they are gainfully occupied.
... 11. Where medical care is provided through a social insurance ... service, all members of the community should have the right to care as insured persons ... 18. Where medical care is provided through a public … service, the provision of care should not depend on any qualifying conditions … or … a means test, and all beneficiaries should have an equal right to the care provided (ILO, 1944b).
The text had been agreed at the Philadelphia ILC, following the ILO's relocation in 1940 from Geneva to Montreal for the duration of the war (Alcock, 1971, p.159-165). The document was signed by President Roosevelt, ILC President Walter Nash (a New Zealand Labour politician), and ILO Acting Director Edward Phelan, in the presence of the American Secretary of Labor Frances Perkins.
What explains the ILO's adoption of this more radical position? First, it reflected the idealistic war aims of the Allies, whose Atlantic Charter (14 Aug. 1941) expressed "common principles ... for a better future." This had promised, alongside peace and freedom, "improved labour standards, economic advancement and social security" (Atlantic Charter, 14 Aug. 1941, preamble, para.5). War also made the imperial powers at the ILO more concerned with ensuring the loyalty of their subjects, thus countenancing positions that potentially compromised their authority (Maul, 2012, p.51-58). Second, the transnational policy discourse about reform accelerated after publication of the British Beveridge Report (to which ILO staff had contributed evidence), with its blueprint for a universalist welfare state (Inter-departmental Committee on Social Insurance and Allied Services, 1942). Similar documents came from Canada and the United States (Marsh, 1975;National Resources Planning Board, 1942).
Third, the Office actively advanced this expansive vision of welfare. After Stein's early death, the Anglo-German economist Laura Bodmer emerged as the section's leading technical expert and hub of its network of advisers; after briefly relocating to wartime Britain, she returned in 1942 to work on the Recommendation (ILO, n.d.). In July 1943, the Office convened a Consultation of Social Security Experts in Montreal, bringing together William Beveridge with Canada's Leonard Marsh, America's Isidore Falk and George Perrot, and representatives from Chile, Cuba, Ecuador and Peru (Jensen, 2013). These were the expert networks from which the Philadelphia proposals for comprehensive, universal "medical security" through either social health insurance or a "public general service" would emerge.
With the peace, the ILO focused on transforming the Declaration into a Convention, "Objectives and Minimum Standards of Social Security," with sickness insurance and medical care included as one of nine branches of social security. It also sought to embed these objectives with the new UN special agency for health, the WHO. Neither of these efforts would succeed.
The Convention was progressively diluted, and the original intent ultimately effaced, through debates amongst delegates to the ILCs between 1949 and 1952 and behind-thescenes lobbying as the Office prepared draft texts for the conference. The first retreat occurred in 1951, with an agreement that ratification could be at either "minimum" or "advanced" standards. This addressed the concern that many members lacked the resources to establish social security systems, and allowed states to restrict the population and areas covered and the range of medical care available (albeit including "essential" drugs) (ILO, 1950, p.24-25, 36-37). The 1952 ILC debate undermined the key principles even further. First, the advanced standards were dropped entirely. In the sections on medical security, the redistributive and integrative elements were removed, with voluntary insurance accepted and cross subsidy of funds rejected. Ratification could now take place with a commitment to only three of the nine branches of social security, potentially excluding medical provision altogether. Even then, countries could claim an opt-out as a "temporary exception." The final text effectively abandoned full medical security and social welfare as a universal right.
Meanwhile, the WHO rejected a strong commitment to health services in debates over its remit. This reflected a division between what came to be called "vertical" and "horizontal" approaches to global health (González, 1965, p.9, 11-12). For some states, faith in biotechnologies, coupled with broader geopolitical calculations, gave priority to centrally planned programmes of disease eradication. For others steeped in the earlier ideas of the LNHO, what mattered was building health infrastructure as an aspect of economic development. In 1951 this legacy of European social medicine was prominent in a joint ILO/ WHO Consultant Group chaired by Rene Sand and containing Henry Sigerist (Gillespie, 2002;Gorsky, Sirrs, 2018). Its report urged the WHO to advocate universal coverage, without means-testing and with services free (or nearly free) at the point of use. It also favoured a salaried medical workforce and a unified national administration to ensure the integration of primary and secondary care (WHO, 1952). These findings were aired in the WHO's Bulletin, but made no further impression, and were quickly dropped.
What explains this reversal of the aspirations of 1944? The Office discussions preceding the ILCs show much hostile lobbying from employers who attacked the proposals as a "monstrosity" which would create "a bureaucratic body of officials with all-embracing tentacles" (ILO, 1952, p.398-401). The doctors also mobilised through the World Medical Association, an organisation largely funded by big pharma, and led by American domestic opponents of social health insurance (ILO,1949a(ILO, , p.114-118, 1949b. At the ILC there was a split between capital and labour, with employers arguing there should only be a Recommendation (ILO, 1953, p.306, 314-315, 325). When this was defeated, they shifted to weakening the Convention, with cover for direct employees only, qualifying periods for eligibility, voluntarism rather than compulsion, a lower range and rate of benefits, and no coverage for migrant labour (ILO, 1953, p.306-309, 325). Worker representatives took opposing views, reiterating the foundational idea that social justice was essential to world peace (ILO, 1953, p.308, 313). Communist members from Poland and Czechoslovakia argued angrily against anything which "puts a brake on legitimate efforts being made by the exploited masses" (ILO, 1953, p.307, 314).
Unlike in 1927 however, the changing geopolitical balance of power in the ILO provided the employers with new allies. The newly independent states of India and Pakistan, along with Brazil, sought on behalf of the "underdeveloped countries" to ensure a "flexible" convention, which allowed members to ratify but claim temporary exceptions without a time limit (ILO, 1953, p.326-327, 330-332). Indian representative Krishna Menon, a London-educated socialist with "little sympathy for the West," was instrumental in shaping the non-aligned position at the UN (Choube, 1964, p.103;Highlights..., 2007). At the ILO, the agenda was to overturn the colonial powers' dominance, decentralise power, and shift the focus from standard-setting to development assistance (Maul, 2012 p.111-118).
Also crucial was the position of the United States, the largest financier (at 37%) of the UN budget, as a dominant player amongst multilateral organisations ( UN Secretariat, 4 Sept. 1952, p.5). By 1952, the New Deal advance of labour's rights to organise, strike and picket had been halted by the Taft-Hartley Act (1947) (Dawley, 1989, p.172). The McCarthyist Red Scare was in full swing, to the extent that American employees of UN agencies were expected to sign a loyalty oath (Farley, 2008, p.186-188). President Truman's failure to deliver social health insurance meant that for US trade unions, medical cover by private or mutual funds was now integral to collective bargaining (Klein, 2003). Hence at the ILO, employer representative Leonard Calhoun eulogised the "American Dream," in which "individual dignity, independence and … personal freedom" were jeopardised when the state provided "entire security for the family," and "socialised medicine" was anathema (ILO, 1953, p.309). Robert Myers, representing the government, similarly opined that "Man does things more effectively of his own volition … instead of doing them from compulsion" (ILO, 1953, p.325). Labour delegate Stanley Ruttenberg tore into the employers' reluctance to act, though he too rejected "socialised medicine" as un-American (p.321). This absence of leadership from the United States proved decisive. The weakened convention passed but proved ineffectual. In its first thirty years, there were just 29 ratifications, of which only 16 (including nine from Europe) accepted both sickness insurance and medical benefits (ILO, 2019).

1952-1975: progressive gradualism
After these setbacks, the ILO's policy towards health security became one of "progressive gradualism." This denotes limited advisory work in developing countries, pioneering comparative research into health systems performance, and in the late 1960s, another push to assert international standards through a convention.
Under David Morse, director-general 1948-1970, the Organisation initially abandoned its standard-setting endeavours. This was partly due to decolonisation, which substantially changed the ILO as the proportion of states from Africa rose from 4 in 1950 to 30 in 1962. Meanwhile, the percentage of democratic member states fell from 64% to 39% (Haas, 1965, p.170). The new nations asserted different concerns in a hitherto Western-dominated arena, switching emphasis from social welfare to economic growth. Morse therefore prioritised "technical assistance" by ILO officials to improve labour's skills and productive capacity. This new direction reflected the UN's doctrine of "development" initiated by the colonial powers, then promulgated by President Truman in 1949 and President Kennedy in the 1960s, whereby Western funding and know-how would usher poorer nations towards modernisation. The guiding assumption now was that improving output was the precondition of welfare spending.
Nonetheless, from the 1960s nascent social security systems began to figure in economic plans of emerging post-colonial nations. This gave opportunities for ILO advisers and consolidated Office thinking about health system structures. Experience in low-income settings also laid bare the challenges involved in extending UHC beyond the advanced industrial economies. For example, the countries taking ILO advice included the Francophone states of Senegal, Gabon and Mali, where it worked in 1961, 1962, and 1966 respectively. In practical terms, its activities included: surveys of business capacity, to establish the potential extent of payroll-based social security; actuarial analysis, to balance deductions and benefits in light of local resources; and advising over regulations, entitlements and staffing. Western bureaucratic models were thus transposed onto newlyminted African states in the pro forma paperwork of claims, record-keeping, and carnets de santé which prescribed norms of health behaviour.
These cases also brought home the difficulty involved. The ILO's post-colonial mantle had been compromised from the outset. For example, prior to independence it had lent support to France for its Code de Travail (1952), which granted workers in its overseas possessions the same rights as those at home. Yet this rhetoric of "progressive colonialism" had carried no financial commitment to extend comparable social security to the indigenous workforce, and the ILO, by providing internationalist sanction, was partially complicit (Cooper, 1996, p.363). The same tension continued after independence. In the absence of direct bilateral aid, extending welfare under "development" was premised on a rapid dash towards self-sustaining growth, either through industrialisation or export-oriented agriculture. Yet growth remained modest for most African nations, while many suffered political instability as independence settlements fractured. Without the necessary resources, plans to extend medical care under social security could not proceed far.
Nonetheless, the ILO observed and enumerated the impediments. First was the gulf between demand for medical services and available resources, as local training in biomedicine remained rudimentary and a "brain drain" to the West began. Second, the salariat that might sustainably fund social security was typically small and geographically concentrated. To begin and then scale up, UHC on this basis would therefore be unrealistic. Third, the alternative providers, the public health services initiated under colonialism, were already inadequate and thinly resourced, and might be undermined further if state spending supported social security. Fourth, there was the question of whether the continental social insurance model preferred in the Francophone nations was indeed optimal. The ILO's position was to support its member state's particular requirements, though privately officials might advocate the desirability of an NHS model on grounds of cost and administrative efficiency.
Alongside these modest technical assistance efforts, more conceptual work was undertaken in Geneva. Laura Bodmer's late career contribution was to initiate comparative health systems statistics, addressing questions about the cost-effectiveness of different models. The original spur came from Chilean WHO consultant Hernán Romero Cordero, which identified the dearth of hard evidence (Romero Cordero, 8 Aug. 1956). This led to a joint ILO/WHO working party in 1958, which brought together Bodmer and two British Fabian socialists, Richard Titmuss and Brian Abel-Smith, who had developed performance metrics for the NHS. Also involved was the American Milton Roemer, a social medicine advocate and New Deal veteran who had worked for WHO until he was forced to resign in the McCarthyite loyalty scare (Abel, Fee, Brown, 2008). He then moved to the Canadian province of Saskatchewan, advising on the establishment of UHC by Tommy Douglas's socialist administration.
This effort laid the groundwork for two WHO technical reports led by Abel-Smith, which are generally regarded as milestones in harmonising cross-national comparative statistics. However, the earlier (and unsung) fruit of the initiative was an anonymous ILO publication presenting quantitative data on fourteen countries from 1945 to 1955. Prepared and written by Bodmer, this responded to the "apprehension" about rising costs of medical protection under social security (ILO, 1959, p.1). Though principally empirical, it demonstrated the sustainability of universalist systems such as those in the UK and New Zealand, showing that care provided under social security "does not appear … to have been more expensive … than care privately obtained, or provided at the expense of public funds, in the United States" (p.2, 156).
The problem of evaluating different models also informed a commissioned text by Milton Roemer which provided an early typology of health systems according to criteria such as funding mechanism and ownership (Roemer, 1969). Although the ultimate goal of relating different models to health outcomes remained elusive, Roemer favoured the NHS approaches of Britain and Chile as "administratively easier to combine organised efforts under a co-ordinated or unified authority." This stance resonated with the interwar social medicine tradition, which regarded statist systems founded on health as a right of citizenship as an evolutionary trend. A later ILO study by a British consultant, Derick Fulcher, took a more centrist line (Fulcher, 1974). Comparing various NHS and social health insurance systems showed that neither model was "better," and that policy actors should work with the grain of historical determinants rather than advocate one optimal model. Both types were, however, superior to market forces.
In the late 1960s, the ILO's "progressive gradualism" gave cause for optimism, and the Office began planning another convention. By now, most Western nations had achieved UHC by expanding social insurance entitlements, while the communist bloc, including China, had instituted full coverage under public systems. Even the United States began edging towards universalism, with Medicare and the means-tested Medicaid providing for older and poorer citizens.
Also, the ILO was responding to international currents. In 1968, the UN's International Year of Human Rights, Morse argued that "civil and political freedoms" must also include the "economic, social and cultural rights" endorsed by the Philadelphia Declaration. These were "not brought about automatically by development or economic growth" (ILO, 1968, p.3, 7). Instead, updated conventions on social security were needed, "adapting them to new conceptions" and "specifying more accurately the scope and level of rights they guarantee" (p.88). Morse's dissatisfaction with the meagre welfare gains of "development" led him to prioritise the social needs of labour, though his World Employment Programme (1969) of job creation for the poor.
The Medical Care and Sickness Benefits Convention therefore arose from a progressive wave of policy-making concerned with social rights (ILO, 1969a). It went beyond the 1952 convention with respect to the extension of population coverage; the range, levels and duration of benefits provided; the acceptance of either publicly-administered (i.e. NHS) or social insurance systems; and limits to qualifying periods for entitlements. However, delegates at the ILC rehearsed familiar conflicts, auguring poorly for the normative influence of the convention. The employers inveighed against broad coverage, over-generous wage replacement rates and migrant rights, and advocated lower standards as the "flexibility" necessary to encourage ratification (ILO, 1969b(ILO, , p.442-443, 1970. Workers resented co-payments and qualifying periods and condemned "flexibility" as a smokescreen for dilution (ILO, 1970, p.440-441, 443). African trade unionists protested the treatment of "developing countries like little children who are told 'you cannot do that yet'" (ILO, 1969b, p.445). Public systems were advocated by Cuba and the USSR, which eulogised Lenin's decree on the right to free health care (ILO, 1970, p.443-444). American business delivered an ideological peroration, by Lyle Fisher of Sellotape firm 3M, for whom a "government-controlled" system "appeals to the weaknesses in man − self-pity and fear. ... it encourages dependence on others. This is contrary to sound, progressive principles of self-respect" (p.442).
História, Ciências, Saúde -Manguinhos, Rio de Janeiro The 1969 convention came at a "high point" for the ILO (Standing, 2008, p.359). In December, it was awarded the Nobel Prize for its efforts in furthering peace through social justice. The World Bank's new President, Robert McNamara, embraced Morse's "basic needs" concept and steered its lending strategy towards poverty reduction as well as productivity (Maul, 2012, p.250-254). The Swedish economist Gunnar Myrdal argued for a "welfare world," in which Western nations would abandon the tariff policies that sustained underdevelopment in solidarity with the Global South (Myrdal, 1970, p.298). Post-colonial nations amplified this call for economic justice in 1974, when the UN declared a New International Economic Order, with policies explicitly favourable to developing countries (UN, 1974). However, this "high point" for the ILO's ideals would be short-lived.

1975-1990: universalism in retreat
From the mid-1970s the ILO found itself on the defensive. The oil shocks of 1973-1974 heralded the end of the West's postwar boom, and curbed expansiveness in social welfare. In the East, the faltering command economies and impending demise of Soviet communism strained medical systems, while China's economic liberalisation encouraged market incursions and eviscerated rural health insurance. By the mid-1980s, a "neo-liberal" policy discourse about the welfare state's moral and fiscal crisis emerged, challenging the ILO's advocacy of social security by right. Human rights discourse in international politics also shifted, jettisoning concern with equality and social security and foregrounding civil and political rights (Moyn, 2010(Moyn, , 2018. The power and legitimacy of labour also waned. In Latin America, for example, the debt crisis from 1982 depressed real wages and drove unemployment, while states adopted restrictive or (as in Pinochet's Chile) repressive policies (Roxborough, 1995, p.368-373). In the West, Margaret Thatcher's breaking of the miners' unions, and Ronald Reagan's defeat of the air traffic controllers marked decisive ruptures in the postwar settlement.
As for global health policy, the WHO/UNICEF Alma Ata Declaration (1978) had initially seemed promising. This championed universal primary health care to achieve "Health For All" by 2000, and WHO had pledged to collaborate over the "planning and organization" of services (ILO, 1970, p.641). However, momentum soon lapsed. For WHO it was outcomes that mattered, not health financing strategies, so pluralist medical economies were acceptable provided primary health care was enhanced. The official WHO/UNICEF position was that: the classical social security systems applied in some of the industrial countries may, in developing countries, tend to favour very limited population groups and thus lead to discrimination against the majority … Every country has to evolve its own methods, based on its own circumstances and judgments ... National non-governmental organizations should be encouraged ... External financing may take the forms of loans and grants from bilateral and multilateral sources (WHO, UNICEF, 1978, p.42-43). This was a misplaced hope, for the debt crisis afflicting many African and Latin American nations from 1982-1983 drained local resources and made foreign donors cautious. The Alma Ata vision dissipated: training and organising the requisite personnel was a barrier; existing systems remained heavily oriented towards urban hospital medicine; China's inspirational "barefoot doctor" model was difficult to emulate; and efforts by outside NGOs to foster local community engagement proved misplaced (Packard, 2016, p.249-266). Policy makers again turned to more limited vertical programmes comprising key interventions of proven cost-effectiveness. This new direction was sealed in 1982 when UNICEF launched its GOBI package featuring growth monitoring, oral rehydration therapy, breast-feeding and immunisation (Grant, 1982).
Thus, the ILO's health security goals were marginalised and assaulted. The 1969 convention remained the ILO's blueprint for technical assistance, but had little influence, with only 12 ratifications to 1982: five from Latin America, one from Africa (Libya), and six from Europe (principally Scandinavia). In a unipolar world, American development policy was increasingly channelled through the World Bank, in preference to the UN agencies. The emergent "Washington consensus" on deregulating global markets promulgated by the Bank and the International Monetary Fund held that debt relief should be contingent on "structural adjustment" policies. This meant liberalising economies with hitherto large state sectors, and freeing capital and labour markets for globalised trade, without regard to the "level playing field" between nation states that the ILO had once arbitrated.
As for health care financing, the World Bank became a direct ideological opponent. Its health portfolio had advanced since the 1960s from loans for sanitary infrastructure, to population control and thence maternal and child health services, and finally health systems financing. Bank economists now challenged in principle the idea of medical care as a human right under social security (De Ferranti, 1985, p.III;Akin, Birdsall, De Ferranti, 1987, p.1). Public goods in health, they argued, were limited to those such as sanitation and isolation of infectious diseases, where free rider tendencies could undermine collective interests. Meanwhile, personal utilisation of medical services was reconceptualised as a private good: "Individuals are generally willing to pay for direct, largely curative care ... The financing and provision of these private types of health services ... should be shifted to a combination of the nongovernment sector and a public sector reorganised to be more financially self-sufficient" (Akin, Birdsall, De Ferranti, 1987, p.2). In practice, state health spending was curtailed in countries subject to structural adjustment agreements and patient user fees were developed as income sources, while charity or private sector providers grew in importance.
Meanwhile, in the heartland of the European social model, the Organisation of Economic Co-operation and Development increasingly usurped the ILO's role as source of public policy advice. Its discussion papers drew on substantial comparative datasets of national accounts, and accepted the assumption that the general economic pressures challenged the viability of welfare states in their current form. Now the "embattled standard-bearer," the ILO defended the sustainability of social security systems, marshalling champions of postwar European social democracy such as Britain's Abel-Smith and France's Pierre Laroque to oppose ideas like the application of co-payments to restrain demand under health insurance (ILO, 1984, p.59-66;Leimgruber, 2013).
In both the West and Global South, the ILO's promotion of universal, comprehensive health care was under assault. Despite sporadic links with the World Bank, the language of welfare economics was alien to the ILO's social security experts, whose professional backgrounds tended to be in law or actuarial science, with expertise in pensions rather than health. Those like the division's director Giovanni Tamburi, whose remit included advocacy rather than just technical expertise, were now on the defensive. Collaboration with the WHO offered some opportunities for the ILO to press its case for first establishing social security for a limited population, then expanding from that base. By the mid-1980s, both organisations agreed that "it was essential for the 'social' specialised agencies to present viable alternatives to the IMF and World Bank's stringent adjustment policies" (Seth-Mani, 1984; emphasis in the original). In the 1990s these links would become more significant as the landscape of health system reform changed again.

1990-2018: "co-operative pluralism"
As the twenty-first century approached, the ILO relinquished its distinctive role in globalising models of UHC and aligned with the World Bank and WHO. Its acquiescence to pluralism in the health care economy allowed it to re-establish itself as a source of technical advice. In this position, its officials were able to further the Office's agenda of ensuring that social insurance systems were redistributive and favourable to beneficiaries. In some ways, this is paradoxical for national developments illustrated in the ongoing political potential for expansive models of UHC, such as Brazil's establishment of its universal and public Sistema Único de Saúde (1990), following the promise in its 1988 Constitution of health as a human right (Paiva, Teixera, 2014). However, the ILO's ongoing marginalisation under globalisation left it unable to capitalise on such developments.
The rationale for its tripartite structure, with national representation of states, employers and workers, was increasingly redundant in a world where capital flowed through multinational corporations and integrated financial networks, and where production was offshored to low-wage economies. Trade unions focused more on maintaining existing gains rather than defending the vulnerable "precariat" in overseas labour markets. The post-Cold War discourse of human rights had narrowed further to signify civil and political rights, particularly of peoples under communism or authoritarian rule. As social democracy paled before neoliberalism, so political idealism fastened onto this attenuated vision of human rights, now void of the social and economic claims it had once encompassed (Moyn, 2018, p.173-220). The millennial ILO needed to rediscover its raison d'etre.
Arguably, it did not succeed. Its Declaration on Fundamental Principles and Rights at Work (1998) represented the response of Michel Hansenne (director-general 1989-1998) to the challenges of globalisation. It emphasised a small number of core rights (freedom of association, acceptance of collective bargaining, no forced labour, abolition of child labour, no workplace discrimination) which member states would support (Hughes, Haworth, 2010, p.46-53). Instead of enshrining these in a new Constitution, there would be an upgraded form of recommendation that encouraged member states to change. The criticisms of this approach were first that retreating to "fundamental" principles scaled back earlier aspirations without challenging neo-liberal philosophy, and second that exhortation was a weak instrument that exposed the ILO's lack of influence (Standing, 2008). Next came the Decent Work Agenda, launched in 1999 by Juan Somavia (director-general 1999-2012), which was intended to build on the "floor" of fundamental principles by mapping four components of "decent and productive work" which the ILO would champion (employment, rights at work, social security and social dialogue) (Hughes, Haworth, 2010, p.74-76). The ensuing "Global Campaign on Social Security and Coverage for All," launched after the 2001 ILC, concentrated largely on work-related rights rather than welfare (Rodgers et al., 2009, p.168;Standing, 2002). Critics again found that this agenda failed to challenge dominant discourses and urged bolder solutions, like a basic minimum income (Standing, 2008(Standing, , p.365-371, 2011(Standing, , 2017. In the end its effect was "limited" (Rodgers et al., 2009, p.168).
This was the context in which officials shifted their position. The Director of ILO's Social Security Department, Colin Gillion, noted that "the times of the grand designs for health care systems based on economic theory or philosophical and political convictions are over" (ILO, 1993, p.V-VI). Although focused on Africa, Gillion's observation resonated with the "third way" discourse later associated with President Clinton and Tony Blair. It signified a critique of the neo-liberal experiment, whose inadequacies were now becoming apparent, while also acknowledging that the left's optimism about UHC through public structures had been utopian. What mattered now was what worked: "health care financing and delivery in heterogeneous societies cannot be monolithic ... different forms ... can, and most likely will have to, co-exist in one country" (ILO, 1993, p.V-VI). The turn towards "a pluralistic, cooperative system" was also spelled out by Gillion's deputy Michael Cichon.
While social security can directly or indirectly subsidise health care delivery systems for uncovered persons, it cannot alone provide and finance comprehensive care for all in developing countries with a relatively small formal sector. The resulting contribution burden ... would simply be too big and the collection of contributions from the informal sector (has) often insurmountable administrative obstacles (Cichon, 1992a, p.15).
The implication was that the ILO would assist both with funding arrangements, whether social security, regulated insurance, or public health, and with the challenge of integrating these arrangements. In a sense, it was a return to the 1930s and the ILO's earliest interest in coordinating fragmentary services in the incipient phase of municipal and social insurance systems.
Thailand provides a salient case study, for here was a lower-middle income country which successfully achieved universal health coverage. Working through the ILO's Regional Office in Bangkok, officials had helped write its Social Security Plan, which established mandatory insurance for private sector employees (ILO, 31 Oct. 1991). Thailand's route to UHC followed a classic ILO model, beginning with public sector health cover, then the formal private sector, then, from 2001, via the "30 bhat" scheme, extending to the whole population via tax funding with small co-payments. Its success was attributable not to international aid, but primarily to its own political will (Sakunphanit, 2008;Patcharanarumol et al., 2011;Harris, 2015;Kuhonta, 2017). First, Thailand achieved successful economic growth, interrupted only briefly by the Asian financial crisis c.1997-2005. Second, an early decision to oblige newly trained medics to undertake rural service created a cadre of doctors concerned about care for the underserved. Once ensconced in the bureaucracy they lent consistent momentum to scaling up, which the populist Thai Rak Thai government finally delivered. Health leaders were also well-networked internationally − the early creation of HITAP (Health Intervention and Technology Assessment Programme), Thailand's "NICE," (National Institute for Clinical Excellence) illustrates their proactive policy-making (Culyer, Podhisita, Santatiwongchai, 2016). Nonetheless, the ILO's local experts contributed to system design, helping shape payment, benefit and remuneration structures to maximise benefits to labour (Burns, 21 Sept. 2017;Cichon, 29 Nov. 2017).
The other region where ILO officials adjusted to pluralist models was Central and Eastern Europe after the collapse of communism. Several governments, including Poland, Romania, Slovakia and the Czech Republic, requested ILO assistance in changing from statist systems to social health insurance. This reflected their doctors' desires for greater freedoms, the wishes of nationalist politicians to rebuild pre-1945 Bismarckian arrangements, and, above all, the popular revulsion at "monolithic, unresponsive central bureaucracy," now synonymous with communist regimes (WHO, Dec. 1991). Consequently, ILO officials had to accept that their once-preferred option, a NHS model subject to democratic governance, was no longer viable. The task therefore became one of maximising efficiency and equity within social insurance structures, both through the scrupulous actuarial advice offered previously as well as the newer skills of pricing and contracting which were then current in Western systems.
In these settings, ILO technical experts became less isolated. There were relatively few such specialists, networked in with academic advisers working under contract. These formed a "caravan of experts" during the post-1989 transition, variously working for the ILO, World Bank, or WHO and thereby developing interpersonal links (Normand, 20 Apr. 2017). Officials of the ILO and WHO by this point had set aside their differences, pooling the capacity of their "understaffed and underfunded health financing units" and strengthening their "mutual political power base" (Cichon, 11 Jan. 1991). At the same time, the World Bank shifted away from purist market policies, first in the early 1990s to "trickle-down-plus," which acknowledged the need for safety nets for the poor, then later to a rejection of the doctrine that economic growth must precede social expenditure. This followed the success of the East Asian "tiger" economies, whose governments had accompanied market liberalism with judicious social investment (Kanbur, Vines, 2000, p.92-104). In place of a pure neo-liberal credo, a pragmatic consensus emerged on which a new epistemic community could flourish.
Greater optimism also prevailed about engaging the informal sector in low-income countries by means of community-based health insurance. These were small, locally based funds with prices, services, and collecting facilities geared towards rural populations, formally akin to friendly societies and Krankenkassen in pre-Bismarckian Western Europe.
The assumption was that once established, they could be scaled up until universal coverage was achieved, and possibly complemented by social security for others (Barnighausen, Sauerborn, 2002). Within the ILO this approach was compatible with the microfinance initiatives currently favoured in rural development, and its promise further validated the concept of co-operative pluralism.
The new pragmatism also drew from the cross-national policy-learning characteristic of Western health systems since the 1980s. NHS countries like Britain and New Zealand, for example, had introduced an "internal market" in which services were commissioned by "purchasers" in primary care or regional bodies using pricing models first developed in the US to reimburse doctors serving Medicare or health maintenance organisations. A British civil servant tellingly extolled this as "perestroika" for the NHS (Hurst, 1991). Social insurance nations with strong cultures of corporatist negotiation, like Germany and Japan, also refined their reimbursement processes and drug lists, and introduced cross-subsidisation between prosperous and weaker insurance funds. Such techniques made management of pluralist structures possible, using common standards to ensure cost-effectiveness and greater equity. The ILO's task was now to create "an integrated, coordinated financing system which consists of different financing subsystems with clearly defined scopes and mandates" (Cichon, 1992b, p.8-9). In this way, its core ethic of social solidarity could be sustained, even if universal care by right fell by the wayside.
Nonetheless, by the 2010s the ILO's long campaign for health security seemed a low priority. Amongst its five "flagship" programmes of 2018, "occupational safety and health" took precedence over access to medical care, while "social protection floors" were concerned mostly with basic income security (ILO, 9 May 2018). Within the many strands comprising the Decent Work Agenda, UHC fell within the Social Protection Floors Recommendation. These stipulated basic minima of essential services and benefits, and rather than specifying optimal levels, allowed entitlements and universality to be "nationally defined" (ILO, 2017, p.103). There was a particular focus on expanding formal long-term care, both for needy recipients and to provide "decent work" opportunities. The recommendations on financing urged equity and sufficiency with only "limited" out of pocket payments (p.108-115). Whether this current incarnation of co-operative pluralism by persuasion will achieve the desired results will only become clear in the future.

Final considerations
The ILO's centenary year of 2019 coincides with the UN's major push to advance UHC. The main players today are the WHO and the World Bank, whose multi-volume Disease Control Priorities 3 is the basis for designing the "essential" UHC package: this currently prioritises 218 interventions, costed according to platform of delivery (Jamison et al., 2018, p.6-14). Rough estimates based on this package suggest that in addition to national resources, a doubling of donor assistance will be needed (Sachs, 2012, p.945). Even without the populist isolationism currently afflicting the West, this would seem a tall order. Time will tell whether global health policy stays the course or reverts back to selective, vertical approaches (Agyepong, 2018). In any event, today's policy discourse finds the ILO, the agency which first placed UHC on the international agenda, reduced to a supporting role.
As this survey has shown, when the ILO began its work on health care, a variety of national models were favoured by its member states. These ranged from pluralist systems blending sometimes minimal public health provision with social or private insurance, to the notionally comprehensive and universal public systems first developed under communism and then adopted in the liberal democracies from the 1940s. Its bureaucracy initially promoted the expansion of social insurance within mixed economies before striving, at least from 1944, to develop a particular model of universal coverage. It was one in which rights to health security were human rights, and where social policy merited primacy alongside economic development. Its "labourist" social insurance model first broadened to include comprehensive public systems, then contracted to accommodate pluralist and partial approaches. Yet throughout, it stood for designs which favoured the worker and the poorer citizen, initially through insurance arrangements that maximised benefits and minimised restrictions, and later through redistributive social security systems with strong cost controls. However, this history has also shown a protracted marginalization of the values which the ILO championed. Opposition came from employers seeking to diminish obligations, post-colonial nations unwilling to accept Western standards, and ideologues from the United States, the heartland of "free enterprise medicine." The result in the later twentieth century was a retreat to narrow technical expertise. This ceded the ground to others, who rationalised universalism as "productive" and "pro-growth," and were moved by compassion for "needs," not acknowledgement of rights (Universal..., 2018). The ILO's role had become one of advice and exhortation, defending a "floor" of basic standards, rather than advancing the higher ambition to which it once aspired.