Obstacles to sUs universalization : tax expenditures , labor union demands and health insurance state subsidy

In the light of the comparative analysis of health systems, we discuss three strategic phenomena for the SUS universalization, as follows: a) health tax expenditures; b) State funding of private plans for public servants; and c) trade union’s demand for private health plans. Among the ideal types of health systems, SUS is universal in law, but hybrid in practice: Beveridgian in primary health care (PHC) and mixed in specialized/hospital care; without really being universal (public spending is only 43% of total health expenditure). There is a massive state subsidy to the private sector, through health tax expenditures (30% of the federal health budget) and financing of private plans for public servants, which generates incoherence, segmentation of the health system and inequities. Despite the general support to the SUS, the union movements have been using private health plans in collective recruitment (76% of them), reinforcing the private sector. Reducing health tax expenditures including state funding of servants’ private plans would significantly increase the SUS budget and facilitate articulation between health workers and trade unionists, bringing the high strength of unions closer to the long struggle for the universality of the SUS and PHC.


Introduction
Despite the 1988 constitution and laws affirming the Unified Health System (SUS) as public and universal, the private sector receives strong government incentives. Recent provisions have deteriorated the situation of the SUS, such as Constitutional Amendment (EC) 87/2015 (unlinking of union revenue extended to 30%) and EC 95/2016 (limiting spending). According to Ocké-Reis 1 , the Brazilian health system (SUS), similarly to the US private model, began to operate as a duplicate and parallel system. Menicucci 2 affirms it is a dual system in which a public and a private system coexist with different modalities of access, financing, and production of services (p. 1401). Silva 3 uses the minotaur as an illustrative metaphor for the SUS: a monster with a human half that provides the poorest with the real SUS; and an animal half, which transfers public resources to the private sector.
Two SUS issues are consensually recognized. One is its chronic underfinancing [4][5][6] . From this standpoint, a substantial evasion of public resources to the private system means direct damage to public services. Another problem is the insufficient political basis of social support for the realization of (administrative, political, and financial) investment in the SUS, linked to the previous one. It has not been possible to drive the expansion/structuring of the SUS towards universality, equity, and integrality 4,6 . This paper is an essay that addresses these two interconnected problems by articulating three socio-political-institutional phenomena that are hypothetically required to address, albeit not sufficiently: a) the large and growing health tax expenditures; b) state funding of private health plans for public servants; and c) the poor articulation between the health and trade union movement, which historically demands private plans.
Our argumentation develops from the comparative analysis of health systems to synthetically characterize the Brazilian health system against the types of international literature. Next, based on data and analyses produced by other public health authors, we discuss and progressively articulate the three mentioned phenomena.

Comparative health systems
The comparative analysis in the field of health uses health system types to describe the institutional context of care and health policies in different countries 7 . The construction of ideal types allows the understanding of reality through the unilateral emphasis of one or more viewpoints (p. 106) 8 . These types help to outline a health system based on analytical realms such as financing modalities, professional remuneration types, organization forms, and provision of services, among others.
Comparative studies have become a consolidated research field [9][10][11][12][13][14][15] . Pioneer in this area, Field 16 described the historical development of health systems and identified a health system, as care was more or less considered consumer goods, insured services, and state or state-supported services. Frenk and Donadediam 17 proposed a type of state intervention in health care based on two realms: 1) form of state control over the production of medical services; 2) State's relationship with the beneficiaries: with the right to citizenship; social insurance or special categories (military, civil servants, among others); poor and less privileged groups.
A very influential study by the Organization for Economic Co-operation and Development (OECD) 18 identified three basic health system models: 1) Health system with universal coverage (Beveridge model), financing from general taxation and provision of public health care (born in the United Kingdom and spread in Sweden, Norway, Denmark, Finland, Iceland, New Zealand and successively in Italy, Portugal, Spain, Greece); 2) Health system with compulsory social insurance (Bismarck model), with universal coverage, funded by employer and employee contributions through non-profit insurance funds, and provision of public or private care (Germany -prototype, with similar versions in Austria , Belgium, France, the Netherlands, Switzerland, and Japan); iii) Health system with private insurance, funded by voluntary contributions from individuals and employers, with predominantly private health care benefits (US -prototype, with approximations in Mexico, Chile, Australia). This OECD type is widely used to this day 7 .
Moran 15 considers three important governance arenas: consumption, supply, and technology. Consumer governance is concerned with mechanisms to ensure access, including resources allocated to the health system. Care governance addresses mechanisms that regulate hospitals and professionals and public-private relationships. Technology governance focuses on the mechanisms that regulate health innovations and the medical equipment and technology industries.
From the OECD 18 model and Moran's arenas 15 , Wendt et al. 19 identify three health system responsibilities or realms: 1) financing: via taxes, social or private insurance; 2) health care: provided at state facilities by public professionals, or in third sector facilities ("societal-based facilities"), or by private professionals; 3) regulation of the modalities of financing and provision of services. The elements 'state' , 'society' , and 'private' tend to coexist in all three realms. Based on the possible variations, the authors established a taxonomy of 27 possible health systems, three of which are defined as ideal-typical health systems: State Health System, Societal Health System, Private Health System. From these, they identify three forms of health system transformation over time 19 : 1) a "system change" occurs (rarer). For example, a social insurance-based health system switches to a National Health Service, as was the case in Spain, Italy, Portugal, and Greece in the 1970s and 1980s; 2) an "internal system change" occurs (most common). For example, care provision shifts from the state to private services, but state funding and regulation remain, as was the case with the English NHS reform, with the introduction of the internal market, which did not involve the replacement of the state as the principal regulator and financier; 3) an "internal change of one or more realms" occurs (milder), which does not imply a change of system characteristics, as in private health systems with public initiatives to address specific social groups (e.g., poor and elderly/disabled via Medicaid and Medicare in the U.S.).
In the face of health system reforms in industrialized countries, a debate arose in the 1990s over whether they would have generated a process of health system harmonization towards a common model 13 . In this line, Chernichovsky 10:340 identifies an "emerging paradigm" ("technocratic") that combines the advantages of public systems ("equity and macro social efficiency") with those of private and mixed systems ("consumer satisfaction and micro efficiency in the provision of care"); traversing ideological (private versus public) and conceptual (market versus centralized planning) lines, and considering the OECD's tripartite division as obsolete. Other authors, however, emphasize the divergence between health systems as a result of cultural diversity and local contexts 16,20 .
Despite this debate, the coined OECD type reformulated by Moran 15 and Wendt et al. 19 is still the most widely used and identifies three ideal-typical health systems ( Figure 1).
Save for rare exceptions, there are no pure systems, as we mentioned in the analysis of trans-formations. However, this classification is useful for establishing how close each national health system is to the three ideal types.

sUs between segmentation and privatization: a comparative perspective
How do we categorize the Brazilian health system? If we consider economic indicators (public vs. private spending), the proportion of the Brazilian public health expenditure (43,3%) is currently lower than the U.S. (46,6%), a mostly private health system model 21 . Brazil is far from the countries with a National Health Service, whose public spending is more than 70% of total expenditure. A Brazilian health system Americanization process (p. 1351) 22 was commented: a clear paradigmatic approximation of the Brazilian case to the American (p. 153) 23 was noted as early as the early 1990s; state action did not seem to include the whole population, as in the English model, but cater to the lower purchasing power social sectors(p. 155-156) 23 . This situation does not coincide with particular optimistic and overly evaluative views of the SUS(p. 754) 24 , which see real SUS as universal.
The public-private relationship is a vital realm to understand the structure of the subsystems that underpin the health system. Thus, high 'segmentation' and 'fragmentation' are striking features of the Brazilian health system. Scholars have analyzed the contradiction between the original design of the SUS (constitutionally ruled as a universal and hierarchical public system based on primary health care (PHC)) and its segmentation into three subsystems: the SUS, the supplementary subsystem, and the direct disbursement system [25][26][27][28][29] . Gurgel et al. 30 recognize three parallel systems: a universal health system, mainly PHC; a system in which the public buyer purchases secondary and tertiary care from private providers; and one where private policyholders buy services from private providers, with state subsidies, as we will see.
A naturalization of the asymmetries of coverage, access and use of health services(p. 1388) 25 was noted between SUS users and health plan users 2 . However, Bahia 24 stresses the insufficient academic reflection that conceals the role of the state as an active agent [...] to preserve segmentation(p. 1394).
However, we should clarify that some health system segmentation is not an exclusively Brazilian phenomenon. It is relatively common in small size. The problem with the Brazilian health system is the level of segmentation. For example, private care via plans/insurance in the European Union accounts for less than 5% of total health expenditure (p. 1443) 28 . Another example: in Portugal, besides the National Health Service, the voluntary public or private subsystems cover 25% of the population 31 . However, the segmentation of the Portuguese health system does not imply a significant public-private funding gap as in our country. Brazil and Portugal have a similar population fraction served in the public system, about 75% of the population. However, in Brazil, this fraction absorbs only 42% of health spending 32 , while absorbing 66% 31 in Portugal. The Portuguese public health expenditure is one of the lowest in the European Union (EU). The gap is even more evident if we compare Brazil with the most prominent EU countries (Table 1).
Despite universalist legislation, the SUS is a non-universal, underfunded public system, structured with predominantly state provision in PHC (under municipal responsibility), a mixed provision in specialized and hospital care (with private predominance), and large tax expenditures promoting the private system. Therefore, the Brazilian health system is a mixture of the two extremes of Figure 1, if we consider that social insurance is based on compulsory contributions; or as a mix of the three ideal types, considering that most private plans are collective, with payment by employers, and that their state funding via tax expenditures is the rule, tending to characterize the tripartite financing… of the Bismarckian insurance model (p. 160) 33 , although plans are voluntary.

Brazilian health tax expenditures
Tax expenditures are waivers of government revenues focused on a narrow group of taxpayers, sectors, or regions 34 . Its most essential forms in health in Brazil include tax deduction of spending on private health plans and insurance for legal entities and individuals and the financing of private health plans for the federal government, state, and municipality public servants 35,36 .
Tax expenditure is relevant in the public budget, and its effect is equal to that of a direct expense. In the health sector, it escalated in the military dictatorship and has grown in recent decades 27 . Machado et al. 33 calculated that health tax expenditure (GTS) was equivalent to 14% of total federal tax expenditure in 2018 (R$ 39 billion).
Medical expenses in the Individual Income Tax (IRPF) are persistently the largest item in the GTS, almost half of the total in 2013 (46.8%) 35 , and benefit the richest 37 . In 2013, 60% of the deduction for medical expenses was concentrated on those who received more than ten minimum monthly wages 35 . In 2016, health insurance spending was 71% of the IRPF 38 medical expense deductions.
GTS is a crucial player in the private sector reproduction 27 and does not influence the regulation of the National Supplementary Health Agency (ANS). Also, there is no ceiling for the health expense deductions of the IRPF and IRPJ, which depend solely on the magnitude of these expenditures 39 .
GTS accounts for 30% of the Ministry of Health expenditure, induces health market

National Health Service
Social Insurance Private Insurance growth, and is highly unfair 1 . Benefiting the wealthiest 10% of the population 1 , GTS with private plans is a factor generating inequalities in access and use of health services, as these plans generally cover higher-income, white, more educated people, within specific industries and residing in the capitals/metropolitan regions 40,41 . This chronic problem must be tackled if we are to pursue a universality with health care equitable in Brazil.

Organized workers, sUs and health plans
The construction of the SUS has a critical and flawed point in its relationship with union movements. In 1990, Faveret and Oliveira 23 described an adverse effect of SUS universalization without proportional expansion of the public service network, whose deteriorated access and quality standards for formalized workers induced massive migration of these workers to private health plans. Their unions guided the inclusion of these plans in negotiations with bosses. This migration eliminated from the PHC users the social groups with strong political and social pressure power; pressure diverted for private sector expansion, which was already subsidized by the public. The late 1980s, regardless of the creation of the SUS, were a time of intense health insurance growth 42,43 .
Studies such as those by Dias Filho 33 and Santos 44 that analyzed in detail the standing of organized workers contextualized in various historical moments, indicated that, already in the birth of SUS, although supporting the construction of the SUS (primarily via union confederations), the union movement carried ambiguities vis-àvis this system, holding a somewhat corporate stance in the discussions of the Eighth National Health Conference, the Health Reform Commission and the 1988 National Constituent Assembly. Gallo et al. 45:416 also warned of the poor adherence of social, trade union and popular movements: The Health Reform is not reinforced by popular participation except in its discourse (p. 416). Menicucci 46:76 stated that, although union leaders join SUS in their fora, a contradiction between immediate interests and the universalist political-ideological orientation (p. 76) is note, with growing "health insurance culture" in unions. In 2001, about 75-80% 42 of Brazilian health insurance contracts were collective. In 2007, this proportion was 70% 47 , and 76% 48 in 2018.
There is some recognition of the strategic relevance of the distance/disengagement issue of the health movement of trade union movements, whose absence (of its political force) hinders the struggle for the structuring of the SUS. Theoretically discussing the centrality of a labor base organized as a strategy for the struggle for the SUS, Santos 49 concludes that while being in a common and converging field of dilemmas, the dialogue between health and unionist reflections remains refractory and discontinuous par excellence (p. 224). Considering, empirically, the Brazilian constituent debates, the persistent and high rates of work accidents and their underreporting in the country, and the growing and robust union demand for private health plans, the author 50 concludes that the health argument has laterality" and a "lapse of theoretical maturation (p. 243) in the trade union movement, beyond the sparse and discontinuous dialogue. The universalist civilizing argument of the health reform is not inserted in grammar and labor values, prohibited by corporate union rationale. It seems coherent and empirically appropriate to consider that the low emphasis placed by organized workers on the struggle for the SUS is a significant loss that significantly weakens the efforts of the post-1988 health movement. Pro-SUS political and social forces have not been able to overcome the huge funding difficulties(p. 2010) 51 .
One aspect of this question relates to the PHC filter function as a mediator of universality, equity, and integrality. While consensual in the international literature, SUS regulations and public health's discourse, PHC seems to be rarely thematic in its relationship with society in general and the trade union movement. Most Brazilian health plans provide direct access to specialized medicine, but such access is much more expensive and does not result in benefits to users 52,53 . However, this access seems established as desirable and coveted in the common sense, in Brazilian culture, public servants in general and the SUS, as well as among the highest paid workers 46,54 . The low penetration and use of PHC in these social groups are shown by general data and relative absence. In 2009, only 16% of people with incomes above five minimum wages were registered in a Family Health team 55 . PHC is absent in union demands. Recent filter function offers via Family and Community Doctors in health plans are not labor demands (as far as we know). We found scarce literature on this subject in public health. In 2005, one paper advocating PHC's filter function 56 was debated: while converging with this line, essential hygienists partially relativized this function 57,58 .
If middle classes, the highest-paid workers, public servants, and the SUS directly access the specialists via private plans (pressured by the queues and delay for specialized care in the SUS 55 , but also perhaps by preference), the theme of PHC becomes a relevant problem, because the SUS should ensure its access via the Family Health Strategy (ESF) 53 , which should enable, when necessary, specialized care. This is heightened by the view of SUS and PHC services, especially in the middle and upper-middle-class income sectors, such as "poor quality" public services (p. 250) 4 .
A relevant factor in the complex set of difficulties and challenges involved there seems to be the adherence, in practice, of unionized workers and public servants to private health plans 43,46-50 . Although civil servants seem to be small stakeholders given the size and complexity of national policy, this argument deserves attention.

Inconsistencies and contradictions: arguments for an agenda
Within the GTS is a symbolically significant portion that hampers coping with underfunding and the insufficient social base to fight for SUS expansion through PHC: the public servants of the three spheres of government, state-owned companies, the three governmental branches, including professionals from the SUS, employees of educational and research institutions (the 'health intelligence'), enjoy state subsidy for the use of private health plans (besides deductions in the IRPF, accessible to all).
We did not find studies on this preference and use by public servants and the SUS, but this is an indicator of the low investment in research on this topic by public health. Analyses on the use of private plans do not clarify subsidies to public servants. Data from IBGE (National Health Survey) and the ANS also did not 41 . Also, we could not find any federal legislation on this subject, only its authorization in Article nº 230 of Law Nº 8.112 of 11/12/1990, which addresses the legal framework of federal civil servants. Two Ordinances of the Ministry of Planning, Budget and Management govern the public subsidy to the private plans of the civil servants of the federal Executive Branch: one 59 establishes the rules, while another 60 establishes allowance values by age and salary range. States, municipalities and other branches have their own rules.
Ignorance about this universe seems to be widespread. As an illustration, we provide some data from the Federal University of Santa Catarina (UFSC), the Municipality of Florianópolis (PMF) and the state government of Santa Catarina, 2019. UFSC provides its employees and dependents with a subsidized health plan (Unimed Grande Florianópolis), with 15,285 users, used by 76% of faculty and 65% of administrative technicians, and is Unimed Grande Florianópolis' largest beneficiary portfolio. The PMF has about 35,700 employees and offers a subsidized health plan (SC-Saúde -also available to state civil servants and the military), with 12,372 users (34,6% of municipal servants). Of the 46,680 state civil servants and the military (including active and retired) and pensioners living in Greater Florianópolis, and their 60,066 dependents (106,746 people), 65% use SC-Saúde. If we consider only the servants of the State Health Secretariat, out of a total of 26,170 people (active, inactive, pensioners, and dependents), 60% are SC-Saúde users (Chart 1). . a) Personal plan: UNIMED-UFSC -Contract 010/2014. Servants receive allowance by age and salary range (Ordinance Nº 8), and the plan is solidary: equal price for all users, regardless of age . Total of 15,285 active beneficiaries, relative to 6,634 holders (the others are dependents), of which 48% are faculty (76%) of the total UFSC faculty) and 52% technical-administrative (65% of them). It was not possible to collect specific data on health faculty promptly . b) Financial reimbursement: requested by servants to use the allowance in another health plan: 727 beneficiaries (on 26/07/2019) PMF 2 . Health insurance provided to municipal and dependent servants (SC-Saúde) is the same as that provided by the state government of SC to state servants . Of the approximately 35,700 PMF servants, 12,372 are users (including dependents) of SC-Saúde . The insured holders pay a monthly contribution as per their salary range regardless of age (if > R$ 3,600.01, their contribution is R$ 323.47, plus R$ 40.00 for each dependent) . It was not possible to specify data on the Municipal Health Secretariat servants promptly State servants residing in the Greater The sum of these three groups of public servants and their dependents amounts to 33,0% of Greater Florianópolis health plan users (details and sources in Chart 1). Also, a subsidized portion of civil servants (judiciary, other federal institutions, among others) were not included in the calculation, which makes it an underestimated figure.
The magnitude of the number of these users shows the relevance of the phenomenon and the need for further research on this subject. This situation causes public servants and SUS to experience a contradiction that undermines part of their adherence to the struggle for the construction and universalization of PHC and the SUS, and injects a significant level of inconsistency in their universalizing health discourse/argument. A metaphor can illustrate such a situation: the situation would be similar to that of chain restaurant chefs who do not eat the food they prepare, and routinely have lunch in other restaurants. Even with the many caveats, such a metaphor conveys a crystal clear message: it seems unlikely that the population and workers/unions will value, believe and fight for the construction of SUS via PHC if their professionals/intellectuals/managers/leaders do not use the ESF and do not expressly and concretely advocate, through research and political/legislative proposals, through their institutions and associations (CEBES, ABRAS-CO, Public Health and Social Medicine departments) the end of privileges of subsidized use of health plans by public servants, beyond the end or reduction of the GTS.
The credibility and legitimacy of these professionals/leaders/institutions are reduced if in practice they use private plans for themselves (and thus subliminally support them) with state subsidies, directly accessing specialized medicine, which is denied to the citizen in PHC. If we want the middle class and labor unions to stop fighting for private plans, it seems consistent, educational and exemplary that hygienists and their institutions do the same, and advocate for the end of their privileges of subsidized use of health plans. In this regard, we looked for manifestations in the public health literature and hardly found any evidence.
In a recent document, ABRASCO 61 mentioned an increasing trade of private plans ... with higher tax subsidies given to buyers of these plans (p.10); and advocates ending the subsidy to health plans, and transferring to the SUS the new resources that the state will collect from companies and people who buy health plans (p. 17). Howev-er, there are no specific proposals on subsidized private plans from public servants. Exceptions are scientific papers that generically show the situation: 2015 data estimated that about 20% of private health plans were directly funded from public funds 25 . In 2010, 22.5% of private policyholders were state-funded at the three governmental levels 55 .
However, we must do justice to managers and researchers who specifically discussed the problem. Dias Filho 43:159 argues that this criticism applies to past and present and "should be transmuted into a self-criticism" of the health movement, which does not relinquish these privileges, leading to a paradoxical situation in need of in-depth review. The paradox is that the union confederations stand side by side with the health movement advocating [the SUS], but, like the latter, they adhere to private plans (p. 166) 43 .
Thus, the criticism that applies to the positioning of the working class extends to relevant portions of the health movement itself. The universal model of health care is advocated, but in fact, it adheres to the boundary between the Bismarckian and residual-liberal insurance models (p. 160) 43 . That is, the health movement speaks of "a SUS" that is "advocated" but not "used" (or even worse, one says "uses", because water is treated and medicines controlled...) (p. 12) 32 . Dias Filho (p. 159-160) 33 credits Maria Lúcia Werneck Vianna and José Gomes Temporão with statements about the state bureaucracy having tax-funded private plans, which affects the indispensable solidarity for the universalization of the SUS. Tesser and Norman 62 assume: instead of a universal SUS that "we" use through the ESF, what is being done [...] seems more like PHC to the "others", the poor (p. 878).
We emphasize the scarce research focusing on thematic issues and contradictions and concrete proposals for their transformation. We argue the need to develop research and policy and legislative proposals on the subject by health reform advocacy entities, researchers and groups, with a view to an institutional solution to this specific problem -one among many others to be faced in order to overcome the complex historical and structural and political heritage of the Brazilian health system, its segmentation and inequity.
The struggle for the construction of the SUS via universal PHC would greatly benefit from the adherence of the trade union movement, but it seems sensible and logical to consider that the health movement will only be able to feature or participate in something like a "political formation program" of trade unionists, suggested by Santos (p. 1419) 48 , if it solves most of this genetic ethical-political problem by advocating, as ABRASCO still did, in general, for the drastic reduction of health tax expenditures, explicitly including the end of state-funded private plan privileges for civil public servants.

Final considerations
The Brazilian health system is hybrid and highly segmented, with severe chronic underfunding problems. Paradoxically, the state finances the private sector through tax expenditures(a), including health plans subsidized by public servants and SUS professionals (b). The effort to universalize PHC and SUS lacks the social and political strength of the trade union movement, whose demands have been directed to the private sector (c). The same nature of (a) and (b) means historical state stimulus to private plans, with chronic SUS budget bottlenecks. The persistent articulation between (b) and (c) means politi-cal-ideological sabotage of the social valorization of PHC and SUS and promotion/advocacy of the private sector by example. The articulation between (a), (b) and (c) maintains politically powerful social groups, especially trade union movements, far from the use of PHC and the SUS, and the struggle for their universalization. Dismantling (a) can mean a significant increase in the SUS budget; it is difficult to change (c) without dismantling (b). Disarming this triple articulation is strategic/necessary to build the PHC and SUS universality and equity. Advancing in this direction requires public health, as an academic field and binding element of social and political movements, to recognize and denounce (a) and (b) as unfair; produce research and specific proposals for its extinction and actively fight for them. This will facilitate the work of transforming [c] into making the SUS and its PHC concrete union claims. Together, the trade union, health, and other movements could advance the long way towards the universalization towards the constitutional Beveridgian public SUS.

Collaborations
CD Tesser and M Sarapioni jointly conceived the article and wrote the introduction. M Serapioni wrote the first version of the second and third topics; CD Tesser the fourth and fifth topics; the sixth and seventh being written together. The authors also participated in the bibliographic updating, critical review, writing and approval of the final version of the article.