In the wake of new evidence of market power of firms over workers, as a widespread economic fact, it is worth revising how the unusual interplay between merger control and labor market regulation in Brazil has played out over the past years. The analysis of recent decisions by CADE, the Brazilian antitrust agency, reveals persisting inconsistencies in remedial practice since 2000s, about relevance of “employment measures” in merger review procedures, which could lead to conflicts between the CADE and public institutions responsible for enforcing labor laws. This short essay explores the consequences of such problem for the implementation of antitrust and how legal argumentation would be key to improving competition policy in Brazil.

Alberto Barbosa Jr1


The perceived success of competition policy in Brazil compared to other jurisdictions in Latin America2 hides inconsistencies in the decision-making of the Brazilian antitrust agency, Conselho Administrativo de Defesa Econômica (“CADE”) regarding an unusual topic: the interplay between the merger control regime and labor market regulation. While such policy interactions could be seen as an old miscarriage of “antitrust justice” in developing countries, influential scholars are now advocating, in a turn of the tide, for merger remedies to prevent monopsony power in U.S. labor markets.3

In the wake of new evidence of market power of firms over workers, as a widespread economic fact, it is worth revising how the intersections of those different regulatory domains have played out in Brazil over the past years. As I have argued elsewhere,4 the unresolved contradiction in CADE’s remedial practice about the relevance of “employment measures” in merger control may lead to conflicts between the antitrust agency and public institutions responsible for enforcing labor laws.

This short essay explores the consequences of such a problem for the implementation of antitrust and how legal argumentation would be key to improving competition policy in Brazil.


During the 1990s, CADE official documents indicated an explicit concern with the possible impact of transactions on the employment level in the economy, as cost-cutting mergers could allow for the discharge of workers made redundant after businesses’ reorganization. CADE’s 1996 Annual Report, for instance, stated that competition policy promotes social welfare by minimizing both private and social costs, so that merger control should involve a more comprehensive analysis of the outcomes of transactions under review.5

According to that Annual Report, the relevant social costs to be considered include negative externalities of mergers affecting labor markets, such as the frictional unemployment of low-skilled workers that, as a short-run byproduct these transactions, could eventually become structural unemployment. In view of the risks of unemployment in the long run, legal measures (i.e. antitrust remedies) applied by CADE in merger review procedures could thus neutralize macroeconomic inefficiencies produced by transactions deemed efficient in a microeconomic sense.6

To that end, CADE and the Brazilian Ministry of Labor entered into a cooperation agreement to jointly elaborate and monitor job training programs, which would be implemented by merging firms as condition for clearance of the transaction under review by the antitrust agency.7 The legal basis for mandated job training programs, as stated in the CADE’s Annual Report, would be provisions of the Brazilian 1994 competition legislation determining that employment levels are taken into account in the design of remedies against anticompetitive transactions.8 Again, according to that Annual Report, only then would the enforcement of competition laws converge with the constitutional clauses that establish the promotion of full employment in Brazil as a guiding principle for economic regulation.9

An extensive analysis of CADE’s merger decisions revealed that antitrust remedies were eventually applied to minimize harmful effects of mergers on the workers employed by the merging firms. In fact, two types of these “employment measures” were identified, both of which were designed to cope with possible plant-closings and mass layoffs: (i) temporary maintenance of the employment level at the company’s plants and facilities; and (ii) implementation of job training programs for workers dismissed due to the internal reorganization of the merged company. These cases are indicated below:

Nonetheless, as CADE’s merger decisions developed after 2000, those employment measures seemed to be tacitly excluded from the agency’s repertoire of antitrust remedies. This could mean that, for the last decade, the agency has adopted a different reading of the competition statute and the constitutional provisions that set the basis for economic regulation in Brazil. Any such departure from CADE’s early remedial standard in merger cases, however, was not followed by further notice or account for the new administrative practice.

To be sure, legal remedies requiring companies to maintain their employment levels were applied after 2000 as a sort of preliminary order, meant to assure the effectiveness of CADE’s final decisions. In these cases, however, the remedy’s purpose does not seem to be the promotion of full employment nor the defense of competition in labor markets.

For example, in Sadia/Perdigão,10 the transaction was cleared in 2011 under commitments by the merging parties to refrain from discharging, without cause, the personnel of manufacturing facilities that would be sold as a divestiture remedy. The same kind of commitment of “employment level maintenance,” ancillary to the divestiture of production units, was conditional to CADE’s approval in merger cases like Camargo Correa/CIMPOR11 and Rede D’or/Hospital Sta. Lúcia,12 decided in 2012 and 2013, after the enactment of new legislation13 that omitted any explicit reference to employment level in CADE’s legal remedies.

In any event, it remains unclear whether the Congress, after the 2011 legislative reform, has deliberately excluded concerns with the regulation of labor markets from the scope of antitrust policy. Despite the lack of a relevant provision in the current competition statute, commentators still hold that, in view of the Brazilian Constitution and its “full employment” clause, the interests of workers affected by potentially anticompetitive transactions should be reflected on the legal remedies adopted in merger control.14


Recent merger cases have once again led to discussions of the relevance of remedies to the maintenance of employment levels after the transaction is closed and cleared by the antitrust agency. Particularly, it seems that the 2016 decision in HSBC/Bradesco15 was a lost opportunity for CADE to define the contours of the policy interactions between merger control and labor market regulation in Brazil.

In that case, unionized bank employees, represented by the Sindicato dos Bancários e Financiários de Curitiba, joined the merger review procedure as an interested third party. According to the employees’ union, after the acquisition of HSBC’s operation in Brazil, Bradesco should be required to maintain its pre-merger employment level, in order to secure the transaction’s “global efficiency.”

In response to the union’s claim that the maintenance of employment levels had already been utilized as a merger remedy, CADE stated that such measures could not be understood as efficiency-enhancing in antitrust sense. And, for that reason, according to the reporting commissioner, these legal remedies could not be imposed by the antitrust agency nor demanded from the merging parties when negotiating commitments for the conditional approval to a transaction.

More generally, the reporting commissioner for HSBC/Bradesco and other members of CADE’s internal tribunal concurred in their understanding that concerns relating to a merger’s impact on the labor markets (e.g. redundancy of employees) should be addressed by specific public policies, for which the antitrust authority is not responsible. That understanding, however, along with the statement by the reporting commissioner that commitments to maintain employment level could only be adopted voluntarily by the merging parties, seem inconsistent with CADE’s decision-making in the 1990s.

In the early days of the 1994 Brazilian antitrust statute, parties to a merger review procedure had little room to bargain over the remedies proposed by CADE. The fact that CADE and the Ministry of Labor had cooperation agreements to implement job training programs also reinforces the conclusion that, until 2000, there was an explicit goal in Brazil’s antitrust policy that was mandatory also for merging firms: the preservation of the employment level in the economy via remedies that “neutralize macroeconomic inefficiencies” arising from anticompetitive transactions.16

Finally, the “employment measures” associated with structural remedies were considered in the 2019 case Prosegur/Transfederal,17 although promptly dismissed because, for one of CADE’s commissioners, the workforce employed by the merging parties was not particularly specialized. This could mean that the divestiture of assets from the merged firm, should it be necessary for antitrust clearance, would be attractive to actual or potential rivals regardless of the employment level at the divested facilities.

Again, this recent case seems to reinforce CADE’s current practice, as established after 2000, to regard employment measures as remedies not meant to address competition concerns in labor markets, but rather as ancillary to the success of structural remedies.

Interestingly, Brazilian scholars have identified CADE’s awareness of employers’ anticompetitive conduct that could harm competition among employees.18 For instance, concerns with no-poaching agreements19 and the exchange of sensitive information about salaries and benefits have been already investigated in CADE’s conduct enforcement procedures.20 This comes in tandem with recent law-and-economics literature proposing that U.S. antitrust authorities make use of remedies to cope with monopsony power in labor markets.

However, those same commentators concede that, as regards for CADE’s current merger practice, the agency has not consistent or

explicitly raised concerns relating to the effects of potentially anticompetitive transactions in the relevant labor markets. 21


This nearly forgotten contradiction in Brazilian antitrust, about the discontinuation of employment measures in merger control, remains unresolved as a legal matter to the extent that the policy reasons behind any such change in CADE’s administrative practice are still missing or, at best, not entirely articulated with the agency’s previous decisions. Unsurprisingly, the absence of explicit grounds to motivate regulatory decisions raises concerns as to the transparency and accountability of decentralized policy-making in the administrative state.

Political analyses addressing the functions of independent administrative agencies commonly focus on issues of democratic legitimacy of non-majoritarian institutions.22 In the procedural dimension, the lack of justification for abandoning the employment remedies means that CADE has failed to meet the “reason-giving requirements” for policy-making,23 which ensure the legitimacy of regulations via judicial review, public participation, or democratic debate.

Likewise, to the extent that any such regulatory shift has not yet been justified, CADE’s decisions in merger cases remain contradictory, reducing the legitimacy of Brazilian antitrust in a more substantial dimension, related to social expectations about policy consistency and technical expertise of regulators.24 However, while legitimacy problems become relevant in contexts of diverging regulatory decisions, these concerns do not tackle the specific consequences of inconsistent decision-making for the policy implementation process.

In the traditional framework for implementation analysis,25 the regulatory output of agencies involves, along with adjudicatory decisions and regulations, some “general transitive goals,”26 which include both statutory objectives and the declared intentions of agency officials about the policy pursued. The same analytical framework suggests that the successful implementation of regulatory policies depends on supportive public institutions – the sovereigns of the regulator – that control legal and financial resources provided to implementing agencies.27 Considering the provision of legal resources, a non-supportive attitude by sovereign institutions (i.e. legislatures, chief executives, and courts) could impact negatively on policy implementation, creating conflicts over the specific goals to be carried out by the implementing agency. Courts in particular, when opposing statutory policy objectives, can restrict the implementation process by adjudicating cases according to adverse interpretations of the relevant legislation.28

Getting back to Brazilian antitrust, the unmotivated interruption in the use of employment measures confers some ambiguity on the regulatory outputs, for CADE has stated in official documents from 1990s that employment levels are relevant in merger control. This inconsistency in CADE’s decision-making aggravates the institutional risk that policies outside the domain of the governing statutes affect the implementation of the regulatory goals pursued by the agency.29 In the final analysis, it means that the ambiguous antitrust practice in Brazil regarding the role for merger control in the regulation of labor markets could even allow labor policies to impair the current implementation of the competition policy.

There is room for two examples here. In 2012, CADE reviewed a merger between two airline companies, Gol and Webjet.30 Despite the expected economic efficiencies, the agency found that this transaction would harm competition in several markets, so antitrust clearance was dependent on legal remedies to protect consumers of air transport services from the market power held by the merged firm. Following the merger review procedure, after the transaction was consummated in compliance with CADE’s approval conditions, 70 per cent of acquired company’s employees were discharged.

In response, the Labor Prosecutor’s Office brought a civil action against the airline companies, claiming that the layoff amounted to unlawful discharge in violation of the labor laws. After a first trial decision and appeal proceedings, the reviewing court of the Brazilian Labor Judiciary found the layoff to be illegal. The appellate court also declared − though not as primary ground for its ruling − that the protection of workers against mass layoffs is an implicit condition for antitrust clearance of mergers.31 It means that, regardless of the legal remedies designed by CADE in that merger case, the labor courts’ reading of the antitrust statute would imply some extension of the competition policy also to labor markets.

A second example is a merger involving two large orange processing companies, Fisher and Citrovita, review by CADE in 2011.32 Because the transaction was found to have anticompetitive effects on the demand-side of markets for raw fruit, the conditions for antitrust clearance involved legal remedies meant to control the buyer power of the merged company, in order to protect orange farmers and, indirectly, juice consumers. After the merger review process, a juice processing plant was shut down as a stage in a process of business reorganization, and 173 employees were discharged.33

This time, the Labor Prosecutor’s Office responded by starting administrative proceedings to investigate whether CADE could be held liable for mass layoffs resulted from merger approved by the antitrust agency.34 Based on allegations that the agency had not cooperated with the investigation, the public prosecutor in charge brought a legal action in 2013, requesting a court order to access confidential records of merger cases.35

It is worth noting that both of those cases do not correspond to the typical judicial review by courts judging the legality of policy outputs of administrative agencies. The jurisdiction of labor courts in Brazil covers disputes related mainly to employment relations and collective bargaining, having no power to review the lawfulness of decisions by the antitrust agency.36 Nevertheless, it seems clear that an escalating conflict between CADE, labor courts and public prosecutors could impact the implementation process in a way equivalent to the shortage of legal resources produced by non-supportive sovereign of the regulator.

The institutional scenario conducive to those institutional conflicts between CADE and other public authorities, regardless of the its dimensions, remain present in Brazilian antitrust policy even after HSBC/Bradesco, despite the fact that, in such case, both employer firms and employees’ union had their say in the merger review procedure.


Granted, the lack of justification for changes in CADE’s administrative practice since 2000 cannot be seen as to immediately cause any threat against the authority of its own regulatory decisions. What can be argued, thought, is that, had CADE been more transparent about its decision-making, the remaining inconsistencies in merger control could have been resolved, and courts and public prosecutors would be less likely to interfere with the implementation of the competition policy carried out by the agency.

Considering again the legitimacy problem, the ideal of transparency in agency decision-making would imply not only presenting reasons that inform the administrative activity, but, most importantly, making arguments to persuasively justify policy decisions.37 This point is premised on the recognition that regulatory decisions by independent agencies are embedded in a legal discourse. It means that, although legal rules alone do not entail determinate and objective outcomes of regulatory policy, the law does influence the implementation process.

Following this line of reasoning, two theoretical perspectives become pertinent to the implementation issues analyzed here. From an institutionalist approach, law can be understood as a particular constraint on bureaucratic behavior, i.e. an institution that structures regulatory action, shapes preferences of agency officials, and is reciprocally affected by their actions.38 This constraining effect of law in agency decision-making, however, does not derive only from the meaning of legal texts or the content of legal categories.

Moving on to a discursive approach, the law extends its effects on bureaucratic behavior also at the “less tangible, but no less significant, level of affecting expectations as to the form and operation of a regulatory or decision-making process.”39 These expectations, concerning the values associated with law, play an important role in the functioning of a regulatory system, as they bear on the way that legal rules are expected to be applied.40 This cultural aspect of legal discourse, therefore, reveals the conventions, assumptions and world views that constitute the tacit understanding shared by those who interpret the law.

Accordingly, the perception of the intangible effects of law on agency decision-making leads to the notion of “community constructed interpretations.”41 This idea becomes useful as basis for regulatory strategies meant to overcome legal uncertainty problems by creating an interpretative community among agencies, courts and target-groups of regulation.42 Although the creation of such regulatory interpretive communities demands long-term development, mostly based on the uniform education and training of their members, another contributing factor would be deepening the argumentative process underlying the decisions by the implementing agency.

In fact, legal disputes on the lawfulness of policy decisions, involving agencies and reviewing courts, and recurrent complaints by regulators regarding the judiciary’s alleged misunderstanding of regulation, illustrate the contrast between interpreters with different interpretive pre-understandings.43 In Brazilian antitrust, the problem of distinctive legal rationalities utilized by CADE and labor courts to interpret the Brazilian competition statute becomes clear when it comes to remedies like the “employment measures.” CADE’s regulatory action presupposes an economic analysis of the market power that may result from anticompetitive mergers; to labor courts the unbalanced bargaining power favoring companies is the very assumption behind the laws that regulate the industrial relations between employers and workers.

However, in construing the meaning of statutory texts, agency decision-makers could build comprehensive arguments about interpretation44 to convince those participating in the legal discourse, such as courts, and justify regulatory decisions. To that end, agencies would, for instance, engage in justification efforts based on the coherence of legal system, interpreting the law and arguing that their decisions fit a rationally constructed set of rules, principles and past decisions.


Having said all that, this essay does not claim that CADE and labor courts and other public authorities integrate to the same regulatory system, let alone propose that an interpretive community should be established between them. Nevertheless, in taking regulatory transparency as an instance of legal argumentation, the purported consistency in antitrust practice becomes an implicit requirement for the argumentative justification of decisions by the agency.

In view of the economic analysis inherent to regulatory decision-making, consequentialist arguments45 about market efficiency come to be as important to antitrust law as the coherence arguments mentioned above. Thus, if potential risks to antitrust implementation are to be tackled via legal argumentation, as suggested, the question about possible intersections between competition policy and other regulatory domains would be better answered with a plausible combination of arguments based on both coherence and consequences.

Accordingly, the recent evidence of monopsonistic or oligopsonistic competition in most labor markets should be accounted for in CADE’s argumentative efforts to justify implementation choices in competition policy. Most importantly, the contradictions arising from inconsistent remedial practice about the “employment measures” in merger control could be finally unraveled through interpretative arguments addressing four topics:

(i) how the reading of the 1994 antitrust statute had changed in the late 2000s with respect to the maintenance of employment level at the merging parties, as legal remedy against potentially anticompetitive transactions;

(ii) to what extent the Brazilian competition law, as construed after 2000, incorporated economic knowledge (i.e. theoretical models and empirical analyses) to evaluate the efficiency of the “employment measures”;

(iii) the impact, if any, of the 2011 legislative reform with respect to CADE’s remedial practice in merger control; and

(iv) the legal and economic grounds for the current implementation of antitrust and how it can be reconciled with CADE’s past policy choices to promote the interplay of merger control and labor market regulation.


1 SABZ Advogados, São Paulo.

2 Francisco Ribeiro Todorov & Marcelo Maciel Torres Filho, History of Competition Policy in Brazil: 1930-2010, 57 Antitrust Bull. 207, 254 (2012).

3 See Ioana Marinescu & Herbert J. Hovenkamp, Anticompetitive Mergers in Labor Markets, 94 ind. l.J. 1031 (2019) and Suresh Naidu, Eric A. Posner & Glen Weyl, Antitrust Remedies for Labor Market Power, 132 HArv. l. rev. 536 (2018).

4 Alberto Barbosa Jr., About Law, Economics and Argumentation: The forgotten case of labor concerns in Brazilian competition Policy and Why it still matters, 5 u. BAlt. J. int’l l. 137 (2017).

5 CADE, relAtório AnuAl 1996 37-38 (1997).

6 CADE, relAtório AnuAl 1996 39 (1997).

7 CADE, relAtório AnuAl 1996 37 (1997).

8 See Law No. 8,884 of June 11, 1994 [Derogated], Art 58, Parag.1: “Performance commitments [legal instruments for antitrust remedies] will take into consideration the extent of international competition in a certain industry and their effect on employment levels, among other relevant circumstances.” 

9 See BrAziliAn FederAl Constitution oF 1998, Art. 170, IV.

10 Ato de Concentração No. 08012.004423/2009-18.

11 Ato de Concentração No. 08012.002259/2012-18.

12 Ato de Concentração No. 08700.004150/2012-59.

13 Law No. 12,529 of November 30, 2011.

14 See João G. rodAs, direito eConômiCo e soCiAl – AtuAlidAdes e reFlexões soBre direito ConCorrenCiAl, do Consumidor, do trABAlHo e triButário 160 (2012).

15 Ato de Concentração nº 08700.010790/2015-41.

16 CADE, relAtório AnuAl 1996 39 (1997).

17 Ato de Concentração nº 08700.003662/2018-93.

18 Amanda Athayde Linhares Martins Rivera et al., O improvável encontro do direito trabalhista com o direito antitruste, 24 revistA do iBrAC 65-93, 82 (2018).

19 See Processo Administrativo nº 08012.002812/2010-42.

20 See Processo Administrativo nº 08700.006386/2016-53.

21 Amanda Athayde Linhares Martins Rivera et al., O improvável encontro do direito trabalhista com o direito antitruste, 24 revistA do iBrAC 65-93, 84-85 (2018).

22 GiAndomeniCo mAJone, reGulAtinG europe 292 (2002).

23 Giandomenico Majone, The regulatory state and its legitimacy problems, 22 West eur. polit. 1–24, 14 (1999).

24 Giandomenico Majone, From the Positive to the Regulatory State: Causes and Consequences of Changes in the Mode of Governance, 17 J. puBliC poliCy 139, 191 (1997).

25 Paul Sabatier & Daniel Mazmanian, The Implementation of Public Policy: A Framework of Analysis, 8 poliCy stud. J. 538–560 (1980).

26 Paul A. Sabatier, Regulatory policy-making: Toward a framework of analysis, 17 nAt resour. J 415, 420 (1977).

27 Sabatier & Mazmanian, supra note 25 at 551.

28 Paul Sabatier & Daniel Mazmanian, The Conditions of Effective Implementation: A Guide to Accomplishing Policy Objectives, 5 poliCy AnAl. 481–504, 499 (1979).

29 Sabatier, supra note 26 at 552.

30 Ato de Concentração No. 08012.008378/2011-95.

31 See Recurso Ordinário em Ação Civil Pública No. 0001618 39.2012.5.01.0023 (TRT 1).

32 Ato de Concentração No. 08012. 005889/2010-74.

33 Gustavo Porto, CitrovitA FeCHA unidAde de mAtão, interior de são pAulo – eConomiA estAdão,,citrovita-fecha-unidade-de-matao-interior-de-sao-paulo,104529e (last visited Mar 14, 2015).

34 Livia Scocuglia, mpt proCessA CAde por não entreGAr doCumentos em inquérito Consultor JurídiCo, (last visited Mar 14, 2015).

35 Processo No. 0011506-28.2013.5.15.0081 (Vara do Trabalho de Matão).

36 BrAziliAn FederAl Constitution oF 1998, Art. 114.

37 See FrAnk FisCHer & GerAld J. miller, HAndBook oF puBliC poliCy AnAlysis: tHeory, politiCs, And metHods 223 (2006).

38 Julia Black, New Institutionalism and Naturalism in Socio-Legal Analysis: Institutionalist Approaches to Regulatory Decision Making, 19 lAW poliCy 51–93, 81 (1997).

39 Id. at 83.

40 Id. at 91.

41 See JuliA BlACk, rules And reGulAtors 18 (1997).

42 Id. at 31–32.

43 Id. at 34.

44 Neil MacCormick, Rhetoric and The Rule of Law 121 (2005).

45 See, e.g. MacCormick, supra note 33 at 101.