The political economy of the Workers’ Party governments in Brazil (2003-2014) – institutions, ideas and the main determinants of economic policy

The article intends to understand key features of the economic policy trajectory followed by the Workers’ Party governments in Brazil. The period was characterized by advances in social conditions, but the governments failed to advance key social reforms, reduce high interest rates or pursue a coherent industrial policy. The policies adopted led to a deep economic crisis, a key component of the subsequent political impasse. The article adopts a historical-institutionalist approach, establishing a dialogue with theories that deal with the role of international constraints, the institutionalization of the party system, economic ideas and the inﬂ uence of interest groups. The main contribution is to provide an interpretation which, by integrat-ing economic, political and institutional factors, signiﬁ cantly adds to the understanding of key economic policy decisions.


Introduction
In 2002, a former trade unionist, the leader of the Partido dos Trabalhadores (Workers' Party -PT), was elected Brazilian president. This was a landmark, given the elitist nature of the political system and the backgrounds of previous presidents. The Workers' Party (PT) remained in power for 14 years, ending in a controversial impeachment process in 2016. Economic performance was positive during most of the period and social conditions improved signifi cantly. The strengthening of income transfer policies, real increases in the minimum wage and positive labor market performance promoted a substantial reduction in poverty and extreme poverty.
It was also a period characterized by political stability. The two main parties, the PT and the Partido da Social Democracia Brasileira (PSDB), were signifi cantly strengthened and the respective presidents succeeded in negotiating their agenda in Congress. This gave the impression that Brazil had made substantial advances towards democratic consolidation.
However, those governments were also characterized by choices which triggered interesting political-economy issues. Firstly, the governments held a strong commitment to monetary stability, an agenda which drastically departed from previous PT ideals. Secondly, generally speaking, they were conspicuous by their low degree of defi ance to the elites, with no signifi cant advances in programs such as the promotion of a progressive taxation system and land, media and political reforms (Kaufmann, 2011;Hunter, 2011).
Another key puzzle was the incapacity to reduce interest rates, which are much higher than those prevalent in most other countries. This is highly disturbing when considering the negative impacts on fi scal accounts, economic performance and income distribution. High interest rates meant the transfer of a tremendous amount of resources to the richest groups.
Another interesting issue was the revival of industrial policy. Despite receiving a considerable amount of resources, there were inconsistencies in the policy and the governments were unable to produce signifi cant results in terms of competitiveness and innovation.
Finally, a careful investigation gains importance in light of the radical infl ection in Brazilian politics. The Brazilian economy has suffered, since 2015, the most severe crisis of its history. Accordingly, democracy suffered a big blow: traditional political parties lost votes and reputation, while an outsider, not committed to the preservation and advancement of democracy, became president.
The present article intends to promote the comprehension of key features of the economic and political trajectory. It embraces a historical institutionalist approach, as developed by Peter Hall (1986), an authoritative approach in political economy analyses. Hall (1986) shows the prevalence of national patterns of economic policy, critically infl uenced by institutional factors related to the organization of the state, the organization of the social groups and the channels of interaction between state and society. 1 In order to specify the main institutional factors analysed, the article deals with four key variables, widely emphasized in recent Latin American comparative studies: the constraints caused by the fi nancial capital fl ows (Campello, 2015); the degree of institutionalization of the party system (Flores Macias, 2012); the role of the different concepts of economic policy; and the impacts of interest groups, the fi nancial bourgeoisie and industrialists in particular (Cimini;Guimarães, 2018).
So, a key objective is to show the suitability of an interpretation which combines economic, political and institutional factors. This interpretation gains relevance when one considers that most prevalent analyses of the Brazilian economic-policy trajectory concentrate either on the infl uence of blocs of power and class segments (Pinto et al., 2019;Fonseca et al., 2019) or on the relationships between the executive and congress .
The analysis is based on books, articles and other bibliographical sources. Section 2 summarizes the main characteristics and constraints in three other Latin American center-left governments, providing parameters to investigate the Brazilian case. Section 3 makes a panoramic description of the main economic characteristics of the PT governments. Section 4 summarizes the main features of economic policy, while section 5 deals with the different views taken by Brazilian economists. Section 6 and 7 seek to explain key economic policy decisions. Finally, Section 8 summarizes the article's main contributions and deals with factors of Brazil's recent political trajectory.
2 Center-left governments in Argentina, Chile and Uruguay Cimini and Guimarães (2018) analyse the similarities, differences and limits faced by four center left governments in Latin America. A key objective is to understand why those governments did not go further towards promoting a more equal income distribution. Although substantial social advances were achieved, the governments, in general, failed to promote ambitious structural reforms. Something similar happened with industrial policy, which failed to promote consistent transformation.
Argentina was the country which comes nearest to a traditional left agenda. Economic policy was centered on the combination of expansionist monetary policy and devalued exchange rates. Fiscal policy found a critical pillar in the taxation of exports, while certain exports were restricted and their prices controlled. From 2003 to 2011, social expenditure grew 600% and minimum wage (in real terms) increased 200%. Programs for the unemployed and informal workers were strengthened, the pension system was re-nationalized and a strong stimulus was provided to the productive sector (Etchmendy;Garay, 2011). 2 The economic and social advances revealed to be highly dependent upon the favorable global context. From 2008, the fall in exports reduced the government's capacity to obtain fi scal resources and foreign exchange. Infl ation rose and farmers reinforced demonstrations against taxation on exports. Consequently, capital outfl ow increased and economic performance deteriorated; the heterodox experiment paid a high cost for challenging private fi nancial capital (Campello, 2015).
The Chilean center-left government also strengthened means-tested income transfer policies and measures aimed at facilitating access to social programs. Moderate advances took place in health and education and timid measures were adopted towards a more progressive tax system. The poverty rate decreased from 26% in 2000 to 7.9% in 2015, benefi ting from economic performance. A key characteristic, nevertheless, was the maintenance of orthodox macroeconomic policy, which included fi scal discipline, trade openness and privatization (Roberts, 2011).
Uruguay promoted substantial advances in social policies. The Plano de Equidad institutionalized support for low income families, reduced the qualifying age for access to non-contributory pension systems and tied income transfers to children's school attendance. A key advance was the adoption of a tax reform which signifi cantly reduced valued added tax on essential goods and expanded the share of the richest people in the tax effort (Lanzaro, 2011). From 2003 to 2014, the income of the poorest 40% of families increased 7.8% a year. From 2006 to 2014, poverty rate fell from 32.5% to 9.7% and extreme poverty from 2.6% to 0.3%. Like Chile, Uruguay's Frente Amplio (FA) maintained a commitment to orthodox macroeconomic policies. Cimini and Guimarães (2018) highlight the main variables responsible for those characteristics. The fi rst is the role played by fi nancial capital volatility, ready to punish governments which deviate from conventional measures. Campello (2015) shows how fi nancial capital veto power tends to constrain governments and narrow the differences between left and right agendas; only in periods of international high liquidity do governments have the autonomy to implement something different.
A second variable is the degree of party systems institutionalization, considered by Flores Macias (2012) the most important to explain differences among Latin American left governments. Institutionalization of party systems changes the politicians' calculus, reinforcing the necessity of fi nding agreement and consensus. Similarly, elections led the incumbents to moderate policies in the attempt to be closer to the median voter.
As Flores Macias (2012) shows, this was a key variable in explaining the Chilean parties' commitment to mainstream policies. Chile developed a very stable system, characterized by coalitions with deep roots in society. In face of the constraints inherited from the authoritarian regime, parties learned that negotiation was essential to approve legislation. This variable also helps us to understand the more radical policies in Argentina. The party system had been critically disrupted by the previous economic crisis. The lack of strong opponents gave the Kirchners autonomy to follow less orthodox policies (Kaufman, 2011).
A third variable is the timing of liberal reforms and the ideas embraced by policymakers. The institutionalization of liberal reforms in Chile and their results in terms of economic growth explain the support for conservative macroeconomic policies. In Argentina, by contrast, the disastrous results of the neoliberal experience gave strong support to heterodox policies. Uruguay, like Brazil, was typifi ed by an intermediate position and conciliation between different views (Cimini;Guimarães, 2018).
Finally, the fourth variable is the role played by specifi c social groups. In Chile, the Unión Demócrata Independiente, the main representative of business interests in Congress, had a strong infl uence in advancing a business agenda, while trade unions were weakened by Pinochet and by the neoliberal policies. In Uruguay, by contrast, trade unions were strengthened by the FA governments and were an important basis of support. In Argentina, the impact of the social crisis gave the Kirchners the support of unemployed and informal workers, the main trade unions and most of the Peronist Party.
In brief, those were the variables highlighted by the comparative analysis involving the three center-left governments. They offer parameters to investigate the Brazilian case.

The Workers' Party (PT) in government -general directions
The fi rst year of the PT in offi ce was characterized by uncertainty, mostly surrounding the party's previous ideas. The possibility of Lula's victory, in 2002, saw outfl ows of capital and high levels of instability. Thus, once in offi ce, a conservative macroeconomic orientation was adopted in the attempt to achieve market confi dence. The signifi cant exchange rate devaluation and the performance of the international economy stimulated economic recovery.
From 2003 to 2005, exports were responsible for 43.5% of demand growth, contributing to an increase of GDP growth from 1.1% in 2003 to 5.8% in 2004. From 2004 to 2008, exports grew 105%, pushed by the boom in commodities; average annual GDP growth was 4.8% (Prates et al., 2020). Furthermore, the entrance of foreign direct investment increased from US$ 15 billion in 2003 to US$ 116 billion in 2010 (Loureiro;Saad-Filho, 2018).
Another vector was the increase in consumption, stimulated by the income transfer policies, the minimum wage increase and a substantial expansion in credit. A key innovation was the introduction of pension deductible loans, a form of credit available to civil servants and pensioners, who use their monthly income as a guarantee. In addition, credit for consumption increased substantially. Thus, credit as a share of GDP grew from 23.8% in 2002 to 55.8% in 2014 (Corsi, 2016).
Economic growth was also favored by the social programs, which brought millions of people into the consumer market. Previous programs were merged into the Programa Bolsa Família, a cash transfer program which benefi ted 11 million families. Other programs, which focused on the disabled and elderly, were expanded, stimulating the economy in rural areas and small municipalities. From 1995 to 2011, people who benefi ted from social programs increased from 14.5 to 24.4 million; social expenditure grew 200% in real terms. The number of people living in poverty fell from 60 million in 2003 to 30 million in 2012 and the Gini index decreased from 0.6 to 0.53 (Loureiro;Saad-Filho, 2018).
Another measure was the deliberate decision to strengthen the minimum wage. A large proportion of the Brazilian population earned minimum wage or had their earnings indexed to it. 3 In addition, economic recovery and an improvement in fi scal conditions allowed the government to increase investment in the housing sector and in infrastructure. Within the context of economic recovery, private investment also recovered (Loureiro;Saad-Filho, 2018;Corsi, 2016).
Consequently, from 2004 to 2010 GDP grew at an average rate of 4.4% a year. Household consumption grew 5.3% a year, investment in the housing industry 5.8% and investment in fi xed capital 8% a year. 4 Household consumption played the key role, representing 60% of the aggregate demand increase from 2006 to 2011, while private investment was responsible for a 23.3% increase (Serrano;Suma, 2015;Corsi, 2016).
In brief, those policies sustained a virtuous cycle marked by economic growth and reductions in poverty and inequality. Millions of people joined the consumer market, increasing the PT's support base. 5 Additionally, conservative macroeconomic policies benefi ted the fi nancial and economic elites.
Nevertheless, economic expansion suffered a blow with the outbreak of the international fi nancial crisis. In 2009, the GDP fell, followed by a substantial recovery in 2010, pushed by China's imports and fi scal stimulus. The period 2011-2014 was marked by a signifi cant deceleration.
A key factor was the exhaustion of credit expansion, given the high levels of indebtedness. In 2005, personal debt represented on average an 18% share of annual disposable income; by 2014, this had increased to 46% (Serrano;Summa, 2015). This was critically aggravated by the exorbitantly high interest rates charged by the banks. 6 It is estimated that 22% of household income in 2013 was used for debt payment (Corsi, 2016).
Consequently, from 2011 to 2014, the growth in household consumption decelerated to 3.1% a year, investment increase dropped to 1.8% and investment in the housing industry to 2.8% a year. There was also a signifi cant deceleration in government investment and in exports, where the annual growth rate fell to 1.6% a year (Serrano;Summa, 2015).
Finally, the period was characterized by a substantial increase in wages and employment. From 2003 to 2014, unemployment fell from 12.4% to 4.8%, while wages as a share of GDP increased from 36.7% in 2003 to 46.3% in 2015 (Marquetti et al., 2019). Economic growth was above the average of previous periods, but industry's share of GDP fell from 16.9% in 2003 to 11.4% in 2015 (Prates et al., 2020).

Economic policy in the period -main vectors
A key policy feature was the commitment to monetary stabilization, preserved in different degrees across the entire period. The PT's fi rst steps in offi ce were characterized by the reinforcement of orthodoxy. Fiscal surplus and interest rates were increased and a partial reform of the pension system was adopted. The market reacted very positively (Campello, 2015;Hunter, 2011).
The economic agenda at that moment included support for central bank autonomy, while an ambitious microeconomic agenda included revision of bankruptcy laws, attempts to reduce legal uncertainty and an improvement in the regulatory environment (Giambiagi, 2016). Once macroeco-nomic stability was considered to be under control, measures to stimulate demand were adopted.
An important infl ection took place in 2006, following the exit of the economic minister Antonio Palocci, involved in a political scandal. Palocci had played a key role in convincing the PT to adopt an orthodox economic policy orientation. The new minister, Guido Mantega, intensifi ed the measures to stimulate the economy. From 2006 to 2010, public investment increased 17% a year (in real terms), while tax exemptions were increased in sectors such as the building, heavy manufacturing and high technology industries. From 2008 on, tax reductions were extended to the automotive, durable consumer goods, equipment, furniture and food sectors (Carvalho, 2018).
An important direction was the reduction in interest rates, which fell from an average rate of 20% in 2003/2004 to 9.83% in 2009. The reduction took place in a context characterized fi rstly by abundant international liquidity and later by the effects of the international crisis. Nevertheless, the commitment to preserve primary surpluses was maintained, praised by some economists as a sign of fi scal responsibility, but criticized by others for restricting the possibilities of economic growth (Corsi, 2016).
The Rousseff government (2011-2014) was characterized by frequent changes in the way macroeconomic policy was conducted. The fi rst year was marked by reinforcement of orthodoxy: interest rates rose from 7.5% in 2010 to 13.5% in August 2011 (Serrano;Summa, 2015). Credit was restricted, compulsory deposits increased and public expenditure contained.
In 2012, nevertheless, a strong offensive was launched to reduce interest rates and stimulate the economy. Interest rates fell to 7.25% in April 2013, 1% in real terms (Corsi, 2016). Public banks were then employed to reduce the exorbitant spread rates. 7 Measures which included control over foreign fi nancial operations were employed to devalue the exchange rate. From the second semester of 2011 to 2013, the Brazilian Real lost 20% of its value.
The "new economic matrix" was also characterized by a strong increase in subsidized credit and other incentives. Tax exemptions were intensifi ed and cuts in fi rms' payroll taxes were introduced aimed at stimulating investment. The National Development Bank (BNDES) loans, at highly sub-sidized rates, were reinforced. Carvalho (2018) estimates that from 2009 to 2014, R$ 412 billion (at 2017 prices -approximately US$ 127 billion) was transferred from the treasury to the BNDES.
In addition, measures were adopted to reduce electricity prices. Reduction in fees and a controversial renewal of concessions resulted in an 18% tariff reduction for families and in a 32% reduction for industry. Public purchasing policies, including an aggressive policy by Petrobrás (Brazil´s oil company), were employed to stimulate industrial niches, while protectionist measures were intensifi ed (Singer, 2015).
From April 2013, nevertheless, the Central Bank reversed the policy and resumed a new cycle of interest rate rises. Nominal interest rates rose to 11.75% in December 2014, an increase to 5.3% in real terms. Regulations were also removed from fi nancial capital fl ows. The tax exemption policy, however, was expanded to other sectors. To keep infl ation under control, the government increased controls on utilities and fuel prices. In 2014, fi scal policy was very expansive due to electoral reasons.
The period was also marked by the revival of industrial policy. An initial effort was made to create councils and increase state capacity. Lula joined business', trade unions' and civil society's representatives to elaborate a development agenda. The National Council of Industrial Development (CNDI) appeared initially to be very effective. From 2003 to 2006, horizontal industrial policies were prioritized and a landmark policy was the passing of the Innovation Law (Guimarães, 2021).
From 2007, nevertheless, industrial policy signifi cantly changed direction. BNDES channeled billions of dollars to stimulate a wide range of sectors, protectionism and public purchasing policies were intensifi ed and "national champions" were promoted. Despite positive initiatives in stimulating R&D investment, the policy presented many inconsistencies: precise objectives were not established, almost every sector was contemplated for assistance and the policy suffered from coordination problems (Guimarães, 2021).
In brief, the period was marked by the alternation between initiatives to stimulate demand and infl ation control measures. The fi rst years were marked by emphases on microeconomic reforms, improvement in regulation and horizontal industrial policies. An infl ection took place in 2007, marked by the embracing of developmentalist positions. The period was also marked by the lack of key structural reforms (administrative, tax and political reforms, for example) or any articulated program to increase tax progressiveness.

Orthodoxy, heterodoxy and different interpretations
Another hallmark of the PT governments was the interaction between economists with highly distinctive views. A group of orthodox economists occupied the Finance Ministry in Lula's fi rst term. They prioritized the promotion of an adequate business environment and a conservative macroeconomic policy. According to Lisboa and Pessôa (2016), such an emphasis was justifi ed by the belief that development is a result of an increase in productivity achieved through improvement in workers' skills, in systemic productivity and in the quality of the institutional setting.
According to Lisboa and Pessôa (2016), a key turning point was the appointment of a heterodox fi nance minister in 2006, which resulted in an interventionist approach and an immense waste of resources. The huge transfer of resources and subsidies, in a context of full employment, had, according to them, little impact upon investment and resulted in infl ation and balance of payments pressures. Lisboa and Pessôa (2016) blame the heterodox policies adopted under the Rousseff's government, which included insuffi cient attention paid to fi scal accounts, for the profound economic crisis from 2014 on. According to Lisboa and Pessôa (2016), subsidized resources led to investments in sectors where demand had stagnated. The automotive sector, for example, was working with an idle capacity of 50%, whilst the naval industry faced huge fi nancial diffi culties. 8 In brief, the main problem was, according to this view, the change in economic policy orientation and the interruption in the necessary reforms. Lisboa and Pessôa (2016) associate the diffi culties to the inadequacies of heterodox theories which, according to them, place excessive emphases on the state's promotion of industry and on the capacity of fi scal policy to induce investments. 9 Heterodox economists have a different view. Paula and Jabbour (2016) strongly criticize the attempts to attribute economic crisis to heterodox economic ideas. Paula and Jabbour (2016) agree that productivity is critical for growth, but consider that it is necessary to go beyond property rights and the business environment. Productivity tends to be favored by industrial growth, which permits the incorporation of machines and new techniques. Another point is to emphasize the relevance of exchange rate policy for economic growth, well-illustrated by previous economic miracles (Germany, Japan, South Korea and China).
Paula and Jabbour (2016) also criticize orthodox economists for failing to recognize the role played by the state in successful development experiences, including the promotion of competitive international enterprises in East Asia. They defend the conciliation between active state intervention with the promotion of macroeconomic stability.
Paula and Jabbour (2016) accept Rousseff's mistakes, but argue that they were less a consequence of the heterodox emphases than of wrong policies. There was excessive emphasis on tax exemptions, which had a low multiplier effect (Carvalho, 2018). In addition, they add that the ensuing economic crisis was critically aggravated by political scandals. 10 Another set of heterodox economists' criticisms referred to the infl ation target policies. They argued that infl ation during most of the period was not related to demand pressures, but rather to the effects of exchange rate devaluations and administered prices (Serrano;Summa, 2015). Infl ationary pressures were also a result of the fi rms' reluctance to increase investment, opting for increasing the degree of capacity utilization. The problem with the automatic increase in interest rates was that it frustrated economic growth and inhibited decisions to increase investment, perpetuating a vicious cycle (Kupfer, 2019).
Thus, infl ation targets' policy was considered contradictory, since monetary policy tended to weaken the effects brought about by economic recovery, keeping the economy on a low economic growth trajectory (Morais; Saad-Filho, 2011). The policy worked when international conditions were favorable due to the positive impact which overvalued exchange rates had upon infl ation. The opposite, nevertheless, occurred when the exchange rate was devalued, leading to interest rate rises. Thus, the coun-10 Petrobrás investments, which responded for 7.9% of total investment in 2013, were cut by one third from 2014 to 2015 (Marquetti et al., 2019). try was condemned to stagnation, while high interest rates provided gains to fi nancial investors, at the cost of the general population.
In brief, this section highlighted the very different concepts embraced by Brazilian economists, providing a view on the lack of consensus regarding the key directions to promote the country's development.

Determinants of Economic Policy
This section addresses the general economic policy trends that were followed. This requires an understanding of fi ve main issues: the adoption of a conservative macroeconomic policy; the relatively low degree of defi ance to the elites; the inability to reduce interest rates; the policies' oscillation under Rousseff and the industrialists' position towards Rousseff's government. Economic ideas are part of the answer, intertwined with other variables.
Several factors help to understand the option for a conservative macroeconomic policy. Brazil had a previous trajectory characterized by high and persistent infl ation. Achieving stability was a hard-fought process, which left a legacy of a prevalent inertial component of infl ation, where price increases tend to be transferred to other periods. This all contributed to the prioritization of maintaining infl ation at low levels.
A second component referred to the constraints created by the vulnerability to capital fl ows. These are most severe in countries with low savings and a high dependence upon foreign fi nancial capital (Campello, 2015). The PT had a glimpse of such potential instability in 2002, which convinced them to embrace a market-friendly position.
A third component is related to the fi nancial groups' capacity to advance their own agenda. According to Teixeira and Pinto (2012) and Fonseca et al. (2019), the adoption of a policy characterised by infl ation targets could be interpreted as an indication of the structural power of fi nancial capital and the PT's surrender to a situation of class hegemony.
An inter-related point is the relatively low degree of defi ance to the economic elites. Lula was very popular and achieved a kind of Bonapartist autonomy, which increased the potential to adopt ambitious policies. Important issues such as land, tax and political reforms, nevertheless, did not make any signifi cant advances, while previous liberal directions were not reversed (Hunter, 2011).
According to Flores Macias (2012), this is in large part explained by the degree of institutionalization of the party system, forcing the party in offi ce into negotiations to improve the chances of getting its agenda approved. In such a situation, the leader in power is induced to adopt a moderate position as a strategy to increase chances of reelection. As the PT already had the support of labor and peasant movements, the challenge was to gain support from middle and higher classes segments (Kaufman, 2011). Flores Macias (2012) emphasizes how PT positions changed substantially, recognizing the importance of compromise to gain approval for the necessary measures. 11 Finally, Lula revealed himself to be a pragmatic politician, who perceived that a responsible macroeconomic program increased the chances of electoral victory.
This moderation is also explained by the characteristics of the political system, marked by the combination of a presidentialist regime with proportional representation elections (Schneider, 2013). During the period, the PT did not have more than 18% of seats in Congress and had to form coalitions with heterogeneous parties. This increased the veto points and the necessity to fi nd minimal common denominators (Flores Macias, 2012).
Another factor is that the PT was able to achieve a trajectory of growth and redistribution without pushing for controversial reforms. As evidenced, life for a large part of the population improved. In the short term, it worked, having, nevertheless, profound consequences later on (Loureiro; Saad-Filho, 2018).
A third puzzle is to understand the persistence of very high interest rates. Teixeira and Pinto (2012) emphasize the role played by the banks and fi nancial capital, key components of a "bloc of power" whose capacity of infl uence had been strengthened by the confi gurations of liberal capitalism. Although Lula tried to attract industrialists, he was not able to break free from the infl uence of fi nancial groups.
Another explanation, associated with orthodox economists, explains high interest rates as a consequence of the challenges to keep infl ation under control during expansionist fi scal policy. 12 This explanation criticizes aspects of the Brazilian fi nancial system, arguing that a large part of credit was controlled by state agencies, implying that larger increases in interest rates were necessary to reduce demand and keep infl ation under control (Pessôa, 2019).
The point to emphasize is that the presence of different views explains the lack of an orchestrated policy to reduce interest rates. It is important to stress that a successful offensive to reduce interest rates, even if it resulted in less economic expansion in the short term, could have reduced a key source of spending and released resources to be invested in other areas. 13 It is relevant to highlight the diffi culties faced by Rousseff when trying to lower interest rates. Resende and Terra (2020) believe that such difficulties were explained by the incorrect manner in which Rousseff implemented the policy, failing to convince economic agents that interest rates would indeed fall. A point to emphasize is that the Temer government (2016)(2017)(2018), much more closely aligned to fi nancial capital interests, was able to implement a substantial cut in interest rates. The economic recession was a contributor, but it is pertinent to ask why something in a similar direction had not been attempted previously.
Another interesting aspect is the lack of a consistent coalition engaged in interest rate reductions, which is somewhat disturbing in light of the PT's industrialization project. A convincing hypothesis is that the industrialists were also fi nancial investors and in the fi nancial sector had a safe haven of profi ts.
Fourthly, Rousseff's government was characterized by oscillations in economic policy direction. Firstly, she increased interest rates to curb infl ation. Secondly, an offensive to reduce interest rates and banking spread and to devalue the exchange rate was introduced, complemented by other heterodox policies. Finally, she abandoned this strategy and started a new wave of interest rate rises. The main puzzle is to understand the second and third decisions.
The decision to embrace a heterodox orientation was, according to Fonseca et al. (2019) and Prates et al. (2020), critically infl uenced by changes in the international context, which reduced the appeal of previous policies. In addition, the process of growth centered on credit expansion had reached its limit, demanding other forms of demand stimulation. Accord-13 From 2007 to 2011, public sector expenses with interest rates accounted for something between 5% and 6% of GDP (Giambiagi, 2016) . ing to Bastos (2017), Rouseff had the deliberate objective of confronting the power of fi nancial capital, intending to promote private and public investment and, through exchange rate devaluation, strengthening exports. Singer (2015) identifi es in 2011 the emergence of an "alliance for production" aimed at stimulating investment in industry and in the productive sector. According to Singer (2015), the Industrial Federation of São Paulo (FIESP) associated with trade unions and signed a document which espoused a coordinated action for industrial revival. This mobilization resulted in the package which included an increase in BNDES loans, protectionism, reduction in energy prices and the intensifi cation of public purchasing policies.
Singer's interpretation has received much criticism. Fonseca et al. (2019) and Prates et al. (2020) recognize the rupture with orthodoxy, but disagree that the measures represent a developmentalist agenda, since Rousseff`s government lacked both a profound diagnosis and the necessary policies to strengthen industry. Carvalho (2018) sees Rousseff policies as much more centered upon subsidies and rent seeking than on articulated policies to tackle the challenges faced by industry. Carvalho (2018) strongly criticizes the emphasis placed upon incentives and tax exemptions, giving the limited impact upon aggregate demand. The cost of tax exemption policies rose from R$ 140 billion in 2010 to R$ 250 billion in 2014, without a signifi cant impact upon investment. Carvalho strongly disagrees with Singer's view that such policies could have worked.
In 2013, as argued, there was another change of direction, leading to a new wave of interest rate increases. According to the interpretation shared by an infl uential portion of the press, this represented evidence that infl ation was rising, putting Rousseff's reelection chances at risk. Singer (2015) has a different interpretation, arguing that Rousseff had opened too many fronts and became isolated. Rousseff outraged powerful interests by attempting to reduce banking spreads, controlling profi ts in infrastructure concessions and introducing arbitrary changes in the electricity sector. Bastos (2017) adds an important point: Rousseff failed to communicate the positive aspects of her strategy in terms of income distribution, stimulus to the economy and government's capacity of action. By not doing so, she allowed the fi nancial capital counter offensive, which blamed her gov-ernment for populist policies considered irresponsible in technical terms. According to Bastos, this led Rousseff to increase interest rates in an attempt to attract the support of fi nancial capital.
A key point is to understand why industrial business people did not provide the necessary support. Although their main demands were contemplated, they failed to support Rousseff´s policies and did not mobilize against the new offensive to increase interest rates. It is also interesting to note that, despite innumerous incentives granted, the investment rate decelerated and industrial investment fell (Bastos, 2017) Several reasons help to explain the investment performance. The fi rst is the increase in uncertainty, amplifi ed by the vicissitudes of the international context. 14 A second factor is the decline in profi t rates. According to Marquetti et al. (2019), a decline in profi ts began in 2008 and accelerated after 2011, a result of economic deceleration, growth in wages and a stagnation in productivity. 15 According to Bastos (2017), profi ts were also negatively affected by the impacts of exchange rate devaluation on input prices and by increased foreign competition, which negatively affected industrial exports.
According to Singer (2015), the industrialists' position may be explained by a "class issue". Previous policies, state intervention and full employment had increased labor power. The number of strikes increased from 873 in 2012 to more than 2000 in 2013. Additionally, Rousseff demonstrated a solid determination to reject any policy aimed at reducing wages and labor rights. In other words, Rousseff was not considered trustworthy by part of the business class. A similar interpretation is shared by Pinto et al. (2019), who see in the reduction of labor costs the business sector's main priority.
A key point in Singer's interpretation is widely shared: the divisions within the business class and its incapacity to support a national project. The Brazilian business class tend to pursue individualist interests, connected, in many instances, to those of the international elites (Cardoso;Faletto, 1975). There is little willingness to engage in coalitions aimed at improving the educational system or increasing the capacity for innovation (Schneider, 2013).

Interpreting Rousseff's impasses and the main implications
A fi rst point to emphasize is the impact of international constraints: favorable international contexts increase resources and make the accommodation of interests easier, while an unfavorable context reduces the economic policy options (Campello, 2015). Nevertheless, the Brazilian trajectory deviates from Campello's conclusions in an important way: an orthodox orientation was adopted in a favorable international scenario, changing to heterodoxy when international conditions deteriorated. As Fonseca et al. (2019) argue, the international changes and the exhaustion of credit expansion affected the appeal of previous policies.
Secondly, the direction given to economic policy seems to have been highly infl uenced by the ideas. Rousseff's team strongly opposed orthodox policies and viewed state intervention as the way to promote development. Considering such an orientation, there is little sense to argue that Rousseff program was not developmentalist, as done by Fonseca et al. (2019). The failure to adopt consistent measures was not evidence of the lack of such an orientation.
As argued, industrial policy had been part of the PT agenda since 2003. The policy resources were substantially increased after 2007. The problem, as reported, was the policy's innumerous inconsistencies, contemplating almost every sector and failing to defi ne precise targets in terms of productivity and competitiveness (Guimarães, 2021). Similarly, the policy demonstrated little concern regarding state capacity issues (Nunes, 2020).
Another diffi culty was the blurring of industrial and macroeconomic objectives. The Production Development Plan, adopted in 2007, aimed to support as many sectors as possible, as a way of stimulating investment in response to the international crisis (Guimarães, 2021). In doing this, it failed to prioritize sectors with the best chance of success. The idea was to stimulate the animal spirit amongst businessmen, which helps to understand Rousseff's emphasis on tax exemptions and incentives.
A third point deals with the deceleration in investment. The fall in profi t rates is indeed relevant, but it also tends to be infl uenced by other factors, such as commodities and share prices (Carvalho, 2018). In addition, profi t rates are also infl uenced by investment decisions (Kalecki, 1977). Thus, a reduction in business willingness to invest may also have played a key role in explaining the fall in profi ts.
A fourth and decisive point is the relationship between business and the PT. Although a cooperative relationship was developed under Lula and important business demands were contemplated by Rousseff, it is argued that business always maintained a certain mistrust and did not consider the PT as a reliable option. The PT supported an active role for the state in the economy and refused to adopt measures aimed at reducing wages and labor rights. In addition, part of the business class feared the potential for radicalization. According to Bastos (2017), this explains the alliance of business class segments in 2015 to support Rousseff's impeachment.
Nevertheless, a very different hypothesis could be raised, highlighting the opposition against a party which had spent so many years in offi ce and failed to deal with important items of the business agenda, such as the high costs of "doing business" and stagnated productivity. It is a relevant hypothesis to be tested, although it is more relevant to understand the opposition against Rousseff's policies than the decision to support the impeachment. This leads to a key issue: the need to better understand the main business demands and their specifi c positions regarding the PT governments. Although much has been said about business interests, very little was based on empirical research.
Rousseff did indeed face substantial constraints. Commodity prices fell, profi ts were in decline, the economy was close to full employment and there were fi scal constraints. Another type of constraint came from the high dependence on commodities and industry's low capacity to penetrate foreign markets. In addition to the constraints, signifi cant mistakes in economic policy need to be emphasized, such as industrial policy failures, an excessive emphasis on tax exemptions and the strategy's failure to reduce interest rates.
In addition, it is necessary to add Rousseff's limited ability to negotiate and compromise, which leads to the characteristics of her coalition. As Bertholini and Pereira (2017) demonstrate, it was a highly heterogeneous coalition in ideological terms and very costly to manage. Another point relates to the government's fi nancial diffi culties and diminishing capacity to implement the congress members' amendments. This combines with Rousseff's tendency towards centralization and resistance to delegating functions. Guimarães et al. (2019) describe Rousseff as a subversive type, who insisted on changing the rules of the political game.
Rousseff's fi rst term ended in 2014 in a very diffi cult and paradox situation. Her center-left government distributed hundreds of US$ billions to the business sector, without clear compensation for society at large. The amount was equivalent to 10 times the most popular income transfer program. In addition, the economy faced very negative prospects and in 2014 GDP growth was only 0.5%. Rousseff's disorientation is very well illustrated by the radical change in economic direction adopted in 2015. After a very polarized election, she embraced the program of her main rival, in an attempt to replicate Lula's 2003 policies. The context, nevertheless, was very different and the results were disastrous, in both economic and political terms (Bastos, 2017).

Final Considerations
This article offers an interpretation of an important period in Brazilian history, characterized by signifi cant social advances, but which ended in an economic and political crisis which in turn created the conditions for an extraordinary retreat of the democratic process. Our main objective was to explore the complex range of factors which infl uenced the economic trajectory.
A critical role was played by international constraints, which decisively infl uenced the available choices. Political interests also played a key role: every government needs political support and certain interests were better positioned to infl uence policy making. However, the infl uence of interest groups is molded by political institutions, including business organization and the channels of interaction between the state and the social groups. As Flores Macias (2012) has shown, a key role is played by the institutionalization of the party system. The features of the political system also help to understand the diffi culties in promoting radical changes.
Ideas also play a highly important role. A key contribution of this article was to emphasize the deep ideological differences between the groups of economists in charge of the economy. This is decisive to understand the lack of coherent action to tackle certain issues.
The role of these variables needs to be complemented, as emphasized, by structural factors related to the organization of the Brazilian economy, which explains why business people fi nd so few incentives to engage in collective action to improve the education and innovation systems. This is critical to understand the lack of support for a national project aimed at strengthening industrial and technological capacity (Schneider, 2013).
The present article carefully reviewed insightful interpretations of the recent political and economic trajectory. Its main contribution is to defend the suitability of an approach which integrates political, economic and institutional factors (Hall, 1986). As argued, interests are a highly important variable, but they are mediated by the institutional features which characterized the organization of the state, the economic groups and the political system. Accordingly, constraints are important, but not deterministic, and institutional factors are key to understanding the differences in national responses.
Our point is well illustrated by commenting upon one of those interpretations. Pinto et al. (2019) criticize interpretations which argue that the problems originated from economic policy mistakes. They contend that it is a mistake to assume that the state has autonomy in relation to economic classes. For Pinto et al. (2019), states are in general constrained by class fractions and blocs of power directed by dominant groups. According to Pinto et al. (2019), party systems are on the surface, being necessary to investigate the dominant sectors within the bloc of power.
A key contribution of this article is to highlight the limits of such an interpretation. The new confi gurations of capitalism strengthened the infl uence of certain groups, but this does not suggest such a deterministic explanation. Countries respond differently to similar economic challenges and some countries obtain better results than others (Hall, 1986).
In brief, to say that a government acts in a given way because it is controlled by a bloc of power is not saying much. A vivid and richer policymaking approach, able to deal with national differences and the vicissitudes of specifi c governments, needs a broader approach which addresses institutional variables present in the national systems of political economy.
Another conclusion is the need to go beyond the impacts of macroeconomic policies. We agree that infl ation target policies constrained investments and economic growth. Nevertheless, heterodox policies are not by and of themselves a panacea, as illutrated by the Argentinean experience.
The best option is a developmentalist policy conciliated with low infl ation and export stimulus, but the big challenge is how to achieve it. A favorable development strategy also needs to consider the state capacity requisites and the institutional factors inherent to the organisation of the economic system. The recognition of institutional rigidites is crucial for the design of a long term strategy with a chance of success.
The article concludes with some considerations related to Brazil's ensuing political trajectory. Loureiro and Saad-Filho (2018) emphasize PT failure to mobilize the workers and social segments and to reform the media and the rules of electoral funding. According to the authors, the PT committed a mistake in avoiding confrontation and believing that the elites would accept their policies.
Francisco Panizza (2013) has a different view. He emphasized that leftist governments in Latin America should be prepared to lose power as part of the democratic process. They should refuse to stay in power at all costs, focusing on their legacy and political programs as the shorter way to return to power. Panizza (2013) did not underestimate the obstacles faced by the left, but emphasized that the dispute must take place within the democratic process.
In Brazil, during the 2014 elections, the opposition, supported by part of the media, made powerful attacks accusing the PT of involvement in political scandals. In a similar vein, Rousseff employed a huge marketing apparatus to attack adversaries and defend her policies. The election was decided by a narrow margin and the defeated candidate refused to accept Rousseff's victory. Once in offi ce, Rousseff radically changed her discourse and adopted the policies of the main rival.
These events and what followed demonstrate that democracy was not yet consolidated. There is evidence that opposition forces articulated a movement to remove the PT from power (Fausto, 2019, p. 73). Despite Rousseff's irregularities, her impeachment process was exaggerated. So, there is little doubt regarding the intentions of part of the political and economic elite.
However, the dynamics that weakened Brazilian democracy started earlier. As Levitsky and Ziblatt (2018) argue, informal rules are key to the preservation of democracy. A key problem comes when politicians treat rivals as enemies or encourage the use of offensive terms to describe them. Another risk is when leftist parties describe democracy as a bourgeoisie anachronism. In addition, it is also pathological when one party contests the legitimacy of other candidates or fails to recognize electoral defeat. Finally, democracy also loses when the executive attempts to become too independent or when Congress opts for an impeachment process without clear motives.
The PT never explicitly threatened the balance among powers or tried to control the judiciary or federal policy. Lula could have attempted to run for a third term but refused to do so. However, criticisms may be made against certain positions which threatened the "informal rules". As Pessôa (2019, p. 34) argues, the PT built a biased narrative against the previous Cardoso government. According to Pessoa (2019), the criticisms were based on fragile accusations, which tried to paint mere disagreements as crimes. This was ironic, as Cardoso had played a role in producing certain conditions for the subsequent economic recovery.
The fact is that the PT espoused certain authoritarian features and that a large part of its supporters interpreted the political process as a fi ght of good against evil. In offi ce, the Party adopted a range of illegal practices: resources from state enterprises were used to buy off Congress members and fi nance electoral campaigns. The public machine was widely employed to benefi t the party and, in several cases, also private interests (Fausto, 2019, p. 51).
According to Fausto (2019, p. 52), segments of the PT say that this was an inevitable cost for improving social programs. Fausto (2019) strongly disagrees: the Brazilian political system is indeed highly corrupt, but the PT emerged to defeat it and systemic corruption did not absolve it of its mistakes. Such practices severely harmed the PT's image. They permitted the resurgence of the extreme right with a discourse of skepticism, which was decisive in attacking institutions and gaining power. As Fausto (2019) concludes, if the party had resisted those practices, it would have faced obstacles to govern, but it would have been less vulnerable to attacks from the elites.
We are aware of the asymmetries inherent to democracy and to capitalism, which are severe in Latin America. It is not our intention to claim that both sides committed irregularities and therefore to legitimize the opposition's strategies. Our point is that democracy was weakened and that the PT had a role in this. Despite the asymmetries, the best form of achieving political and social gains is through the democratic process and the best strategy is to work towards strengthening it. Democracy has the idea of equality at its core, which should be mobilized to tackle the deep inequalities inherent to capitalism (Panizza, 2013;Fausto, 2019).