Business Power in post-authoritarian Chile: explaining regulatory continuity in the pension system, 1990-1994

Es difícilmente controversial sostener que las elites económicas en América Latina tienen una gran y desproporcionada influencia política. En este contexto, Chile ofrece un estudio de caso único para el análisis del poder empresarial. La Concertación, la coalición de centro-izquierda que gobernó Chile entre 1990 y 2010, fue electa con un programa de gobierno que prometía la creación de un sistema de seguridad social con elementos de solidaridad. Sin embargo, esta coalición terminó finalmente consolidando el sistema privado de pensiones impuesto en 1981 por la dictadura militar. El presente trabajo aborda esta paradoja entrándose en el debate sobre pensiones desarrollado durante la primera administración de la Concertación (1990-1994). En resumen, esta investigación demuestra que, a pesar del nuevo contexto democrático, la elite económica continuó controlando el proceso de toma de decisiones por medio de múltiples mecanismos.


ABSTRACT
It is hardly controversial to contend that economic elites wield a disproportionate political influence in Latin America. In this context, one critical area which we require a better understanding of is the influence of business on the development of social policies. Chile offers a fascinating puzzle in this regard: why did a centre-left coalition, elected on a platform of "creating a social security system with more elements of solidarity" end up passing reforms that consolidated the private, individually managed pension scheme imposed by the military regime (1973)(1974)(1975)(1976)(1977)(1978)(1979)(1980)(1981)(1982)(1983)(1984)(1985)(1986)(1987)(1988)(1989)(1990)? This paper attempts to shed light on this question, focusing on the debate around the old-age pension system during the first Concertación government, the Aylwin Administration (1990)(1991)(1992)(1993)(1994). The analysis is based on processtracing. Primary sources for the analysis include interviews with key informants, and congressional and press records. In summary, this work shows that despite the new democratic setting, the Chilean economic elite continued to control the policy process by means of multiple mechanisms. Keywords: economic elite; business power; Latin America; Chile; pensions.

INTRODUCTION
It is hardly controversial to contend that economic elites wield a disproportionate political influence in Latin America. Large, diversified domestic business groups are politically connected and economically dominant in the region (Cárdenas, 2016;Schneider, 2013). In this context, one critical area which we require a better understanding of is the influence of business on the development of social policies. Most analyses done so far on pension and health reforms tend to consider the role of business as incidental in comparison to other variables (e. g. Borzutzky, 2002;Brooks, 2009;Castiglioni, 2018;Garay, 2016;Martínez Franzoni and Sánchez-Ancochea, 2016;Pribble, 2013). This gap in the literature is not trivial since the last four decades have witnessed a growing business involvement in the arena of social policies (Orenstein, 2008).
Chile offers, in this context, an instructive puzzle for the study of business power. In March 1990, a centre-left coalition -the Concertación-rose to power marking the end of seventeen years of civic-military dictatorship (1973)(1974)(1975)(1976)(1977)(1978)(1979)(1980)(1981)(1982)(1983)(1984)(1985)(1986)(1987)(1988)(1989)(1990). The newly elected coalition promised "to create a social security system with more elements of solidarity" (CPD, 1989: 37). While the coalition planned to preserve the foundations of the pension system imposed in 1981 -which required workers to save for their retirement in accounts managed by private providers-, it promised to implement significant equityenhancing measures. Yet no significant progressive reform was passed in the subsequent decade; in fact, the few changes approved closely resembled business preferences (e. g. El Mercurio, 1989). Why did this happen? What were the mechanisms that the economic elite used to prevail in the debate on the old-age pension system?
In an attempt to answer such questions, the literature on business power has mainly focused on the classical categories of instrumental power (also known as political power) and structural power (or investment power). Through the lens of the framework developed by Fairfield (2015aFairfield ( , 2015b, this research indeed identifies significant and durable sources of power for the Chilean economic elite. One of the main contentions of this work, however, is that such variables are not enough to explain the outcome mentioned above. The Concertación not only abandoned its equity-enhancing agenda, but it also ended up supporting the consolidation of the market-oriented pension scheme bequeathed by the authoritarian regime. In order to explain such "reversal of fortunes", the analysis needs to take into account additional factors that may act as sources of business power, or at least might modulate it. The evidence gathered for this research point toward two defining factors that restricted the space for equity-enhancing reforms: i) ideology, and ii) collective memory (i. e. fear of authoritarian reversal).
In this context, this paper makes analytical, methodological and empirical contributions to the literature on business power in Latin America. Analytically, I identify new sources of business power, highlighting the role of ideology and collective memory in shaping contemporary policy debates. Methodologically, following recent insights from Bril-Mascarenhas and Maillet (2019) and others, I demonstrate the need to consider long-term dynamics to explain business power even over a short period of time. For instance, in order to capture the role of ideology in this case study, it was necessary to carry the research inquiry back to the 1960s, when the Chilean economic elite began to invest in the development and spread of neoliberalism. Empirically, I explore a period in Chile's pension history that has not received much attention in the literature.
This study will focus on the debate around the old-age pension system during the first Concertación government, the Aylwin Administration (1990)(1991)(1992)(1993)(1994). This debate concluded with the passage of the "Reform to the Capital Market" in March 1994. This research aims, in this context, to establish systematic links between the independent variable (i. e. business power) and the observed outcome (i. e. regulatory continuity). Process-tracing is especially suited to discovering and understanding the causal effect that links an independent variable under study to a particular outcome (Collier, 2011;Waldner, 2015). The empirical evidence for conducting the process-tracing comes from different sources. The work includes extensive archival research of press records at the National Library (i. e. the daily newspaper El Mercurio, and RES n.º 29 (3) (2020) pp. 503-521. ISSN: 1578-2824 the magazines Hoy, Análisis and Qué Pasa) and primary documents from the Library of Congress of Chile and state agencies. In addition, interviews with key players such as high-ranking government officials and business leaders provided a wealth of information to understand the policy process.
The remainder of this work is structured as follows: Section 2 presents the background of the case study. The third section discusses the canonical approach to analysing business power and the need to expand such a framework with additional variables. Then, the paper proceeds in the fourth section with the analysis of the policy-making process for the period under study, and finally it draws some conclusions from the analysis carried out.

Background and structure of the new pension system
Chile constitutes a key case since it has served as a model for market-oriented pension reforms elsewhere. Indeed, Chile's pension reform represents the world's most notable case of pension privatisation (Brooks, 2009;Orenstein, 2008). In the midst of the neoliberal turn in the 1990s, for example, countries in Latin America and Central and Eastern Europe adopted the "Chilean model" to varying degrees (Brooks, 2009;Mesa-Lago, 2012;Orenstein, 2008).
In the early 1980s, the civic-military dictatorship imposed the fully funded, defined-contribution (hereafter DC) pension system. Under no public scrutiny, the Military Junta enacted the Decree-Law 3500 on 4 November 1980, bringing to an end the traditional pay-as-you-go (PAYG) system. Then, on 1 May 1981, the new pension scheme started operating. The reform amounted to a radical departure from the previous system (Borzutzky, 2002). Financially, the new scheme eliminated employers' contributions and transformed the common fund into individual capitalisation accounts. The reform legally mandated each worker to transfer 10 % of their taxable income to their individual accounts. Regarding the administrative structure, the new scheme transferred the management of those Pen-sion Funds to the private sector, which involved the creation of a new type of enterprise: Administrators of Pension Funds (AFPs). Critically, the reform did not include mechanisms for workers' participation in the management of their funds.
The new scheme established AFPs as for-profit corporations. Their responsibility was to invest workers' savings in capital markets. AFPs' profits arose mainly from a commission charged to every worker for the management of their Pension Funds. Most of them charged this commission as a flat rate (known as "fixed commission") plus a percentage of workers' contributions. Under the new system, workers would receive a pension based solely on their individual savings. The new scheme, then, did not include any elements of mutual insurance.
This overall structure ultimately meant that a constant, growing stream of revenue from workers' savings were to be managed by for-profit providers. Since its establishment, Pension Funds have undergone impressive growth (see Table 1). In 1989, after eight years of operation, the combined value of their assets reached $4,4 billion (or 17.8 % of GDP). By 1994, Pension Funds exceed $22 billion in assets, which represented more than 41 % of Chile's GDP (SVS, 2016). For-profit corporations, therefore, were going to take investment decisions over funds which were proportionally extremely significant to the Chilean economy. As such, while these funds had the potential to constitute an engine of economic growth, the system also entailed a potential source of conflict of interests. When the military regime launched the system in 1981, the two largest conglomerates at that time, the Cruzat-Larraín and BHC groups, created two AFPs each (Análisis, 1981). In this way, AFPs, controlled by already mammoth business conglomerates (see Dahse, 1983), were now in charge of managing Chilean workers' savings. As such, the privatisation of the pension system meant huge opportunities to develop economies of scale and vertical integration for these conglomerates with dominant holdings in the banking and insurance sectors. Tromben (2016: 79) reports, in this context, that during the first years of the system, AFPs committed irregular practices. For example, in what constituted a completely irregular financial operation, the two AFPs controlled by Cruzat-Larraín extended sizable credits to companies controlled by the same conglomerate (Tromben, 2016: 146, 178).

The new pension system and the economic elite
Business conglomerates were involved in the shaping of the new system since its origin. Business representatives had privileged access to the policymaking process (Silva, 1996b;Teichman, 2001). The press, for instance, reported that some business conglomerates had started to register trademarks for AFPs even before the Military Junta enacted the actual Decree-law 3500 (Hoy, 1980). Business conglomerates aimed to reap massive benefits from the new system.
Later on, AFPs played a crucial role in the second round of privatisations (1985)(1986)(1987)(1988)(1989) carried out by the military government. AFPs invested massively in shares of former public firms (Hachette and Lüders, 1993). As such, AFPs became influential players in the privatisation process, taking advantage of a loose regulatory framework (Huneeus, 2007;Mönckeberg, 2001). An investiga-tion carried out by the National Economic Prosecutor determined that AFPs indeed colluded among themselves and with ex-executives of former public companies in shareholder meetings, gaining control of privatised companies (Office of the National Fiscal Prosecutor, 1991). This process further strengthened the broad business coalition emerged since 1983 (Silva, 1996b), led by CPC (Confederación de la Producción y Comercio) 1 , the business encompassing association. The new scheme, thus, placed even further capital and power into the hands of the Chilean economic elite. In this context, since its inception, business has had one cohesive voice regarding the defence and maintenance of the scheme.

The AFP system after the Reinstatement of Democracy
Aylwin's campaign platform for the presidential election was explicit about the pension system. While accepting the DC model managed by AFPs, the centre-left coalition promised significant changes. The campaign programme stated: We maintain that to create a broad, solidaritybased social security system constitutes a firstpriority duty. To achieve such a purpose, it is essential to introduce greater elements of justice to the current system by means of various actions (CPD, 1989: 26).
The most concrete actions proposed were the abolition of the fixed commission (given its regressive effect), and to enshrine in law the workers' right to participate in the management of their savings. Certainly, these proposals did not represent a radical departure, but they meant significant modifications. Aylwin's propositions, moreover, were in tune with contemporary evidence, which already showed that the new scheme required significant amendments. For instance, a high proportion of the labour force failed to contribute regularly (Barrientos, 1993), and estimated replacement rates reached just 44 % of final earnings (Gillion and Bonilla, 1992). Even opposition, right-wing politicians, such as senator Sebastián Piñera 2 , accepted that, for instance, the system had exceptionally high costs of administration (El Mercurio, 1993g). Nevertheless, on 25 January 1994, Chilean Congress passed Law 19,301 (Biblioteca del Congreso, 1994), which did not address any of those matters.
This law, commonly referred as the "Reform to the Capital Market", included, among other matters, significant changes to the Decree-Law 3500. The conservative media termed the reform package as "the government's most modernising step" (El Mercurio, 1993b, 1993d. Broadly, the reform extended ceilings for investing in international markets and in domestic companies. Moreover, it established a mechanism to facilitate AFPs' international expansion to other Latin American countries (Biblioteca del Congreso, 1994). As such, the question to address is how, through which mechanisms, the economic elite managed not only to maintain but also to consolidate the scheme. This work now turns to answer that question.

THE STUDY OF BUSINESS POWER
The main challenge in the analysis of business power is to capture the multiplicity of mechanisms and channels employed to influence the policymaking process. Hacker and Pierson (2002) assert that to capture such multiple mechanisms of influence it is crucial not to underestimate the extent of business' power. For analytical purposes, this work employs the classic concepts of structural power and instrumental power. Fairfield (2015aFairfield ( , 2015b has recently developed a landmark contribution by operationalising both conceptualisations of business power in one single framework. Crucially, she shows how both sources of power might interact and enhance each other.

Instrumental Power
Instrumental power relates to deliberate, active political action. It entails political coordination. For the period under study, the Chilean economic elite held every source of instrumental power identified by Fairfield (2015a). Some of them, however, were more significant in the realm of pension reform. Sources of instrumental power may fall into two categories, namely "relationship with policymakers" and "resources" (see Table 2 below).
Regarding relationship with policymakers, institutionalised consultation was of pivotal importance. The authoritarian Constitution certainly limited the scope for progressive reforms in the new democratic setting. As such, businesses viewed the 1980 Constitution as a safeguard. However, according to Silva (2002), the maintenance of their privileged and exclusionary interaction with the government was more (or at least equally) important to the economic elite than the rules established by the Constitution in order for them to accept the transition to democracy. Such privileged access to policymaking developed during the dictatorship persisted during the Aylwin administration (Barrett, 1997;Silva, 1996b). Partisan linkages were also critical. The economic elite had strong ties with right-wing parties, Renovación Nacional (RN) and, especially, with Unión Demócrata Independiente 3 (UDI) (Angell and Pollack, 1990). Programmatic convergence, shared class interests and close ties established during the dictatorship virtually transformed UDI in a business interests' transmission belt during this period (Pollack, 1999). Finally, recruitment into government was also critical. The most obvious effect of recruiting business representatives is that it affords economic elites with direct participation in policymaking (Fairfield, 2015a: 35). In 1990, Minister of Labour, René Cortázar, appointed Julio Bustamante as Superintendent of AFPs. Bustamante was top executive of the Cruzat-Larraín conglomerate (Mönckeberg, 2013), the largest in Chile up to the 1982 crisis (Dahse, 1983). Bustamante participated decisively in the creation of one of the AFPs founded by that conglomerate, and kept his job there as Corporate Lawyer until 1988.
In terms of resources, cohesion and media access were highly influential during the whole policymaking process. Their influence was pervasive and constant over the process, so this section will address its implications. Cohesion refers to the capacity to form and sustain a united front for influencing policy (Fairfield, 2015a: 38). The Chilean economic elite count on a long tradition of high levels of institutionalisation (Schneider, 2013;Silva, 1996b). It counts with six peak business associations (one for each sector of the economy), which in turn coalesce into the encompassing association, the CPC. The CPC aggregate, coordinate and reconcile business interests (Silva, 1996a). In this context, in 1983 AFPs founded "The Association of Administrators of Pension Funds" (AAFP). Although the AAFP has never been a formal member of CPC, there are close ties between them.
For example, José Antonio Guzmán was CPC president from 1990 to 1996. Before assuming CPC presidency, he was president of the Chilean Builder's Chamber. Crucially, the Chilean Builder's Chamber is a business peak association that owns an AFP. Guzmán, as well, was a member of that AFP's board from 2005 to 2014, and president of one of the newly privatised companies (Mönckeberg, 2001). Besides, he maintained affiliations with UDI, as he attempted to run a candidacy for the Senate supported by the party (Qué Pasa, 2014). The case of Guzmán exemplifies very well the broad business coalition forged since 1983, with presence in business associations, the UDI, boards of AFPs and newly privatised companies, and the participation of traditional business groups. Silva (1996b), in this context, documents that since the mid-1980s, the Chilean economic elite has been able to defend a common position in most policy debates. In the case of the new pension scheme, in spite of sporadic disagreements triggered by the banking sector (e. g. El Mercurio, 1990bMercurio, , 1991c, which aims to participate in the "pension industry", there was a strong crosssectoral cohesion during the period under study. Although cohesion does not feature explicitly in the analysis of the policymaking process carried out below, its significance is pivotal. Business organisations were able to process internally differences between the AFPs and the banking sector, such that this particular tension never threatened the cohesive, supportive stance of Chile's economic elite toward the privatised system. Finally, it is difficult to overestimate the degree of influence of the conservative press in Chile. Previ- ous research has tended to focus on the structure of Chile's newspaper market to explain the role played by privileged media access as a source of power (Fairfield 2015b: 433). However, the historical role and influence of El Mercurio, the most traditional conservative daily newspaper, largely exceeds the potential impact given by its market share. Historically, El Mercurio has constituted a "barometer" of the Chilean economic elite's sentiments and political positions (Petras, 1969). In practice, at least for the period under study, El Mercurio sets the daily agenda of Chile's politics, and exerts a powerful effect on individual politicians, who either hope to feature in its pages or fear potential attacks by the newspaper (Personal Interview #1, 2017).
Indeed, the editorial board of El Mercurio consistently articulated a particularly strong and clear position on the new scheme during the entire policymaking process. The newspaper, besides, gave extensively coverage to the debate on the pension system in its specialised section "Economy and Business". For instance, it repeatedly praised the AFP system when Pension Funds showed high returns (e. g. El Mercurio, 1991a). In addition, it offered a reliable platform to AAFP representatives to expand their messages or defend their interests. In the first stage of the policy process, agenda setting, El Mercurio's influential editorial page continuously promoted the alleged scheme's benefits. For instance, in June 1990, the editorial board expressed that the AFP system "was admired in the world", and that Chile was the "leading Latin American country and in a position of global frontrunner" on social security issues (El Mercurio, 1990c). In the second stage, design of the policy proposal, privileged media access afforded economic elite with a source of power continually activated for influencing the content of the reform. For instance, El Mercurio successfully pushed for a raise in the maximum level of investments in international markets (El Mercurio, 1991b). In the last stage, discussion in Congress, privileged media access was pivotal in blocking those initiatives that would mean mild regulations of AFPs' conflicts of interests. Soon after the government introduced the bill, business representatives voiced their concerns about such restrictions because they "infringe upon the freedom of AFPs" (El Mercurio, 1993e). El Mercu-rio as well editorialised a number of times against such measures (e. g. El Mercurio, 1993f). As such, privileged media access afforded economic elite a powerful means of influencing the policymaking on pensions.

Structural Power
Structural power, in contrast to instrumental power, does not require collective action to influence the policy process (see Table 2). Structural power arises, instead, from the profit-maximising investment decisions taken by economic agents. As such, structural power only requires market coordination, not political collective action. It fundamentally depends on policymakers' perceptions. For instance, structural power is at work when policymakers keep a reform off the agenda because they "anticipate" it could provoke reduced investment. Fairfield (2015a: 43) points out that if the sector under analysis plays a crucial role in the economy, the fear of disinvestment affects more decisively policy makers' decisions (i. e. structural power is stronger). Given that structural power arises from perceptions, it is desirable to look at relevant economic data to assess whether stated anticipations were grounded in credible disinvestment threats.
Pension Funds, for instance, have become a crucial source of funds for Chilean companies. By 1990, AFPs had invested more than 11 % of Pension Fund assets in corporate bonds (Superintendencia de AFP, 1991). By December 1994, such investment holdings represented more than 60 % of the stock of bonds issued by Chilean corporations (Superintendencia de AFP, 1995;SVS, 2000) 4 . As such, Pension Funds became critical to corporate debt financing requirements. As for equity holdings from Chilean companies, we have already discussed the importance of Pension Funds for the second round of privatisation. By December 1989, just before the beginning of the period under study, 94 % of AFP's corporate equity investment consisted of holdings of shares in twelve newly privatised companies (Superintendencia de AFP, 1990). The new pension scheme, therefore, constituted a critical gear in the economic structure imposed by the dictatorship, providing cheap and secure funds for the private sector. In this context, policymakers' perceptions of potential disinvestment threats become credible.

Additional factors: ideology and collective memory
In spite of these durable sources of power, this work argues that such variables are not sufficient to explain the outcome of the pension debate in the period 1990-1994. The main contention is that economic elite's instrumental power and structural power help us to explain why the Concertación did not push forward its equity-enhancing agenda. Nevertheless, it remains unexplained the fact that the Concertación ended up consolidating the market-oriented pension scheme imposed by the dictatorship. In this way, structural power and instrumental power play out embedded in a specific, historical context, which include long-term processes with enduring effects. From evidence gathered through interviews, it is clear that both the ascendency of the neoliberal ideology and the fear of an authoritarian reversal greatly reduced the space for equity-enhancing reforms.

Ideology
In their landmark work, Hacker and Pierson (2002) argue that agenda setting is the most critical stage in the policymaking process. As such, influence over the government's reform agenda is usually the most decisive aspect of business power. As we shall see, the hegemony of the neoliberal ideology indeed reduced the political space for equity-enhancing reforms. Roberts (1998: 82) asserts that no ideological vision before the rise of neoliberalism had embedded in such an allencompassing and distinctive way the interests of Chilean the economic elite. From 1975 onwards, the authoritarian regime imposed neoliberalism by means of violence and terror, the education system and the media (Tedesco and Barton, 2004). In this context, the 1981 pension reform came to crystallise the central tenets of neoliberalism, namely: the subsidiary role of the state, the need to augment the role of the private sector, and individualism (see Piñera, 1991). Korpi (1985), in this context, asserts that cultivation and spread of ideologies constitute a major form of investment. This is explained by the fact that the actual exercising of power always entails significant costs (e. g. funding political parties). In extreme cases (e. g. use of violence), the mobilisation of power resources increases the costs but, what is more important, also increases uncertainty. Then, economic elites would have strong incentives "to invest in the conversion of high-cost power resources into low-cost ones", such as ideology (Korpi, 1985: 56). In comparative terms, normative power resources (i. e. ideology) have much lower costs than, say, pressure power resources. The cultivation of ideologies, thus, is an efficient strategy for decreasing the costs of the exercising of power.
In the Chilean case, it is very well documented that, at least from the 1960s, there were concerted investment efforts to cultivate and spread the neoliberal ideology (Herrero, 2014;Valdés, 1995). In the 1964 general election, the Right was virtually swept from Congress, which reduced this group to opposing reformist governments (Correa, 2005;Vergara, 1984). In these adverse circumstances, powerful fractions of the economic elite led a struggle aiming at regaining political influence and protecting their interests (Fontaine, 1988;Fontaine, T., 1992). These fractions were: i) traditional, diversified conglomerates -especially the Edwards group, owner of El Mercurio-, and ii) financial conglomerates with connections to international financial markets, represented by the BHC group. At the time, the Edwards group was the largest conglomerate in Chile (Dahse, 1979), with significant holdings in finance and in internationally oriented companies (Stallings, 1978).
Then, the successful development of neoliberalism in Chile was chiefly the result of a collaborative effort between capitalists and a group of economists trained at the University of Chicago, the so-called "Chicago Boys" (Qué Pasa, 1975). The Edwards group played a critical role in these efforts. Among other activities, Agustín Edwards -head of the business group-supported the dissemination of the doctrines of the Chicago school of monetarism through his daily newspaper. In 1967, for example, El Mercurio created an economic section (Valdés, 1995). As such, Chicago Boys obtained an important foothold in the media, from where they disseminated Chicago School's dogmatic ideas (Fontaine, T., 1992;Soto, 1995). Along with the Edwards group, other conglomerates and peak business associations supported Chicago Boys during the 1960s (Soto, 1995). For instance, Fontaine T. (1992) points out that they started serving in executive positions at the BHC group, the second largest conglomerate (Dahse, 1979). Likewise, other Chicago Boys began working as consultants to the industrialists' peak association, SOFOFA (Fontaine, 1988).
A critical by-product of these intellectual efforts was the development of a cohesive network. From the classic account "The Economists and President Pinochet", written by an insider 5 , it is possible to appreciate how this collaborative efforts to promote neoliberalism nurtured personal ties between Chicago Boys, top-executives and heads of conglomerates (Fontaine, 1988). The development of this network facilitated the preparation of the economic plan for the 1973 military takeover. Months before the coup d'état, representatives of SOFOFA, the industrialists' peak association, appointed a group of Chicago Boys with the aim of drafting a post-coup economic programme (Fontaine, 1988;Valdés, 1995). Then, this group of technocrats prepared the document dubbed "The Brick", which largely defined the agenda during the military government (Silva, 1996b). Then, once imposed, the neoliberal ideology had enduring influences on Chilean society.
Indeed, ideologies tend to have profound and enduring influences over societal consciousness, citizens' aspirations, and norms of fairness (Korpi, 1985). The central tenets of Chicago Boys' ideological view were also strong in the first two Concertación administrations. The conception of offering market solutions to public problems per-sisted in democracy (Tedesco and Barton, 2004). In this context, the role of ideology was critical for the period under study, especially during the agenda formulation stage. The dominance of the neoliberal ideology restricted the scope of the agenda by reducing the set of valid policy options. Policy options not framed in the principle of "market solutions to pu blic problems" would be kept off the agenda, thereby reducing its scope.
To understand the powerful influence of dominant ideologies, consider Figure 1, which depicts the policy options for pension reform. Following Brooks (2009), we might think of a range of pension policy options, ranging from designs based on the principles of risk-pooling and redistribution (e. g. PAYG system) to institutional designs relying purely on market mechanisms (i. e. privatisation, usually linked to DC, individually managed schemes). Between these two polar options we might think of notional, DC pension systems, which link contributions and benefits (as in privatization) but preserve PAYG financing of pension benefits and state management. Such a range of options is represented by the dotted lines in Figure 1, stretching along a simple left-right continuum. In reality, our two endpoints are rarely seen; most pension systems combine some degree of risk-pooling with selfinsurance mechanisms. For example, the radical Chilean pension reform included some risk-pooling through a residual, means-tested minimum pension guarantee. Then, as Brooks (2009) points out, pension reforms typically occur as movements along the continuum defined above and shown in Figure 1, and not as discrete movements from one endpoint (e. g. collective insurance provision) to another (e. g. wholesale self-insurance).
Then, ideology operates as a powerful causal mechanism by reducing the range of "viable" alternatives. Furthermore, it moves the decision making agenda toward the economic elite's preferred end of the spectrum (see Hacker and Pierson, 2002: 284). As Figure 1 conveys, policy options such as the return to a PAYG system were excluded from the debate, thereby moving and reducing the policymaking agenda from "a" towards "b". As such, the agenda's reduced scope would entail just minor changes that would not affect the core of the pension system. That was the case, for example, with the moderate rise in minimum pensions carried out by the Aylwin administration (Castiglioni, 2005: 94; see Figure 1b).

Collective memory
As in the case of ideology, the participation of business associations in the planning and promotion of the 1973 coup d'état it is very well documented. Indeed, business associations were the first domestic organisations that started operating with the aim of toppling Allende as soon as he came to power (González, 2013). Given the direct, active support of business to the coup d'état and the ensuing application of state terror, it may well be argued that the fear of an authoritarian reversal -especially prevalent during Aylwin administration-is somehow a power resource of this group. At least, we may recognise that these historical events had enduring influence favouring business' interests. Evidence gathered in the interviews suggests that this "trauma" determined the agenda significantly.
Memories of the authoritarian government were usually mobilised favouring business interests. As Verovšek (2016: 529) points out, agents usually mobilise understandings of the past as political resources in the present. Political agents, for instance, may deploy memories strategically in order to legitimise their actions. The transition to democracy and the 1990s, in this context, remained embedded in the traumatic years of the dictatorship (Tedesco and Barton, 2004: 150). However, the threat of an authoritarian reversal arose more out of fear than as a result of a pragmatic analysis of the likelihood of such an event. This fact may be well appreciated through analysing Edgardo Boeninger's 6 account of the transition. Boeninger (1997) recognises that during the negotiations to reform the Constitution in the aftermath to the 1988 referendum, government officials discarded the occurrence of a new military coup from their analyses and calculations. As such, the strategy of Pinochet's ministers was to negotiate minor reforms in order to legitimise the Constitution and safeguard the neoliberal model. Otherwise, once in democracy, the popular majority would push for sweeping changes to the Constitution. "It was difficult to think that such surge of transformative energy could be stopped through a new military coup, given that there were not minimal conditions for another 11 September", sentenced Boeninger (1997: 349). Furthermore, he explicitly recognises  Figure 1a represents an agenda in which all policy options might be potentially considered. Figure 1b represents, in contrast, a restricted policy agenda where risk-pooling options or collective insurance provision are kept off the agenda.
that an authoritarian reversal was not a real threat given that "the US backed democracy". Indeed, Petras andLeiva (1988, 1994) document extensively the role of the US in Chile's transition to democracy. In spite of Boeninger's convictions, nevertheless, the Aylwin administration strategically mobilised the fear of an authoritarian reversal with the aim of disciplining leftist elements of the governing coalition.
The above is consistent with evidence obtained through interviews conducted for this research. For instance, a former right-wing member of the Chamber of Deputies during the 1990s argued that "the two shows of military force by Pinochet (he refers to the 1990 liaison exercise and the 1993 so-called 'boinazo') had to do more with personal issues; his family interests were at stake" (Personal Interview #8, 2017). Consequently, according to this account, they did not represent real threats of "authoritarian reversals". This deputy explicitly stated, "I reckon there was not a chance of a military coup" (Personal Interview #8, 2017). In this context, this work argues that collective memory constituted mainly a mechanism to discipline rebellious members of Congress.

PENSION REFORM IN DEMOCRACY, 1990-1994
For the sake of clarity, the analysis of the policymaking process is divided in three idealised stages, namely: i) agenda setting; ii) design of the proposal, and iii) parliamentary discussion (see Figure 2). In each of these stages, this section will attempt to show how different business' sources of power interplay.

Agenda setting
Business' strong structural power significantly restricted the Aylwin administration's agenda on pension reform. The Aylwin administration gave utmost priority to the development of a business-friendly environment. Indeed, the priority was to boost business investment decisions in new projects to sustain the process of economic growth started in 1985, which they deemed as highly successful (Boeninger, 1997). In the words of a former Socialist Party Deputy during the period: Look, the option followed during the transition was to give priority to the promotion of economic growth as the key to reinstate democracy, and hence the government subordinated everything else to such aim. [...] This option included a certain degree of understanding with big business, which was deemed functional to the achievement of economic growth, and undoubtedly Pension Funds was the capital market (sic), the main source of funds to support the economy (Personal Interview #2, 2017). The government's perceived threat was that some reforms or policies might cause investment to stagnate. In this context, a reform to the pension system on the line of Concertación's programmatic platform was removed from the agenda early. The newly appointed Superintendent of AFPs, Julio Bustamante, a few days after the inauguration day, appeased the private sector. He stated that the government was not planning a way back to the old PAYG system, and it only considered introducing minor modifications (El Mercurio, 1990a). While Bustamante confirmed that the government was planning to abolish the fixed commission, there was no more talk of crucial issues, such as the effective participation of workers in the management of their savings. El Mercurio observed that such views "have generated calm among the main agents related to the system" (El Mercurio, 1990a: A3).
Declarations of one Concertación's senator confirm the governments' public stance and show how collective memory augmented structural power. Interviewed for this research, a former Christian Democrat Senator recounted that during the first months of his administration, Aylwin summoned influential left-leaning members of the Concertación. "In a meeting at the beginning of the administration, we discussed this issue with the president (a pension reform), and when I say 'we' I include Concertación's union leaders and members of the parliament close to them. In an honest conversation, Aylwin requested that we not pursue changes to the pension system, given the serious conflicts that the transition was experiencing; we had public and private military mobilisations and Pinochet was still in power [...]" (Personal Interview #3, 2017). According to this informant, Aylwin also included in his analysis "the importance of AFPs for the economy" to gain the support of his partners. This evidence supports the contention that collective memories of the authoritarian period were mainly mobilised to discipline Concertación's rebel, left-leaning members.
Having kept off the agenda major changes to the AFP system, later on there was a further drive to stop any proposal (including modifications to the pension system) that could harm economic growth or generate "business uncertainty". In an internal policy report prepared by the Ministry General Sec-retariat of Government in august 1990, Minister Boeninger (1997: 390) proposed: "All those proposals with potential economic impact that we reckon we cannot promote in the next six months, simply should be eliminated from our agenda" 7 . Structural power is clearly at work here: the policymaker wants to rule out proposals that he anticipates could harm economic growth. Following that strategic line, on 9 September 1990, Julio Bustamante announced that the government was studying a mechanism to reduce the "fixed commission". Some AFPs executives strongly rebuffed the proposal, saying that could create serious complications to their financial positions (El Mercurio, 1990d). The outcome of this discussion was that government never brought the issue forward again, keeping this modification completely off the agenda.
Finally, the dominance of the neoliberal ideology also reduced the scope for progressive reforms. Boeninger states that the economic success of the military regime influenced the Concertación's economic platform, encouraging a programmatic convergence that the centre-left coalition could not admit publicly for political reasons (Boeninger 1997: 369). In this way, ideology reduced the space for equity-enhancing reforms in two ways: 1. By enhancing structural power. Neoliberalism viewed economic growth as the best social policy. Concertación's technocrats believed that economic growth was the main way (not the only, certainly, but the most efficient) to lift people from poverty and maintain stability.
2. By directly reducing the space of policy options (see Figure 1). The extent to which neoliberal ideas exerted influence on the set of potential reforms is well captured by the words of a former member of the Lower House of Congress: The explanation (for the lack of equity-enhancing pension reforms) does not lie either in the shortage of legislative votes, or in the presence of designated senators, or in the lack social mobilisation either. I mean, all those factors were certainly at work to prevent the passage of a progressive pension reform, but the fundamental explanation lies in that all of us, the whole political elite, some more, others less [...] accepted the neoliberal thinking [...]. And therefore, those who were critics of the AFPs became staunch supporters of the AFPs' individual capitalisation model. I am not referring to the right, I am talking of this political world (referring to the Concertación), the one that was governing the country. And we, those politicians that the neoliberal model did not co-opt, did not have a response. At that time, to speak of a system that radically replaced the AFPs (was unthinkable) [...]. To those people (who had dared to advocate such a thought) would have not received a reply, instead they would have been put directly into the sanatorium (Personal Interview #1, 2017).
In summary, the interplay of: i) structural power augmented by the fear of an authoritarian reversal; ii) the dominance of neoliberal ideology, and iii) privileged media access (see above), kept off the agenda a reform to the pension system as promised during the Aylwin presidential campaign.

Design of the Policy Proposal
The design of the "Reform to the Capital Market" started in early 1991. In this stage of the policymaking process, instrumental power played a pivotal role. Relationships with policymakers -institutionalised consultation, specifically-afforded a critical source of power which business activated constantly. Asked by the relation of business associations and the government regarding the debate on pensions, a member of the Board of Directors of the CPC during this period stated, "[we] held multiple meetings and conversations with Aylwin, Foxley, and Cortázar" (Personal Interview #10, 2017). Foxley, as Minister of Finance, and Cortázar, in his position as Minister of Labour and Social Security, were crucial players. According to a number of accounts given by interviewees, Foxley, as Minister of Finance, established the main guidelines of the reform. Then, Cortázar and Bustamante participated in the direct negotiations with AFPs representatives (Personal Interview #5, 2017;Personal Interview #6, 2017). This setting was facilitated by the fact that, according to the 1980 Chilean constitution, the executive holds exclusive initiative on social security (Siavelis, 2000). In this context, the government negotiated the reform with AFPs' representatives before sending it to Congress.
Ideology also played a significant role in this stage. Neoliberal thinking significantly augmented the instrumental power of the economic elite at this stage. According to neoliberal ideology, social policy was not a political issue, but rather had to be addressed technically based on "economic science" (Borzutzky, 2002). Democratic participation, in this view, could just hamper the process. This top-down, exclusionary way of designing and implementing policies persisted during the 1990s, despite campaign promises of developing a more democratic, participatory policymaking process. Boeninger (1997: 370) points out that the incorporation of more liberal ideas among Concertación's technical cadres was facilitated by the political process' nature in this period: notoriously elitist, restricted only to the upper echelons of political elite who worked with considerable freedom. This elitist nature of the policymaking process obviously augmented the potential influence of the economic elite, since it participation was not counterbalanced by labour power.
Press records also offer evidence of the economic elite's participation in the proposal design. The AAFP led the negotiations with the government, acting in the interest of the whole economic elite. Given the elitist, exclusionary policymaking process, technical expertise constituted a source of instrumental power that enabled business interests to participate in the design process. As it was noted, the Minister of Finance was the branch of the executive in charge of the bill's preparation. Officials of such a ministry tend to be highly trained economists, so technical expertise afforded AFPs the opportunity to dialogue and negotiate with those officials. Then, for instance, by August 1992 the general manager of the AAFP declared that they had commissioned a report to the Department of Economics of Catholic University and Econsult 8 to analyse and propose new investment options (El Mercurio, 1992). For the Association, this report would be important since it was to be the source from which they would "make comments and add proposals to the legal text in progress" (El Mercurio, 1992). As such, at least six months before the executive introduced the bill in Congress, the private sector was well aware of its content. Finally, when the government introduced the bill in January 1993, the president of the AAFP praised the reform package (El Mercurio, 1993c). He stated that the bill addressed demands that the sector was formulating since long ago.

Discussion in Congress
On 25 January 1993, the government introduced the reform bill known as the "Reform to the Capital Market". Since the beginning of the discussion, the government presented the reform as a "modernization" of the capital market, in the words of Finance Minister Alejandro Foxley (El Mercurio, 1993a). Nevertheless, at the same time, the Aylwin administration downplayed the fact that this was mostly a reform to the pension system. True, the reform modified different aspects of the capital market, but changes to the pension system were the most significant. The fact that the government negotiated the bill with AFPs before introducing it to Congress, determined the parliamentary discussion. Several interviewees concur on the fact that private pension administrators simply did not need to lobby in Congress (Personal Interview #7, 2017;Personal Interview #8, 2017). The most decisive business influence, therefore, mostly occurred before congressional debates.
At this stage, partisan linkages afforded the most important source of instrumental power. Considering the composition of the Senate, from the beginning it was clear that the Aylwin administration would have to do deals with the right-wing forces in Congress (Angell and Pollack, 1990). Especially given the higher quorum required to reform the pension system. Some aspects required a quorum of four-sevenths of the congress, given that fell in the category of "leyes orgánicas institucionales". This meant that the government needed 27 senators to enact the law, and the Concertación had only 22 (Angell and Pollack, 1990). As shown in Table 3, the Concertación won a majority of the seats in the Senate in the 1989 election. However, the nine institutional senators -established in the 1980 Constitution-appointed by Pinochet, ultimately meant that right-wing forces held the majority. Institutional senators banded with UDI and RN elected representatives. These dynamics afforded instrumental power to business given that they could impose their terms in negotiations.
In this context, partisan linkages were crucial to water down significantly two articles of the bill proposed to regulate conflict of interests. As mentioned above, AFPs had been involved in grave accusations of collusion and mismanagement, which proved to be true (Office of the National Fiscal Prosecutor, 1991). The specific place in which these changes took place was the Senate's Finance Committee, which must process the bill before the general vote in the Senate chamber. Meanwhile, in October 1993, Sebastián Piñera -member of the Committee-recognised that he had been in touch with representatives of the AFPs (El Mercurio, 1993h). Other parties 3

Sub-total 22
Institutional Senators Armed forces appointments and other 9

TOTAL 47
Source: adapted from Angell and Pollack (1990 Originally, the first article established restrictions to the way in which AFPs exerted Pension Funds voting rights in board elections. The measure established that AFPs should vote for independent candidates in the boards, establishing specific restrictions. The second article originally established that 30 % of AFPs' boards had to be formed by directors subject to specific prerequisites. The main responsibilities of those directors should have "to act for the benefit of the affiliates, who have entrusted AFPs with the management of their Pension Funds", and to "monitor the occurrence of potential conflict of interests" (Biblioteca del Congreso, 1994: 17). The right on behalf of Sebastián Piñera and one institutional senator opposed both articles in the Finance Committee. The rejection was stated on the grounds that those measures would "severely affect the private property right and the freedom of association guaranteed in the Constitution" (Biblioteca del Congreso, 1994: 691). Finally, the Congress passed a significantly watered-down version of both articles. Though it established that AFPs had to inform the way in which they exerted their voting rights, this legislation did not prevent the network of patronage and personalisation established around AFPs. Based on the analysis of board members of newly privatised companies elected by AFPs for the period 1995-1998, Huneeus (2014) established that almost half of them held posts in the civic-military regime (19 out of 40), and that 95 % (38 out of 40) had right-wing political affiliation. As such, with democracy well underway, there was complete continuity with the dynamics of the authoritarian regime.
Alongside political links, technical expertise -augmented by ideology-also played an important role at this stage. As in the case of the design of the bill, the top-down policy-making process enhanced business power. Repeatedly during the discussion in Congress, government officials and some members of the parliament asserted that this was a highly technical project, and, as such, the best option was to leave the issue to "experts" (Biblioteca del Congreso, 1994). Business representatives participated actively in the committees' hearings. The AAFP, for example, at least since the late 1980s, counted on a strong technical department that allowed it to participate actively in the finance committees (Personal Interview #9, 2017). In contrast, congressional records indicate that labour leaders were totally absent during the discussion of the bill in both chambers of the Congress. Furthermore, in what was a very irregular situation given that the reform package modified substantially the pension scheme, the Senate Labour and Social Security Committee did not have a chance to discuss the bill. In spite of a formal request raised by this Committee, the president of the Senate rejected the petition, stating, "the board has serious doubts that a bill with a distinctly financial nature has to be processed by the Labour Committee" (Biblioteca del Congreso, 1994: 1102. This way of doing politics did not represent the Concertación promise of carrying out a more democratic policymaking process.
Finally, in this stage collective memory also operated. Concertación strategists appealed to the fear of an authoritarian reversal in order to procure the necessary votes from Concertación's left-leaning members. The name most associated in the interviews conducted for this research with this dynamic was Edgardo Boeninger, who, as the president's Chief of Staff, was in charge of managing the government's relations with Congress. A Socialist Party Deputy interviewed for this research, for instance, stated that even in the face of minor disagreements, Concertación's members of parliament were compelled to support government's bill appealing very explicitly to the threat of the authoritarian reversal (Personal Interview #1, 2017). The threat of another military coup, the threat of the instability that could cause legislation against business interests, was voiced in order to discipline Concertación's congressional representatives (Personal Interview #7, 2017). In this regard, Senator Ruiz denunciated "undue influences" when he substantiated his vote in favour of the reform (Biblioteca del Congreso, 1994: 1156. As such, the support of conservative, mild reforms was viewed as the price to be paid in order to maintain democracy.

CONCLUSIONS
The debate on pensions in Chile for the period 1990-1994 illustrates the significant impact of business power. Undoubtedly, structural and instrumental power constitute helpful categories for understanding the mechanisms of business influence. Nevertheless, in this case study collective memory and the dominant neoliberal ideology also played a significant role at different stages of the policymaking process. It is necessary, therefore, to supplement the classical framework for studying business power with these two additional factors in order to achieve an understanding of the full scope of the economic elite' influence. The role of collective memory is especially critical, for instance, in enhancing structural power to rule out policy options in the most critical aspect of the policymaking process, agenda formulation. Although less explicit, the dominance of neoliberal ideology also had an effect in blocking an equity-enhancing pension reform in the agenda setting. In this context, this work highlights the need to adopt an intertemporal, historical approach and to consider business power and influence as processes that unfold over time.
Secondly, this work poses a question about the significance of democracy to promote particular policy outcomes. The policymaking process through 1990-1994 closely resembles the exclusionary nature of this process during the authoritarian regime. The Chilean transition to democracy seems to have perpetuated the elitist, non-democratic nature of the policy process instituted during the authoritarian regime. Indeed, the dominant business coalition which emerged in the mid-1980s continued to dominate the debate on pensions.
Finally, from the analysis carried out it seems that the Chilean economic elite was even more powerful under democracy than under the authoritarian regime. To begin with, business had a persistent control over the policymaking process, maintaining a position of quasi co-legislator. Moreover, in democracy the Chilean economic elite did not have to deal with Pinochet's high levels of autonomy, which in some critical moments led him to take harsh decisions against big business' interests (e. g. during the 1982 banking crisis). As such, the establishment of formal democracy does not appear to have secured the development of a more participatory, democratic policymaking process.