PENSION REGIMES AND DISTRIBUTIONAL PRINCIPLES

The analysis of pension regimes moves the focus of research away from the “black box” of social expenditures, towards the distributional principles guiding old-age policy. The mechanisms by which a given pension scheme is set out to allocate rights, benefits and costs across individuals can indicate the equity priorities underpinning policy design. With the use of a dynamic conceptual framework which, unlike existing welfare typologies, is well equipped to capture specific distributional trajectories in pension policy (in other words, policy change rather than stability), and the systematic analysis of social security legislation, this research was able to identify and differentiate the three distributional models that have existed over the history of pension policy in Argentina, each one embedded in the political and ideological logic of the government of the period.

The first pension regime originated with the schemes created at the beginning of the century, and consolidated with the welfare expansion in the 1940s and 1950s. Over the first half of the century, a particularistic distribution of pension rights restricted benefits to workers in occupations that were either economically or politically strategic for the government (notably, civil servants, journalists, and banking employees in the first few decades, and industrial workers later on, with the ascendance of Peronism). Even after the expansion of pension coverage to previously uncovered groups (in the second half of the century), benefit levels and eligibility conditions continued to vary across occupational categories, creating intergroup inequalities that often reflected the bargaining power of each group.⁶

Within occupational categories, however, intragroup solidarity prevailed, especially over the 1940s and 1950s. Progressive benefit formulas granted higher relative-to-wages benefits to lower income workers. This transfer of resources occurred only within occupational groups, in a typical corporatist organization, which indicates the principles of distribution that underpinned the design of the system over this period.

The second pension regime emerged after a substantial reform applied in 1968/69, under the so-called bureaucratic-authoritarian state.⁷

The military government that took power in 1966 (with General Onganía in the presidency), aimed to dismantle the type of corporatist organization inherited from the Peronist past, and to concentrate power in a central state operating top-down. Pension reform in this period reflected the new political orientation. The previously fragmented structure of pension provision was centralized and fully administered by the state. This canceled most of the particularistic features of the system. Eligibility conditions and benefits were largely standardized across occupational groups (with the exception of the military, who kept their privileges).⁸ A single replacement rate was set as the basis for benefit calculation for all workers, thus reducing particularism but also simultaneously eliminating intragroup solidarity and progressive redistribution within occupational categories. The system concentrated on an income-replacement scheme, and left narrower scope for income redistribution.

The third and currently existing pension regime was adopted with a structural reform in the 1990s, which redefined the entire mechanism for benefit administration and distribution. After the pension crisis in the mid 1980s, it was clear that something had to be done to put pension finances back on track, and to rebuild the crumbling social trust in the system. The shape the reform would take was molded by the economic orientation of Carlos Menem's government, a Peronist with a wide neo-liberal reform agenda. As part and parcel of the overall package of structural reforms applied during this administration, the pension reform passed in 1993 (and implemented in 1994), partly abandoned pay-as-you-go for a system of individual funded pension accounts, in a local version of what the World Bank has named the “three pillar model.”⁹

The reform partly followed the Chilean experience as well as the new policy directions coming from the most influential intergovernmental organizations towards an individualized model of resource distribution. Choice, self-reliance, and individual equity consolidated as the new normative features guiding pension policy. Eligibility became more strongly based on contributory capacity and compliance, and benefit levels more closely reflected individually made contributions. Redistribution via pension policy was reduced further, and workers were transformed into “savers.”

LIFECYCLE DISTRIBUTIONAL IMPACTS OF PENSION REFORM

Over the second half of the century, successive reforms have affected the relative gains and losses of different generations of workers. The political relevance of pension policy often means that reforms involving losses for large or influential groups tend to be applied only relatively far in the future, in order to contain public resistance.¹⁰

This generally implies that, given a particular pension system design, if the evolution of economic and demographic variables leads towards financial imbalances in the long-run, younger generations of workers will be more likely to bear the costs of adjustment (with the consequent intergenerational inequalities this generates). This is a pattern that has been found in many countries where the process of population ageing has pressed governments to introduce parametric or structural reforms to their pay-as-you-go systems. In the Argentine case, the distributional impacts of pension reform operated both inter- and intragenerationally. Over history, reform was not just a way to balance the accounts; it was also a mechanism by which the distributional principles of the system were redefined. The paced adoption of an individually actuarial type of system (where benefits depend on individually made contributions) tended to reduce intra-generational redistribution. The progressivity of the system was concentrated at the very bottom, only for the covered population, in the form of a state-provided basic flat-benefit.

In order to evaluate gains and loses for individuals living and working under different institutional contexts, the pension rules existing throughout the entire lifetime of each worker (rules of eligibility, benefits, contributions, and so on), were combined with data on wages, labor market patterns and other demographic indicators. This lifecycle approach is a substantial methodological advantage over most common analyses of resource distribution in cross-section, because it captures the essence of pension policy as a system that transfers resources over the lifecycle. The calculation of pension gains and losses by each worker or generation provided virtually the first piece of evidence on the distributional impacts of the pension system from a historical viewpoint, and showed that pension policy in Argentina had more significant redistributive impacts than those deriving from risk-pooling.

Intergenerational income transfers were substantial. Cohorts retiring during the early stages of the system were relative gainers, whereas later cohorts, who experienced tougher eligibility conditions, were relative losers. A mean-wage worker born in 1910 could obtain real returns on pension contributions that were up to ten times higher than those obtained by workers born in the 1940s. Intragenerational transfers were both horizontal (between gender groups) and vertical (between income brackets). Women tended to obtain higher returns than men (as a result of higher life expectancies and lower retirement ages), although the value of their benefits was often lower than those for men, due to lower wages and shorter contributory histories. Low-income groups benefited from higher returns if retiring before the 1960s, when progressive benefit calculation formulas prevailed. Although some redistribution was maintained after the introduction of individual funded accounts in 1994, low-income workers (who often spend long periods of their working lives in the informal sector) were more directly affected by the longer contributory periods required.

IMPLEMENTATION PROCESSES AND OUTCOMES

While distributional principles and distributional outcomes were analyzed following the rules for the calculation and allocation of pension benefits, these rules were not always applied as written. In periods of substantial economic decline and fiscal crises (for example, during the 1980s), ad hoc adjustments were introduced as emergency measures. Changes in the context of implementation (labor market, prices, economic activity, and so on), also influenced the outcomes of the system. Indeed, the process of implementation can have substantial effects in redefining the distributional impacts of the pension system. The idiosyncratic enforcement of pension rules as well as changes in the political context or socio-economic environment may make initial distributional projections or expectations unworkable. This is particularly important in the case of developing countries where the socio-economic context is less stable and predictable and the rule of law does not always prevail.

Over the history of pension policy in Argentina the process of implementation was found to substantially affect the outcomes of the system, over and beyond the actual dictates of the legislation. Pension policy outcomes were found to be a function of the interaction of pension system design with the context for policy implementation. Contemporary socio-economic environment, policy-feedback from previous policies, and political pressures have affected the application of pension rules as written, and often created what have been called “policy gaps” between observed and expected outcomes.¹¹

Four major “policy gaps” were found over the history of pension policy in Argentina.¹²

First is the “coverage gap” of a system that claims to be “universal” and “standardized” (that is, to provide equal access and equal treatment) but that, by the year 2000, in fact covered only 52 percent of the working population, and in which uncovered workers were more likely to be women, poorly educated, and low income. Second is the “pre-1968 funding gap.” Although up to 1968 the pension system was supposed to be funded, already by 1958 it was dedicating roughly 80 percent of revenues to pay benefits. Accumulated funds fell sharply in only a few years due to inflation and poor investment, eroding the basis of the funded system and making it operate, in practice, as a pay-as-you-go system well before the legislation established the change in the mechanism of resource administration. Third is the “post-1968 financing gap.” The pension system was always designed as contributory and self-financed, but in the mid 1980s, transfers from the general budget represented over 40 percent of pension expenditures, a figure that rose to 60 percent in the year 2000, rapidly eroding the basis of “self-financing” and with it the justification of coverage exclusions based on individual contributory histories. Fourth is the “benefit indexation gap”: although benefits were supposed to follow wage levels—to allow pensioners to share the increases in standards of living enjoyed by workers—the real value of benefits fell by half in the late 1980s as a result of the inadequate application of benefit indexation rules in a context of high inflation.

The existence of significant “policy gaps” in key areas of system performance raises the issue of policy failure and points to the operation of informal policymaking mechanisms, which work beyond the written rules, but can shape policy outcomes. Thus especially in countries where significant “policy gaps” exist, the process of implementation becomes a key variable to consider when designing and evaluating policy performance. As implementation contexts and practices differ across countries, this also means that no single “best practice” model of pension policy is likely to produce “best” outcomes everywhere—“best practice” policy learning need not only operate at the level of policy design, but crucially, at the level of the interaction between design and implementation.