THE LONGITUDINAL STUDY OF SOCIAL PROTECTION IN CHILE

This workshop session considered information on the nexus of work, retirement, and social protection for older people in Latin America. David Bravo, Pontificia Universidad Católica de Chile, began by describing the Survey of Social Protection (EPS) in Chile, a longitudinal study that has been conducted since 2002.¹ The EPS is the longest-running panel study ever fielded in Chile. It is a nationally representative survey of individuals 18 years and older, with a sample size of approximately 20,000 people. Four waves of the survey have been conducted to date, with a fifth wave to take place in 2015. The survey has a face-to-face interview that lasts about 1 hour.

Bravo explained that the origin of the EPS survey stemmed from a public policy question. In 1981, Chile underwent a major reform of its pension system, and an individual-capitalization system was established. For bureaucratic and other reasons, it was difficult to determine how the new system was functioning. Individuals could have information about their own retirement accounts, their savings, and other items, but the government did not have access to this information. For fiscal planning and projection purposes, national-level information was very poor.

At the same time, policymakers in government ministries began to ask about broader topics, such as the relationship between health and pensions and the relationship between family and pensions. It became clear that to answer questions about pension-system trends, better information was needed about mortality risks, numbers of survivors, health, and disability. Teams of academics proposed and developed a longitudinal survey that would help the government with answers to these questions, and the Ministry of Labor and Social Provision funded the survey.

In addition to basic information on individuals and household members, the survey collects information on work history, family income, health, education, family history, financial knowledge, and other topics. Survey data have been linked to data from administrative registries, and a refresher sample of respondents is planned for 2015.

Bravo reported that EPS information has been widely used for academic purposes both within Chile and abroad. Within the country, the survey has had an important impact in terms of further pension-system reform. In 2006, Chilean President Michelle Bachelet established a commission charged with evaluating the pension system. Some of the major questions that the commission considered involved the density of effective contributions; the extent to which workers meet the essential requirements for pension receipt; perceptions of pension contributions as a type of tax; the impact of financial shocks on the labor market and social protection; the characteristics of independent workers and the self-employed (most of whom do not participate in the pension system); and the effect of chronic diseases or catastrophic diseases on the system contribution patterns.

Using data from the EPS, the commission suggested various system reforms, many of which were enacted into law in 2008. The survey played an important role in the design of reform from the standpoint of both diagnosis and assessment. Since 2006, the treasury and budget departments have used an actuarial model that relies on the longitudinal EPS data. And EPS data are now being linked with data from the national death registry in order to analyze mortality and survival by socioeconomic stratum.

Bravo presented data from the EPS showing that pension system contribution patterns were quite different for men and women. The data suggest that, regardless of gender, many people have paid too much or too little into their retirement accounts, which Bravo attributed to a lack of knowledge about the pension system. There is also a general lack of understanding about the pension industry and how pension fund administrators compete amongst themselves.

In conclusion, Bravo stated a need to (1) incorporate biomarkers into survey research to better understand the interactions between health and aging; (2) improve linkages to administrative data; and (3) consider the establishment of an HRS-type survey in light of the growth of the older population in Chile.

NONCONTRIBUTORY PENSIONS

Emma Aguila, University of Southern California, described a randomized controlled trial and related research to evaluate the impact of a noncontributory pension program on the older population in the Mexican state of Yucatan. She began by noting that Mexico does not provide universal coverage of social security benefits. Forty-two percent of the labor force is in the formal sector, which includes public and private sector employees as well as self-employed people who choose to contribute to the system. These individuals receive, at retirement, social security and health care benefits. Most workers, 58 percent of the labor force, are in the informal sector. These are workers in noncompliant small firms and self-employed workers who choose not to contribute to the social security system. Upon retirement, these individuals may receive noncontributory pension benefits that are of smaller amounts than social security benefits; they also receive health care provided by the government. Aguila mentioned that most developing countries have a high proportion of their labor force in the informal sector, and that at least 20 countries around the world have implemented noncontributory pension programs. Studies have shown that these types of programs reduce poverty and inequality (see, e.g., Barrientos, 2003; United Nations Department of Economic and Social Affairs, 2007; Asian Development Bank, 2012).

She explained the thrust of this research project was to design and implement a noncontributory pension program in Yucatan for adults aged 70 years and older. The main goal was to evaluate the impact of this noncontributory pension program on the health, nutrition, and well-being of recipients. A secondary goal was to analyze different implementation designs, for example, the method of payment, cash versus debit card, and differences in the frequency of payment.

Aguila explained that an important aspect of this study is that it was experimental, in contrast to previous experiences in other countries. In South Africa and Brazil, for instance, most studies of the noncontributory pension schemes compared beneficiaries with nonbeneficiaries; if claiming benefits were a choice, this would introduce sample selection problems in those studies. Also, the development of noncontributory pension programs in other countries has not been accompanied by surveys that evaluate the programs, so countries typically try to use data from household surveys (if available) that focus on the household level rather than the beneficiary level. The Yucatan project also affords an opportunity to disentangle the causality between socioeconomic status and health, she explained. The project involves a cash transfer of 550 pesos per month, equivalent to $78 U.S. at purchasing power parity. This is exogenous income to the recipient, and researchers can study the effect of income on health.

The research team conducted three experiments. The first began in 2008 and involved two towns: in the town of Valladolid, all age-eligible individuals 70 years and older were identified and received a pension, while all age-eligible people in the town of Motul were identified but were not provided with a pension, thus establishing a control group. The second experiment began in 2009 as a pure randomized controlled trial in the city of Merida, with treatment and control individuals within a given area. The third experiment began in 2010, also in Merida, and varied the proportion of treatment and control individuals within areal clusters such that researchers could conduct more in-depth analysis and identify potential spillover effects.

The research teams, working with government entities in Yucatan, built two government programs, one to disburse the pension and another to evaluate the impact of the program (Aguila et al., 2014). In the evaluation program, there was a baseline survey for the treatment and control groups before the intervention, and the pension program was announced a month after fieldwork began, to avoid announcement effects. In-person follow-up surveys were conducted approximately every 6 months, which included collection of biomarkers and anthropometric measurements. To date, interviews have been conducted in 16,195 households. Many survey questions were adapted from the Mexican Health and Aging Study and Health and Retirement Study described earlier in the workshop (see Chapter 3). Also included was a community-level survey of prices conducted in 1,987 grocery stores and other establishments, to determine potential effects of living in areas with different levels of poverty and price structures for goods (e.g., medicines).

Aguila summarized the main results that were observed 6 months after implementation of the program. They included an increase in food availability and reduction in the incidence of hunger spells, a 22 percent increase in health care utilization as measured by doctor visits, and a one-third increase in medication access measured by reports of not being able to afford medicines. In terms of health outcomes, there were highly significant improvements in cognitive abilities (immediate and delayed recall), measured improvements in lung function, and a 10 percent reduction in anemia (Aguila et al., 2015). In terms of family transfers, there was some crowding-out effect of family transfers; data suggest that money flows from children and family members to those receiving the noncontributory pension declined by half. Another important result has to do with the implementation of the program, in terms of the frequency of pension payments. Researchers found differences in behavior between those who receive monthly versus bi-monthly payments (the latter are the federal government norm). While a 2-month disbursement is more cost-effective from an administrative standpoint, the salutary effects of pension receipt mentioned above were generally enhanced among those receiving a more frequent payment. There is evidence that people have difficulty smoothing their consumption between pay cycles, both with regard to food consumption and health care usage.

There was also some indication that people receiving the noncontributory pension were more likely than those in the control group to stop working. One workshop participant suggested that it would be very useful to further investigate the interaction between employment and the noncontributory pension, particularly in light of what is known about the importance of work for individual identity, network formation and strengthening, and receipt of health benefits. Aguila agreed and mentioned that one next step is to analyze the type of work that individuals aged 70 and older are engaged in, most of which is in the informal sector and often of a sporadic nature.

PENSION SYSTEMS IN LATIN AMERICA

Roberto Ham Chande, Colegio de la Frontera Norte, focused on what he called “a great need to renew the Latin American pension system, particularly in Mexico.” In his view, the major underlying topic of the entire workshop has to do with interpreting old age as a dependency. He said he believes that old age should be defined as a state of dependence, especially in terms of economic security. Economic security can be achieved through a promise of future support by family, by children, or by the state. One of these promises can be a pension, i.e., an agreed-upon form of support, legally validated, accepted without reluctance, and claimed with a full sense of entitlement. A widespread but misleading belief, he said, is that personal savings are the means of allowing people to purchase the goods and services needed during aging and retirement.

Policies to promote these promises have been discussed during the workshop. However, Ham Chande noted, established policies often do not deal with what is currently happening. Rather, they deal with what will come, with the hope that whatever will come will be better. For this reason, marketing is needed about what old people really need during their retirement, based on data and projections. He divided needs into four parts. The first involves necessities of daily living (food, housing, clothing). The second is health care, including clinical care, medications, therapy, exams, and hospitalization. The third is long-term care, which may be provided by families but may also involve assisted living arrangements and nursing homes. The fourth is family and social life, including a focus on social networks and entertainment. By far, health and long-term care are going to be the most expensive.

In Ham Chande’s opinion, the focus should move beyond the monetary aspect of future needs. Referencing an historic paper by Dawson (1912), Ham Chande noted that there is a fundamental fallacy in the notion that people are supported in old age from their savings. Another focus must be on the distributional aspects of economic, political, and social systems, and on notions of the proper distribution of goods and services, including among older individuals. He noted that this has been a theme in the literature beginning more than 100 years ago, and continuing to the present (see, e.g., Barr and Diamond, 2006; Lee and Mason, 2011).

Ham Chande presented several of his thoughts about pension systems in Latin America. The first is that pensions with a defined benefit represent a burden that is impossible to pay as now structured. This is partly due to population aging and increased life expectancy, and less so to corruption within social security institutions. Another aspect, in his view, is that pensions have become public debt because of labor unions, which have created a series of privileges that cannot be paid anymore. The same fundamental economics underlie a pay-as-you-go tax system and funded private savings, he asserted; all pensions systems, funded or unfunded, involve a transfer of resources from workers to retirees. It is also the case that administrative costs of pension system management are eroding rates of return, with a high risk to individuals who are often unaware of these costs. Hence the transition toward individual, defined-contribution pension systems may not provide an adequate retirement for large numbers of workers.

Investing in government bonds, said Ham Chande, is equivalent to pay-as-you-go, but with higher administration costs. Retirees who try to sell assets to a smaller next generation will realize relatively low prices. This is because longevity and retirement risks actually reside in, and are determined by, the economy. The only way that funded systems can do better than pay-as-you-go systems is by creating a greater gross domestic product, he said. In his view, the best way to assure the future of social security is by investing in youth, health, and education.

One primary task that nations need to undertake is to review actuarial, economic, and social estimates and reconsider their sustainability. Countries must bear in mind that health care costs are going to grow exponentially, and may well be more expensive than pension costs. Ham Chande suggested more attention to econometric models and to use of the “demographic dividend” framework for understanding intergenerational transfers (Crespo et al., 2014). Countries should exploit the first demographic dividend by taking care of young populations so that they can form the social and economic infrastructure that guarantees transfers to aging populations; this becomes the second demographic dividend.

According to Ham Chande, another important task, and the most formidable, is to more openly discuss the critical conflict of vested interests within LAC societies. The privilege associated with some union-related pensions is an obvious topic, and Ham Chande suggested seminars and fora on this subject. Another topic he identified is how to refocus financial systems’ emphasis on speculation toward investment in infrastructure, to help realize the potential of demographic dividends.

Footnotes

1

For an overview of and data from this study, see http://www​.previsionsocial​.gob.cl/subprev/?page_id=7185 [August 2015].