This volume is a compendium of Carmelo Mesa-Lago's decades of study and experience regarding pension reforms in Latin America. Its merits include brevity (the online version is 194 pages), multi-country coverage, including data, and the fact that a free version is available online. The book sets the scene with a useful summary of the debate about pension design, making clear its depth, breadth and complexity (for a fuller discussion, see Nicholas Barr, ‘Reforming Pensions: Myths, Truths and Policy Choices’, International Social Security Review, 55: 2 (2002), pp. 3−36; Nicholas Barr and Peter Diamond, ‘Reforming Pensions: Principles, Analytical Errors and Policy Directions’, International Social Security Review, 62: 2 (2009), pp. 5−29). The rest of the book is organised around four questions.
Q1: What happened? The value of this short chapter lies in the tables with cross-country comparative data on the types of reform in the different countries (Chile, Colombia, Costa Rica, Dominican Republic, El Salvador, Mexico, Panama, Peru and Uruguay), and the extent to which pensions in those countries were privatised.
Q2: Through what process were the reforms introduced? Alongside the reforms themselves is the process leading to reform. Mesa-Lago attaches considerable importance to open debate and transparency, arguing that both elements are helpful to good reform. Though true in general, that argument is only part of the story. First, there are exceptions. Chile reformed in 1981 under the Pinochet dictatorship, yet the system of individual funded accounts – albeit unpopular and under pressure – remains intact. Separately, it is hard, if not impossible, to disentangle two roots of the unpopularity: the origins of the reform under the dictatorship, and the level of pensions. In addition, successful reform also needs a long-term view – reform is not an event but a process. The United Kingdom is a democratic country with reasonably transparent data, yet (with rare exceptions) its pension reforms have been too frequent, too hasty, too incoherent and with too little consultation. In Sweden, in contrast, the cross-party Pensions Group has been very effective as a long-term guardian of the reform.
Q3: How did the reforms perform relative to what was initially promised? This long chapter, the heart of the book, looks at what was initially promised and what subsequently happened in different countries under five headings: coverage, social solidarity and gender equity, adequacy, efficient and cost-effective administration, and sustainability.
Coverage: an important conclusion is that ‘the private pension system in most of the nine countries has not managed to substantially improve the coverage’ (p. 38 of the online version).
In terms of social solidarity and gender equity, Mesa-Lago points out that individual funded accounts provide consumption-smoothing but no risk-sharing either within or across cohorts, hence do not address social solidarity; and given that women's contribution records are generally worse than men's, nor does the design improve gender equity. That would not matter if other parts of the pension system addressed those two objectives. But the initial reforms in Latin America were weak on those aspects (see Indermit Gill, Truman Packard and Juan Yermo, Keeping the Promise of Social Security in Latin America (Stanford University Press, 2005)), hence the importance – both for Chile itself and in the example it set for other countries – of Chile's introduction of a solidarity pension in 2008, improving both social solidarity and gender equity.
In relation to adequacy, a worker's pension accumulation depends on (i) his/her hourly wage, (ii) hours of work and (iii) the number of contribution years. A professional in formal-sector employment is likely to do well on all three dimensions and hence have a respectable replacement rate (i.e. pension relative to the average wage). Many workers, however, do not score well on all three, with a recurring problem being incomplete contribution records. Table 10 show the average replacement rate in different countries, with the unsurprising conclusion that countries that relied mainly on individual accounts had average replacement rates lower than those with defined-benefit pensions as at least part of their system. The analytical bottom line is that it is inherent in the defined-contribution design that benefits will produce a low replacement rate for workers with low earnings and incomplete contribution records.
And regarding efficient and cost-effective administration, the argument at the time of the initial reforms was that competition between pension providers would produce the same result as between supermarkets, i.e. lower costs and increased choice. That did not happen – the findings of behavioural economics explain why (see Daniel Kahneman, ‘Maps of Bounded Rationality: Psychology for Behavioral Economics’, American Economic Review, 93: 5 (2003), pp. 1449–75; Richard H. Thaler, ‘From Cashews to Nudges: The Evolution of Behavioral Economics’, American Economic Review, 108: 6 (2018), pp. 1265–87). Workers do not vote with their feet by moving to cheaper and/or better providers (Table 12), and charges (Table 13) and administrators’ profits (Table 14) remain high.
Finally, on the issue of sustainability, a section considers different dimensions of the finances of pension plans: declining contribution densities, employer behaviour (evasion, under-declaration, etc.), growth of total funds under management, concentration of investment portfolios, rates of return, voluntary saving and pre-retirement withdrawals. The empirical findings align with theoretical predictions in two ways. First, a defined-contribution plan is by definition financially sustainable in the sense that its obligations are tied to what accumulations can finance, though not necessarily politically sustainable in the sense that the resulting benefits may be inadequate. Second, the results are consistent with what has been learned from information and behavioural economics and search theory. Though unsurprising, the findings are useful. First, they reinforce the argument that the problems of individual accounts are more fundamental than that of poor implementation. Second, they are not limited to a snapshot, but document performance over time.
Though valuable as a summary, the findings give only a partial picture. Mesa-Lago looks at performance relative to what was promised. That question is the right one in political terms. In economic terms, however, the question is of the counterfactual, i.e. how did the reformed systems perform on the various dimensions relative to (i) no reform or (ii) a different reform, for example realistic parametric change in a defined-benefit plan. Put another way, how did the reforms perform relative to moving to (say) a notional defined contribution (NDC) plan such as in Sweden or a partially funded national defined-benefit plan such as in Canada? Since that question – though the right one – is unanswerable, failure to address it is not a criticism of Mesa-Lago's analysis, but an indication of what his findings do and do not say.
Q4: What have been the re-reforms and their effects? Many of the countries that introduced individual accounts have since partly or wholly rolled them back (e.g. Argentina), and/or reformed other parts of the pension system (e.g. the solidarity pension in Chile). Mesa-Lago refers to these changes as ‘re-reforms’. The first part of the chapter describes the re-reforms in a range of countries using the five headings above as an organising principle, and the second part seeks to evaluate the re-reforms following the same themes.
The final chapter moves from questions to setting out 18 recommendations, ranging from points of process, e.g. supporting the independence of supervisors, to the fundamental question of which pension model to follow. Spoiler alert: Mesa-Lago – rightly – does not propose a single best system, instead ‘outlin[ing] three alternative models (given as examples that are not exhaustive)’ (p. 179 of the online version, emphasis in original).
That last part, though short, is the right place to conclude. The primary focus of this book is not on which pension design is best, but on setting out and evaluating in an orderly way the wealth of information on the performance of pensions in Latin America since 1980. As such, the book is an invaluable companion to pension reformers by an author with unparalleled knowledge of Latin America.
