1. Challenges for Collective Risk Pooling

There are some challenges for the collective risk pooling systems that, I believe, are underestimated by Otsuka. First, he is of course aware that collective risk pooling cannot work if people have the freedom to defect after they become better informed about their own situation. They may have some freedom to choose or decline a pension at the outset of their working life reaching the age of majority, but only at the outset, because otherwise the efficiency of the scheme in pooling longevity and investment risks would be undermined. However, he is confident that at the point of entry, the (real) veil of ignorance is thick enough because both longevity and investment risks are roughly equal in society. I am afraid that this is too optimistic. The differences in life expectancy between children from different socioeconomic descent are large (and well known or experienced by everybody). More importantly, children who can reasonably expect that they will inherit from their wealthy parents, are aware that they have a stronger buffer to cope with future risks, even if these risks are the same. Moreover, even before inheritance, there is now in most Western economies an increase in the inter vivos gifts from parents to children, and the wealth position of the parents is well known by children reaching the age of maturity. There is therefore a non-negligible issue of adverse selection, which can only be settled through strict rules: if a sufficiently large group of children from a wealthier background set up their own scheme that also may have some (albeit minor) risk pooling advantages, the broad ‘nation-wide’ scheme may suffer.

Second, as emphasized by Otsuka, the good working of the collective multigenerational risk pooling system requires a constant inflow of newcomers into the scheme. I find it surprising that the ageing problem, implying that the size of the retired cohorts is growing relative to the size of the young active cohorts, is not really discussed in the book. It can be solved, of course, by increasing contribution rates, working longer or lowering pensions. As Otsuka rightly emphasizes, promises are not absolute. Yet, to guarantee the stability of an ambitious nation-wide scheme, it seems advisable that the adjustment rules about how to react to structural changes in society are set in advance as much as possible, i.e. that they are stipulated in the (implicit or explicit) social contract. Here again, government intervention would help a lot, and is perhaps even necessary to overcome transaction and information costs in writing the ‘contract’.

2. Inequality and Collective Risk Pooling

Until now, I did in no way question the basic pension design proposed by Otsuka. Yet, I now want to raise the question: is this all? Does justice not require more than just a risk pooling arrangement based on mutual advantage? Otsuka admits the soundness of a redistributive case for pensions, but he claims not to focus on this because he wants to show that a strong case for collective pensions remains, even in the absence of grounds for redistribution from rich to poor. I agree, but I wonder about the ethical status of that case. It can be seen as a kind of public choice (or political efficiency) argument, as a strategy to defend systems of collective risk pooling in societies where interpersonal redistribution (or the old-fashioned term ‘solidarity’) comes more and more under pressure. But this is not the main strategy in the book. Otsuka defends (certainly in the last chapter) that the collective risk pooling system embodies a principle of justice, in that the ‘mutually advantageous move is voluntary rather than coerced, and the stronger party does not take advantage of the weaker party’ (87). And, when there is a conflict with more egalitarian justice principles, ‘when everyone has enough so that nobody is in need the demands of equality need not always trump the strong Pareto improvements of mutual advantage’ (89). This then leads to the full description of a just pension system: ‘A state pension should be sufficient to meet our basic needs for income in retirement. Above that floor, there is a sound case for the mutually beneficial risk pooling of a collective pension even if it arises from a baseline of unequal income’ (89).

In fact, I think he could have kept even more firmly to his collective insurance paradigm when proposing this system. Indeed, a substantial basic pension, covering basic needs, could be part of the insurance arrangement, agreed upon by the people at the beginning of their career. At that moment, they are not only confronted with investment and longevity risks when arranging for their retirement, they also face uncertainty about their future career path. Some will become severely ill, unable to work, or (involuntarily) unemployed – while other parts of the social insurance system (and to some extent progressive income taxation) may help coping with these risks, they do not as such solve the challenge of keeping a decent living standard when retired. Introducing within the pension system a generous minimum may be an answer to that challenge. In fact, allocating a high universal minimum income when retired should not create too many distortions in the economy.

There is of course the difficult question of how to fix the level of that basic pension. Defining what are basic needs in a rich society is a particularly tricky question. Yet, even assuming this question to be settled (e.g. within an insurance arrangement based on self-interest), the question remains: is this enough? If one endorses a principle of egalitarian justice, in which individuals should be compensated as much as possible for outcomes for which they cannot be held responsible, including their productive potential in so far as it is linked to their genetic endowment and the quality of their upbringing when young, then just guaranteeing everybody a minimum when retired is certainly not sufficient. Otsuka describes as a possible egalitarian alternative (that he does not support) a state pension that would not take into account labour income, but would only be based on number of years worked. This is an extreme system, that we do not observe in practice, partly because it would completely fail to avoid a severe fall in the living standard of the high-income people at the moment of retirement.

Yet, it is obvious, certainly in countries with a pay as you go system, that distributive considerations weigh heavily in the social debate on pensions. Different countries have opted for very different answers to distributive questions, but there is most often an intensive discussion about caps (or not) on contributions, caps (or not) on pension benefits, compensations for periods of unchosen inactivity (like disability or involuntary unemployment), different retirement ages for people in arduous jobs, etc. Otsuka notes that pay as you go systems, basically managed by the government, offer the opportunity to redistribute, but then (understandably) does not focus on this, because he wants to interpret pay as you go as an instrument of collective risk pooling. As mentioned before, he basically shows that pay as you go systems indeed allow for collective risk pooling, and from this perspective, converge towards collective defined contribution systems and funded pensions with defined benefits. Which immediately raises the question: if pay as you go offers risk pooling possibilities comparable to those of the other systems, is it then not an advantage that it also allows for some redistribution? Precisely because we have a choice between different collective risk pooling schemes, we can choose the one that offers the best redistribution opportunities.

Of course, caution is needed. Introducing redistributive features would imply that we can no longer count on voluntary participation behind the thin veil of ignorance at the start of the professional career. In the language of the veil, we need a thicker one which also removes knowledge of one’s earning potential. This would bring us into the philosophical world of hypothetical insurance, rather than the real world of ex ante mutual advantage. Formulated in economic terms, the insurance system (if voluntary) will have to take into account a participation constraint, implying that individuals are better off in the collective system, even taking into account its redistributive features, compared with other solutions. This moves us away from pension systems with a clean design (such as the one proposed by Otsuka) into a complex and clumsy system, in which design choices are codetermined by contingent empirical information.

3. A Broader Social Contract?

My feeling is that the limits of the individualistic approach become obvious when one looks at the real-world struggles around pension reform. The Robinson Crusoe metaphor has become very popular among economists. It has proven to be very useful and it has strongly influenced our thinking about social systems. Yet, society does not consist of a countable number of Robinson Crusoes: it is a network of social relations. And the retirement problem was solved in earlier times (and other cultures) through an extremely imperfect pay-as-you-go system within extended families, where contributions were partly inspired by mutual advantage, but equally strongly by social norms about how to behave towards the elderly. No need to be romantic about that solution: very often it did not work at all, and it lost its effectiveness with the break-up of the institution of the extended family. Yet it illustrates how pension design fits into the broader social fabric.

In fact, there is no reason why the broader egalitarian approach could not be applied to individuals in different cohorts. Intergenerational redistribution can be justice enhancing. Of course, the issue of responsibility is tricky in a setting with cohorts, but there are obvious cases where redistribution between cohorts seems eminently just from this egalitarian perspective.

A good illustration is the fate of the ‘first’ generation at the moment of introducing pay as you go in Western European welfare states. This generation received a pension without (in Otsuka’s interpretation) ever having contributed, which he considers unfair in his reciprocity framework. Yet, historically, this does not seem to have played a significant role in the decision-making process and there was not much resistance in society to make a ‘gift’ to that first generation. As soon as we see society as more than a system of self-interested monetary transfers between and within cohorts, but as a network of social relations that go far beyond insurance arrangements, this is not surprising. This first generation had seen the collapse of the existing funded system. More importantly, it had lived through the Depression and the Second World War and had invested a lot in rebuilding the economy and the educational system. The golden sixties did not come about automatically, they were to some extent created by that first generation.

I think two interpretations are possible here (and I believe that they are both relevant and related). First, in society there is room for generosity and/or for gratefulness, and nothing is wrong with that. Trust and social coherence are essential elements of the welfare state, and pensions are an essential component of that welfare state. In fact, the sophisticated nation-wide and multigenerational risk pooling schemes as proposed by Otsuka cannot be built on self-interest alone, but will only be sustainable if there is sufficient trust and social coherence.

Second, if we keep to the contribution-benefit paradigm of Otsuka, it is difficult to see how one can detach the pension system from other examples of intergenerational relations. Contributions to and benefits from the collective efforts of society go further than pensions. Generation G invests (or does not) invest in education, in economic growth and in the conservation of natural capital. Generation G+1 will reap the fruits (or have to bear the consequences) of the decisions of this older generation. As mentioned above, the investments in the human capital of the boomer generation by the previous generations were a crucial element in the economic growth experienced by the boomers. At the other side, the feelings of the presently young generations that the boomer generation has left them with a huge problem of climate change certainly affects the intergenerational relations and the social sustainability of the present pension system.

I believe that the notion of collective risk pooling is extremely important for a well-designed pension system. Otsuka’s book explains this clearly and convincingly. From an analytical perspective, one cannot do better than what he did: focusing as sharply as possible on this specific insight. Yet, the pension system is a crucial component in the broader web of social relations within and between cohorts. To understand this better we have to go beyond mutual advantage and looking at broader justice issues – without losing sight of the importance of collective risk pooling.