In defence of Chile’s privatised pension system
What had gone wrong? Up until the Great Recession, the Chilean pension system had been going from strength to strength. Aspects of it were being copied all over the world, and Chile became something of a ‘Mecca’ for pension reformers worldwide. It is probably a reflection of the system’s emblematic status that the recent protests, although technically strictly a domestic matter, have attracted an unusually high level of international media coverage. BBC News reported on the subject, as did BBC World, the Guardian, the Economist, the Financial Times, the Daily Mail, Reuters, the New York Times, the Los Angeles Times and Bloomberg.
Unfortunately, most of the media just repeated the protesters’ slogans, thus giving a wrong impression. There is nothing wrong with the Chilean pension system as such. The system provides excellent value for money, and decent pensions for those who have saved regularly throughout their working lives. What it cannot do is magically transform a record of sporadic contributions into a high old-age living standard. The system needs to be benchmarked against a realistic counterfactual, not against a magic money tree. This is what my new IEA briefing ‘Prefunded pension systems: recent developments’ sets out to do.
How the system works
The Chilean pension system, introduced in 1981, is a system of mandatory old-age savings, underpinned by a tax-funded old-age safety net. Working-age people pay 10% of their gross salary into a retirement savings account, managed by a pension fund provider – known as AFPs (Administradoras de Fondos de Pensiones) – of their choice. Since 2002, people have also been able to choose between five different risk levels.
Taken together, Chile’s pension funds have accumulated assets worth the equivalent of about two thirds of the country’s GDP. Pension funds are highly diversified, with AFPs investing in a range of sectors, financial instruments and countries. The average annual rate of return since 1981 has been 8.4% after inflation. Pension funds took a hit during the financial crisis, but quickly bounced back.
The system also has a safety net component, which has been reformed and extended in recent years. Its main instrument is a means-tested, tax-funded minimum pension. People who have no pension income of their own receive the full amount. Support is then gradually withdrawn with income, and tapers off at the 60th percentile of the income distribution. The poverty rate among the elderly is lower than that of the population as a whole – 3.9% vs 10.3%, or 8.4% vs 14.4%, depending on the poverty measure used.
Chile’s 1981 pension reform has also given rise to a number of positive economic spillover effects: the prefunded system has been an active ingredient in the accelerated economic development that the country has been experiencing since the mid-1980s. It has demonstrably increased savings and investment; it has increased employment, especially in the formal sector; and it has boosted the development and sophistication of Chile’s capital markets, thus raising productivity.
So what’s not to like?
High returns – low pensions?
How can a system which consistently offers such great rates of return offer low pensions? Where do all those returns go? The system’s critics argue that they end up enriching the companies managing the funds. David Blake, a professor of Pension Economics at London’s Cass Business School, claims:
“Initially the Chilean model appeared to be very successful, but the sting in the tail appears to be that charges extracted by the industry have resulted in pensions being much lower than otherwise would be the case.”
This is nonsense. AFPs fund themselves via an income-related monthly fee. That fee is not taken out of people’s retirement savings accounts; instead, people pay it on top of the 10% they deposit into their accounts. This means that the fee cannot diminish people’s future pensions: It can only reduce their current consumption.
The critics nonetheless have a bit of a point. The AFP industry is characterised by high levels of concentration, customers show low levels of price sensitivity, and market entries and exists are not a frequent occurrence. AFPs have therefore historically shown high profit margins, often exceeding 20%. So there is certainly scope for greater dynamism and competition in this sector. But this would not, on its own, increase people’s pensions.
Latin American welfare systems usually cover only a small part of the population. The main reason is that these economies contain large informal sectors, and people in informal occupations are not covered by formal social protection mechanisms. This is less of a problem for Chile than for its neighbour countries: by Latin American standards, Chile is an unusually formalised economy. Nonetheless, about a quarter of the labour force work in the informal sector at any given time.
A peculiarity of informal employment in Chile is that, unlike in the neighbour countries, it is rarely a permanent arrangement. Chileans don’t spend their entire working lives in the informal sector, but many switch back and forth between formal and informal employment. This means that almost all of them have an AFP account (unless they are part of one of the parallel pension systems), and are, in this sense, ‘covered’. But having an account, and contributing actively, are two very different things.
Only about one in three men, and one in four women, contribute regularly, if we define ‘regularly’ as ‘for at least four years out of any five-year period’. At the other end of the spectrum, one in five men, and one in three women, only contribute very rarely, if we define ‘very rarely’ as ‘for less than a year out of any five-year period’.
This is why, in the Chilean case, ‘the average pension’ is not a meaningful concept. Suppose Chile had a larger informal economy, and less movement between the formal and the informal sector. This could dramatically raise the average AFP pension – simply because some of the most sporadic contributors would drop out of the AFP system altogether.
Among people who have contributed for between 24 and 32 years, the median AFP pension amounts to 46% of their final salary (the pension replacement rate). This is still not high by OECD standards. But then, the average pension contribution rate in OECD countries is just under 20%, compared to a mere 10% – applied to a narrowly defined contribution assessment base – in Chile. Meanwhile, Chile has virtually closed the life expectancy gap with the developed world. With this in mind, the Chilean AFP system offers excellent value for money.
More moderate critics of the Chilean system want the state to become a more active player in the pension fund market, not least by setting up a state-owned AFP. The more radical critics – especially the ‘No+ AFP’ movement – want to dismantle the AFP system entirely.
As for the former, if people are not putting enough money into a privately managed retirement savings account, they are not suddenly going to put a lot more money into a publicly managed one. The creation of a state-owned AFP would therefore not address the system’s main problem. But it could create a host of new ones.
A state-owned AFP could become a tool for channelling resources into the political pet projects du jour. Apart from the economic damage, this would genuinely lower people’s future pensions.
How about a return to a PAYGO system? PAYGO systems essentially come in two varieties, namely as Bismarck systems and Beveridge systems. Bismarck systems are contributory systems, characterised by a strong relationship between current contributions and future pension entitlements. Beveridge systems aim to provide a basic universal standard, which is not, or only loosely, connected to people’s prior contribution record.
The No+ AFP movement does not specify what exact alternative they want, but a Beveridge system is clearly not what they have in mind. They promise higher pensions across the board, not just a guaranteed minimum standard, which is, in any case, already provided by the solidarity pillar. A Bismarckian PAYGO system, however, would replicate the current system’s main problem. Just as a record of low and patchy savings leads to low private pensions in the AFP system, a record of low and patchy contributions leads to low state pensions in a Bismarckian system.
More importantly, a Bismarck system would create massive fiscal problems in the long term. Chile is a young, but rapidly ageing society. Accordingly, modelling shows that while (under restrictive assumptions) a return to a PAYGO system could produce some short-term gains, these gains would dissipate quickly, and after a decade, the cost of a PAYGO system would begin to rise rapidly relative to the cost of the current system.
Few developed countries would introduce PAYGO systems today. But given that they already have them, they shy away from the transition cost of moving to a prefunded system now. Chile is, in this sense, in an admirable position. They have started the transition when demographic conditions were still favourable, and have now paid off the bulk of the transition cost. If introducing a PAYGO system today would be nonsensical, returning to a PAYGO system after having borne most of the transition cost would be madness.
There is room for improvement, but there is no need to ‘think outside the box’ – there are plenty of workable solutions within the box, and plenty of non-solutions outside of it.
Informal economic activity normally results when the cost of entering the formal sector is too high – when there are, for example, too many bureaucratic hurdles to clear for setting up a company legally. Streamlining and simplifying such processes is the way to formalise the economy
The issue of low market entry rates in the pension fund industry could be solved by allowing banks, insurance companies and other financial intermediaries to also offer retirement savings accounts. At the moment, they can only do so by setting up a separate AFP company, entailing a duplication of fixed costs.
Meanwhile, price sensitivity among customers could be increased by allowing group contracts. Trade unions, professional associations, employers and civil society organisations could be allowed to act as intermediaries between their members and the pension fund administrators, shopping around for the most lucrative offers on their behalf.
As explained, lower AFP commissions would not in themselves increase pensions. But ultimately, Chile will need a higher pension savings rate, and raising it will be easier if AFP commissions fall at the same time.
The current retirement age of 65 years for men and 60 years for women is also no longer realistic, and the distinction between a male and a female retirement age is outdated anyway. The retirement age should be pegged to life expectancy, and the male-female gap should be closed.
The impact of unemployment-related career breaks could be cushioned by integrating the pension system and the unemployment insurance system. Rather than just smoothing current consumption, unemployment insurance could be used to also smoothen pension contributions. The integration of the self-employed, meanwhile, is already under way, even if implementation is lagging behind the official timetable.
The idea of prefunded pensions may well have fallen out of fashion. But intellectual fads and fashions come and go. The economic case for prefunded pensions remains as strong as ever.
Read Dr Kristian Niemietz’s IEA briefing paper ‘Prefunded pension systems: recent developments’ here.