In 1981 Chile adopted a new pension system that has set a controversial pattern around the world. Unlike traditional systems, benefits are financed by investment accounts owned by workers. Chileans are sensitive about the starting point of their system, under Pinochet. But they do not worry about whether their system will run out of money as baby boomers retire, because benefits are financed by investments accumulating in the accounts, not by taxes paid by current workers. The funds are privately managed, therefore insulated from political interference. Studies have shown that Chile’s new system contributed to its rapid rate of economic growth during the past 20 years. But that doesn’t mean it can’t be improved. Chileans have been busy fine-tuning their system since it started.
Over the past decade, most Eastern and Central European countries have adopted reforms that follow the Chilean example, but with a relatively smaller personal account and larger remaining role for the public benefit. Several Western European countries are converting or considering converting their funded employer-sponsored DB plans into DC plans (personal accounts). What are the successes and pitfalls of the Chilean model? How can we learn from - and improve upon - Chile’s scheme?
First, a bit of history. Before Chile launched its system of personal accounts its pay-as-you-go social security system was out of control, beset by overly generous promises of benefits, high rates of evasion, early retirement and hyperinflation. A payroll tax rate of 23% covered only 60% of total pension benefits; the rest was covered by government subsidy. Clearly a non-sustainable system. And clearly in worse shape than western European countries are in today, although not much worse than some eastern and central European countries before they reformed.
In 1981, workers in Chile were allowed to choose between the new and old systems. Practically everyone under the age of 50 (and some older workers as well) switched. New entrants to the labour force were required to join the new system, so eventually the old system will be completely shut down. (This contrasts with the UK, where workers can switch back and forth between SERPS and personal accounts, raising costs and uncertainty for the pre-existing state system). José Pinera, then Minister of Labour and now a leading advocate of private accounts in the US, was on the TV every week and warned workers that their contributions to the old system were going down a black hole. This was an easy sell because the new contribution rate - 10% for the accounts plus 2.6% for administrative costs and disability and survivors’ insurance - was about half the combined employer-worker payroll tax it replaced.
Since the government had to go on paying benefits to retirees and older workers who stayed in the old system, this raised a transition financing gap. Chile planned for this transition by building a large fiscal surplus. The surplus and continued fiscal discipline, together with the accounts, were responsible for increasing Chile’s national saving and productivity, and therefore the size of the GDP pie. It’s not clear how the eastern European countries that did a similar reform financed their transition costs. If they used pure debt finance to cover the transition, this would have offset the increase in personal saving and the opportunity to increase national saving and GDP would have been missed.

Has it been a good deal? Yes. The annual rate of return on the account balance during the first 22 years (before taking fees into account - see below) was 10% above inflation - fortunate for the new system but far above the rate that any country can maintain in the long run. But even if that rate of return is cut in half, while wages continue to grow at 2% above inflation, the average Chilean worker who contributes until he retires at 65 will get 60% of his final wage for himself plus a survivor’s pension for his spouse. This is about the same as traditional European countries hope to provide, but at much lower cost.
Of course, higher returns potentially entail higher financial market risks – a point that is often made by critics of personal accounts. Chile reduced risk in several ways. Most important, it established a minimum pension guarantee (MPG), financed out of general government revenues, for any worker who contributes at least 20 years. The minimum is especially important to women, whose lifetime income and pension accumulation is much lower than that of men. Every year the government has raised the minimum pension, roughly in line with wage growth. This keeps the pension floor around 25% of the average wage, way above the poverty line. (A smaller means-tested pension exists for those with less than 20 years of contributions). Only 4% of all beneficiaries collect the minimum pension subsidy so far, but the numbers (and costs) will increase as retirees who didn’t buy annuities use up the money in their accounts. Many European countries (including the UK, the Netherlands, Switzerland and most of Eastern Europe) also offer a minimum pension of some sort and are also worried about rising future costs.
Chile also reduced the volatility in returns and disparity across individuals by penalising funds whose returns deviated by more than two percentage points from the industry average. Of course, this caused “herding” among pension funds, which defeated a major rationale for personal accounts – to give workers control over their retirement savings. Choices were further circumscribed by Chile’s undeveloped financial markets and by regulations that tightly limited allowable investments. Initially foreign investment was ruled out. Was this too little choice? Probably. But at the beginning the government wanted to avoid a disaster, to prevent some workers from making big mistakes. (In contrast, when the UK scheme of personal accounts was put in place a few years later, it gave workers too much unguided choice and they proceeded to make big mistakes). As Chile’s financial markets developed and workers gained investment experience, the range of allowable alternatives increased—see below.
How do people get their money out of the accounts? Through annuities or through gradual withdrawals that follow a formula set by regulators. Regulations gave an advantage to insurance companies selling annuities, salesmen raced around convincing workers to buy, and about two-thirds of all retirees have annuitised. This is good because annuities protect retirees from investment risk and from the risk of outliving their saving. (Most European countries with personal accounts also require or strongly encourage annuitisation. This is a sharp contrast to Australia, where lump sum withdrawals are permitted; some analysts fear retirees will use up their retirement savings too soon and fall back on the means-and asset-tested old age pension).
In Chile all annuities must be price-indexed, so retirees are protected from inflation. Indexed annuities are also required for part of the retirement savings in the UK, but retirees pay a high price for this insurance. The cost is much lower and the money’s worth ratio much higher in Chile, because Chile has an indexed currency in addition to a nominal currency, and many price-indexed financial instruments in which insurance companies can invest, to hedge their risk. The government guarantees a large part of the annuity so retirees are also protected from the risk that insurance companies will become insolvent – and government strictly regulates the insurance companies to prevent this from happening. Married men must purchase joint pensions to protect their wives when they die and women can keep their own pensions as well as the joint annuities. (In contrast, widows in the US and some other countries must choose between their own pension and the survivors’ benefit, so many working women get no additional benefit after years of contributions).
All these guarantees will cost money in the long run, although the costs haven not kicked in yet. But the risks (of low earnings or low returns or insurance company failure) are there and someone has to bear them. Chile made the decision that the taxpayer is the risk-bearer of last resort – a defensible decision. However, moral hazard is always present, these costs were not adequately projected at the beginning and will probably turn out to be higher than expected.
So what’s the downside to Chile’s scheme? What flaws have the Chilean detected and tried to correct? What should Europe do differently?

First, as discussed, workers in Chile had practically no investment portfolio choice for the first 20 years. Each asset manager could offer only one portfolio and all portfolios were very similar. A young worker might want to make riskier investments than an older worker, but he could not do so. In 2002 Chile modified its rules to enable a wider variety of investment instruments and choices. Each pension fund can now offer five different portfolios, with different allowable equity proportions for each one (ranging from 0-80%) and the foreign investment limit is raised to 30% of total assets. The five portfolios allowed in Chile are a far cry from the 600-odd portfolios allowed in the new Swedish system. But the Swedes are reconsidering. Structuring limited choice with clearly differing risk-return trade-offs is probably a good way to go in a mandatory scheme where many workers are inexperienced investors.
Second, Chile has been criticised for its high administrative costs. Administrative costs are surprisingly important in determining final pension size. Fees that are 1% of assets per year, paid over the entire working life, will reduce pensions by 20%. Costs in Chile were indeed high at the beginning – over 10% of assets in the first couple of years, as assets were small and investments in new information technologies and marketing were large. But as assets and experience grew, costs fell dramatically and are now about 1% of assets, lower than in the average European fund. This pattern of high initial costs that decrease dramatically as average account size grows has also been observed in every European country that has started personal account systems.
But a mandatory system that has hundreds of millions of euros running through it should be able to do better, cutting marketing costs and increasing bargaining power by aggregating many small accounts into larger sums. Sweden does this, and allocates the aggregated sum among mutual funds according to worker choice, without revealing the names of individual workers to the fund managers. Kosovo auctioned off the rights to run the funds to two asset managers, which pushes down their fees and incentives for marketing. Mandatory systems could cut costs further by requiring the use of index funds, which are much cheaper than actively managed funds yet earn equivalent returns, on average. Collecting contributions through the existing tax collection system and using a simplified record-keeping system are other possible ways to economise on costs. Sweden is hoping to keep its costs about half those in Chile by a combination of these techniques.
Third, Chile allows workers to stop contributing and start gradually withdrawing once they meet a minimum threshold. While the principle is good, the threshold has been set too low. Most workers qualify before age 60, some even before 50 – and they have taken their pensions as soon as they can. Since the minimum pension rises with wage growth, eventually many retirees will qualify for a top-up from the government. Chile is now in the process of raising this threshold, but more needs to be done.
At the same time, taking the pension does not mean you must stop working. In fact, Chile encourages older individuals to continue working by the close linkage between benefits and contributions and by exempting pensioners from the payroll tax. While Europe’s workers have been withdrawing from the labour market early, Chile’s have been working longer as a result of these incentives – a big plus of the current system.
Fourth, Chile’s minimum pension is good, but could be even better. On the positive side: it has risen with wages, so keeps low income pensioners from falling way below the average standard of living. On the negative side: someone who works for 19 years is not entitled to the minimum pension, while someone who contributes for 40 years gets no extra safety net. Some low earners try to evade contributing beyond 20 years because their additional contributions simply replace the subsidies they would have obtained from the government. The minimum pension also leads workers with small retirement accumulations to avoid purchasing annuities, because other forms of payouts allow them to get their money sooner – and once they run out of money the government picks up the bill. As a result, the costs of Chile’s minimum pension will probably increase over time. These loopholes should be plugged – by tying the size of the minimum pension to years worked and by slowing down the rate at which retirees can withdraw their money.
Finally, Chile’s new system has been criticised for low coverage rates and a high proportion of missed contribution years among members. The number of affiliates actually exceeds the size of the labor force, but only 60-70% of those affiliates contribute at any given time. Many workers will have low pensions from the contributory system, and the MPG liability may be higher, as a result. This is a problem, but it is not due to the nature of the present system.
Coverage rates were similar, and even lower for many years, under the old system – leading to its insolvency. In the new system if a worker doesn’t contribute he undermines his own pension but not the solvency of the system as a whole. Coverage rates are invariably low in less developed countries, where government’s tax collection capacity is low and a larger role is played by small firms, self-employment and agriculture, which are notoriously hard to include in a contributory system. Western Europe doesn’t face this problem in a big way, but some countries in central Europe and the former Soviet Union do. One solution might be to expand the role of noncontributory benefits – but these would have to compete for scarce public funds with other pressing social needs.
In summation: It works. Chilean workers will get a higher pension relative to wages than they would for the same contribution in a pay-as-you-go system. And their system is financially sustainable. But this requires a large regulatory and risk-reduction role for government, including a minimum pension that has been rising over time. The majority of retirees annuitize, but they also start withdrawing their money as soon as they can – so early withdrawal and lump sum withdrawal rules must be chosen with great care. Administrative costs took a large bite at first, but fees have come down dramatically as a percentage of assets as average account size has grown. Other countries should be able to keep costs lower from the start by aggregating contributions and utilising economies of scale and bargaining power. Perhaps most important, Chile’s exemption of pensioners from the pension payroll tax has increased the labour supply of older individuals. Its pre-funding of pensions and fiscal discipline in financing the transition increased total national saving. Its development of institutional investors - pension funds and insurance companies - increased the demand for and therefore the supply of a variety of financial instruments and services, including mortgage-backed securities, secondary bond markets and credit-rating institutions. All of these are credited with raising Chile’s rate of economic growth and GDP during the past two decades. As European countries reform their systems they should keep their eye on the ball - the importance of incentives for work, saving and better functioning labour and financial markets - because ultimately this is the source of growth that will provide greater security both for old and young.