What an Ageing Europe Can Learn from Chile

By Kristian Niemietz, Research Officer for Health and Welfare at the Stockholm Network, Diplom-Volkswirt (M.Sc. in Economics) of Humboldt University, and author of the book “Die kapitalgedeckte Altersvorsorge am Beispiel Chile” (Diplomica Verlag, 2008)

“Consider the difference between a private attendant and a court servant. The latter will put up with much more, for he has a pension to expect.” Otto von Bismarck (1881)

“You are absolutely right, general. If this reform is approved, a giant step into the direction of reducing state power will have been made.” José Piñera (1980)



For decades, Chile’s PAYGO public pension system had appeared impossible to reform. Benefits were determined by the political process but paid from a common pool, so the formation of vested interests was heavily encouraged. Early retirement privileges were a popular election promise. As a result, the system became financially untenable. At the end of the 1970s, its deficit increased up to 40% of its expenditures, or 3% of GDP, and increasing rapidly.

Since the 1950s, there had been a broad political consensus that the pension system was being excessively abused by pressure groups, had become burdensome and bureaucratic, and that it paid low pensions to the bulk of the elderly. These criticisms were repeated by governments from very different political factions, be they conservative, Christian Democrat or Socialist. They were also backed by the results of various expert committees installed by these governments, but no coherent solution was ever proposed. In 1973, out of political and economic chaos, a military government emerged. Paradoxically, this government hesitantly accepted advice from a group of classical liberal economists who were enthusiastic about rolling back state power.

In December 1978, José Piñera took over the Ministry of Labour and Social Security. Initially he ended the most grotesque abuses of the PAYGO system by standardising some key variables such as retirement age and inflation adjustment. But Piñera feared these improvements would likely be reversed by successors and therefore prepared steps to unhinge old age provision from the political sphere once and for all.

The creation of the AFP system.

José Piñera proposed a radical reform entailing the creation of a fully funded, defined contribution, pension system and announced it in his May 1st 1980 speech. But the resistance of assorted “social security experts”, of diverse groups benefiting from the PAYGO system, and, unsurprisingly, by members of the military, blocked it. On September 1980, the government called for a referendum about a new Constitution that included a gradual transition to democracy. The country viewed the referendum also as a poll over the market liberal reforms. The positive result provided the needed back-up. The referendum was accepted with a two thirds majority and Jose Piñera seized the opportunity of the moment and put his reform plan back on the agenda. This time, he got his way.

The legislation which laid the foundation for the new pension system was finally signed in November 4, 1980 and came into effect on May 1, 1981. The managers of the workers individual retirement accounts would be new companies, called the Administradoras de Fondos de Pensiones, or AFPs. They were to manage the personalised savings accounts belonging exclusively to the affiliated workers – not to the AFPs. The AFPs were prohibited from engaging in any other commercial activities, and required to organise disability and life insurance for their clients, via a group contract with a life insurer. A highly technical regulatory watchdog, the Superintendencia de AFP, was created.

It was decided that nothing should change for those who had already retired, while new entrants to the labour force had to join an AFP. Current contributors were given a choice. They could either remain in the PAYGO system under unchanged conditions, though the multitude of existing pension regimes and institutions were merged into a single organization, or they could switch to an AFP. In this case, their already accrued entitlements were quantified by a standard formula and paid out to them in the form of a “recognition bond”, due upon retirement. This recognized the unfunded liability of the PAYGO.

The transition phase had to confront many difficulties. A capitalisation scheme requires sophisticated capital markets, or at least risk rating agencies, bond custodians and an insurance sector, all of which were underdeveloped in Chile. For the transition phase, a governmental risk classifying committee was set up, and bonds custody was given to the Central Bank. Both tasks have gradually been passed to the private sector. The insurance sector was liberalised and oversight was tightened, so that it could grow alongside the AFP system.

Today, a Chilean employee, instead of paying the government a payroll tax, has to deposit 10% of monthly gross earnings, up to an earnings ceiling, into a personal retirement account run by an AFP of his or her free choice. This contribution, as well as additional voluntary pension savings up to the same ceiling (APV), is tax exempt. Currently, six AFPs are competing in the market, each of which offers five types of pension funds. These are arrayed by risk and return, A-Funds can contain up to 80% of instruments with variable returns, B-Funds up to 60%, and so on, E-Funds being composed exclusively of fixed-return assets. Workers aged under 55 can choose any fund they wish, older workers become ineligible for A-Funds and retirees for B-Funds. AFPs charge a wage-related fee which pays administrative expenses and pays for disability and survivorship insurance. A retiree can choose to retain his AFP account, purchase a lifetime annuity from a life insurer, or combine both options. A time-deferred combination, in the sense that a retiree chooses an annuity that begins at some distanced future point of time and lives from his AFP savings until then, could be interpreted as an insurance against surviving one’s own retirement resources.

The results have been impressive. In the last 25 years, AFPs have earned annual real returns of 10% on average. Financial market volatility has not been a major problem because AFPs have continuously diversified their portfolios. Gradually, the capital market has grown in size and diversity, investment possibilities open to the AFPs have been liberalised, and AFPs have made use of these opportunities. Neither the Mexican nor the Argentine crises have been able to do serious harm to the pension funds, and neither have turbulences in the pension market itself. The number of AFPs had risen from 12 at the beginning to 21 in the mid-1990s and then dropped back to 6, but workers accounts have not been affected.

Savings and Investment

Economic theory suggests that moving to a funded pension scheme will permanently raise the level of savings and investment. Indeed, each year, Chileans pay about 4.6% of GDP into retirement savings accounts. But is that really a net increase in the savings rate? Or has the introduction of pension saving simply squeezed out other types of savings? Bennett et al (2001) and Coronado (1997) have econometrically modelled private saving behaviour in Chile, to examine whether a substantial offset has occurred. Neither has found reliable evidence that it had. ‘Saving’ is not a monolithic block, but serves specific purposes. It is unlikely that reserves intended for, say, the purchase of a car, will be liquidated to be transferred to a pension account. Corbo and Schmidt-Hebbel (2003) have shown that about half of the new savings have translated into additional domestic investment, thus enhancing economic growth.

A commonly heard argument against a reform such as the Chilean one is that it necessarily produces a transition deficit. One generation, it is argued, pays twice. They must save for their own retirement, and support the retirees of the old system through their taxes.

This argument deserves a closer look. What is actually meant by “transition deficit” is the process of transforming a debt hidden from government accounts -the promises made to future retirees- into a visible debt. This creation of public awareness is likely to exercise pressure upon political decision makers to reduce spending and sell off assets. In this way, the transition deficit might have an inbuilt self-containing mechanism.

While that effect is, of course, not measurable, the Chilean case makes it at least plausible to assume that it plays a significant role. In 1981, a transitory tax on income of 3% was introduced, but it decreased automatically by one percentage point each year and thus disappeared in 1984. There was also a global budget deficit in the five years following the reform, but it was succeeded by twelve years of budget surpluses. While it is unknown how taxes and debt would have behaved in absence of pension reform, at least they have not generally increased. In contrast, government expenditure contracted. Its share of GDP was 32% from 1974-84, but 25% from 1985-99. The period from 1985-89 also experienced the greatest wave of privatisations in the regions’ history, comprising predominantly areas such as telecommunications and electricity. Government assets worth 7% of Chilean GDP were sold off, more than twice the relative scope of British privatisations in the Thatcher era.

A Model for Economic Transformation

One target of the process of privatisation in Chile was to spread property ownership throughout society, in order to make the market reforms of the previous years irreversible. A programme named Capitalismo Laboral encouraged workers to purchase shares in their own company, which turned 30,000 workers into co-owners of their own workplace. Another programme named Capitalismo Popular favoured small investors and turned 170,000 Chileans into shareholders. But it was the sales to the AFPs, denominated Capitalismo Institucional, which transformed 3 million workers into small-scale capitalists. Thus, 18% of all privatised assets were directly purchased by AFPs. Investment in the affected companies soared. Until 1990, the number of telephone lines in use almost tripled, and energy production went up by a third.

There are reasons to assume that the way in which the pension reform was designed had its share not only in encouraging these privatisations (as explained above), but also in explaining its tremendous and lasting success. The quality of Chile’s judicial institutions was still mediocre even by Latin American standards in 1985. In the absence of sufficiently developed legal institutions, privatisations are often vulnerable to abuse and self-enrichment. The involvement of expert investors representing millions of people might have been a strong check which enforced greater transparency.

But most of all, privatizing state-owned companies after the creation of the AFP system established a nation of property owners, a stabilizing factor in the new market economy. This idea can best be illustrated by comparing Chile to Venezuela, which had for a long time been regarded as Latin America’s success story. In recent years, exactly those sectors privatised in Chile in the 1980s have been nationalised in Venezuela. Such a roll back would be impossible in Chile, re-nationalisation of these sectors would imply expropriating the retirement funds of millions of workers.

Labour Market s
PAYGO systems and defined-contribution, funded, schemes differ substantially with regard to their effects on labour markets. Social security contributions in a defined-benefit system are mostly perceived as a tax, and taxes on labour discourage work. The same is not true for pension savings, because people pay them into their own accounts, which is, to themselves.

There is a second effect. In a transition country such as Chile, the labour market (as in the economy as a whole) is usually divided into a formal and an informal sector. Lacking legal security, these “extralegal sectors” will tend to be less productive. Cutting taxes on formal employment will lower incentives to tax-evading informal employment. Some people will switch from the informal to the more productive formal sector, changing the composition of the economy in favour of the latter. An AFP system can thus increase employment and labour productivity at the same time.

Some have argued that a PAYGO system, too, could be designed in such a way that efforts and rewards are closely connected, so that people would perceive their contribution rates as a quasi-investment and not as a tax. In that case, there would not be a negative effect on labour market participation. But the true reason why PAYGO contributions are not really distinguishable from a tax is that their contributors do not acquire property rights, which they could contractually enforce. Their entitlements are a mere promise of a diffuse nature and will only be made concrete through the political process.

Edwards and Cox Edwards (2002) have found empirical evidence for a positive effect of the pension reform on formal employment. They developed a model simulation of Chile’s labour market, dissecting the components determining its shape. Expanding this model, Corbo and Schmidt-Hebbel show that the pension reform has increased formal employment by 3.2% under the most pessimistic assumptions and by 7.6% under the most optimistic.

As in Chile, an AFP system can be combined with a means-tested minimum safety net, a much more efficient and transparent instrument to combat old-age poverty than the redistribution inherent in a PAYGO formula. PAYGO systems tend to redistribute large amounts of money, but the distributional net effects are opaque. A tax transfer, in turn, can be a targeted instrument, exclusively directed to the focus group of those who fall below a certain income level.

In Chile, there are two types of old-age income supplements, a minimum pension guarantee for people who have contributed for at least 20 years and still fall below a certain level, and a non-contributive assistance pension that ensures subsistence to the poorest. Old age poverty has fallen from 30% in the late 1980s to less than 10%, whilst both forms of payments to the old poor cost about 0.5% of GDP combined. This is much less than government subsidies to the old system pension system. Spending efficiency could be even greater if these two instruments were integrated into one.

Capital Markets

Financial institutions represent a very special economic sector. The ability to channel capital to those points of use where it is most productive is a prime determinant of a nation’s productivity. Creating fully funded pension systems can be expected to be a great stimulus for capital market development. Demand for diverse long-run saving products increases, and professional institutional investors come into play. Chile’s development in the last quarter-century seems to confirm this intuition. Capital markets have greatly expanded in size and diversity. In the early 80s, banks were practically the only actors in Chile’s capital markets. Today, a wide variety of institutions such as mutual funds, real estate funds, infrastructure funds, and, of course, the AFPs themselves, have emerged.

Alongside, the nature of economic growth has changed. Economies grow if the amount of labour or the amount of capital grows. But the most decisive long-run source is so-called Total Factor Productivity (TFP), the efficiency in the interplay of the two factors. TFP-growth has virtually been nonexistent in the past, but became a major source of economic growth after the mid-1980s. This is the major change that has occurred in recent Chilean economic history.

A good measure for the degree of capital market development is the Financial Intermediation Ratio (FIR), the sum of the most relevant assets traded in financial institutions. Corbo and Schmidt-Hebbel have found econometric evidence that the pension reform has greatly raised FIR, and that FIR has greatly raised TFP. An increase in TFP directly translates into an increase in economic growth.

Would it Work in Europe?

Not all of the merits of the Chilean reform would equally apply to Europe. Long-term real returns of 10% could not be achieved here. Promoting savings and financial markets would be desirable, but as our economies are already highly capitalised and our financial markets highly sophisticated, the boost to growth would be much weaker than it was in Chile. At the same time, the transition deficits would be a lot larger in Europe, both because our PAYGO systems usually comprise the whole labour force, and because they provide not only a minimum living standard but often fairly high replacement rates.

But at the same time, the pressure from an ageing society is far greater in Europe. While a rising life expectancy raises the cost of old-age provision in any pension system, and while funded schemes are not fully independent of demographics either, the adaptability of a funded scheme is enormously greater. Two extreme thought experiments illustrate this point:

Ø What would happen if the entire young generation of a country decided to leave it (and no immigrants would follow)? If it happened in Europe, the older generation would be left to either work until their last day of life, or to starve. If it happened in Chile, the retirees would have to transfer the whole capital stock abroad, because there would be nobody left at home to use it. Maybe people would not be happy with it because they would presumably prefer to have their old-age assets geographically close to them. But even in this extreme case, old-age income would be provided by the returns of the investments abroad.

Ø What would happen if a number of miraculous pharmaceutical innovations (gradually) raised the average life expectancy to 100 years? In Europe, fierce political battles over the distribution of the cost burden would break out. Young entrants to the labour force would be unwilling to spend the lion’s share of their income on sustaining the older generation, retirees would be unwilling to see their living standards fall, and older workers would be unwilling to see retirement age raised. As no government would be likely to touch what voters regard as their deserved “entitlements”, the adaptation would surely be delayed and the additional cost financed by debt. If it happened in Chile, the additional burden would be large too, but the adaptation process would be entirely different. The price for the purchase of a lifetime annuity, which represents an “exchange rate” between a stock of savings and a life-long stream of income, would rise. This would act as a signal to all workers, even if they would not actually purchase an annuity themselves. Young people who would observe this rate continually worsening would see an incentive to save a larger proportion of their wage for old-age. To workers close to retirement age, additional working years would suddenly seem relatively more attractive than at the old “exchange rate”. So while people would certainly not be happy about having less money for present consumption and working longer years, the adaptation process would begin sooner, run more smoothly, and take account of differing preferences. It would not involve political battles between different age groups, as it would be obvious that the change was not the result of any deliberate political decision. An arbitrary political decision process would be replaced by a self-steering market mechanism.

Meanwhile, if Europe adapted a Chilean-style reform, it could easily avoid most of the difficulties that the Chilean reform had to face:

Ø The Chilean reform was a risky experiment in the sense that, as was explained, capital markets and the institutional infrastructure had to be developed in parallel. At the same time, many people had no experience with any type of financial institutions. All of this would hardly play a role in Europe, where fully developed capital markets, sophisticated regulatory institutions and a high degree of financial literacy can simply be taken for granted. Some countries even have complementary pension funds already in place.

Ø As there is no noteworthy “informal sector” in Europe (moonlighting is not the same thing), universal participation could easily be guaranteed. Even the unemployed could be covered if a share of their benefits were transferred to the recipients’ retirement account, immunising the system against frictional imbalances in the labour market.

Ø In 1980, the share of government spending in Chilean GDP was about 27%. In many European countries today, that share is almost twice as high. So there is definitely room to finance the transition deficit, especially in those countries with large public sectors.

So, many of the positive economic side-effects of the Chilean reform would be less pronounced in Europe. But so would the initial problems. After all, the role of a pension system is not to boost the economy, but to provide a reasonable and stable old-age living standard. In that regard, Europe would need a Chilean-style reform far more desperately than Chile did.

Bismarck vs Piñera? The importance of the framing of the debate for Pension Reform

Why, then, is a Chile-type pension reform so difficult in (Western) Europe? Political-economic decision models usually argue that a transition creates winners and losers, but these models are unconvincing. Pension privatisation can be a pure gain. The market interest rate is usually higher than the indirect yield of a PAYGO system, the capital stock in an economy with a funded scheme is permanently higher, and the transition deficit partially offsets itself. In short, this reform creates the means by which those who lose from it can be compensated. The trouble with these models is that they reduce the decision process to a weighing of economic variables, while any observer of everyday politics knows that social policy issues are not debated in this way. They are more of a battle of ideas. A historical retrospect is enlightening because it shows that they have been so since their modern origin.

The debates that accompanied the introduction of Otto von Bismarck’s welfare state in the 1880s were not technical, but ethical in nature. Chancellor Bismarck had spoken of “dividends from human tragedy” when referring to private insurance. Social insurance issues should “from the moral viewpoint not [be a] subject matter of speculation and a source for the payout of dividends”. His premise was that government action, aspiring to the “common good”, was ethically superior to private actions driven by base motives.

No less founded in ethical value judgements were the arguments of the adversaries of Bismarck’s welfare state, led by the Reichstag deputies Eugen Richter and Ludwig Bamberger. Since the late 1840s, numerous mutual self-help initiatives were emerging in the German states, on a voluntary and decentralised basis, covering a number of risks such as accidents or illness. They were organised as cooperatives or were run by labour unions, and provided an alternative to commercial insurances. Richter and Bamberger feared that compulsory monopoly insurances would bleed these voluntary initiatives to death - and assumed that this was even intended by Bismarck. In some of these associations, people developed a spirit of loyalty and personal belonging. According to his critics, it was Bismarck’s very intention to weaken this culture of social self-organisation, for being out of reach of government control, and replace it with a culture of state-dependency. Workers should be bribed into projecting the mentioned loyalty upon the state instead. Ludwig Bamberger asked whether “human individuality, self-determination, the free initiative of the responsible citizen” were to be replaced by “the supervision of the police and the caring hand of the state“. Eugen Richter indicted that workers should “learn to regard themselves as state pensioners”, in order to “develop the sentiment of dependency” typical for these.

In this view, the welfare state does not appear as a humanitarian institution, but as an instrument of power and control. Detmar Doering (2006) finds proof for this perspective in the fact that the welfare monopolies of the 1880s did not cover those risks for which voluntary insurance mechanisms were indeed underdeveloped, such as unemployment. Instead, they covered –surprisingly- those areas where private initiatives were already relatively advanced. Richter and Bamberger embraced the voluntary alternatives because of their spirit of independence and self-determination. For the specific issue of old old-age provision, the alternative they held against Bismarck’s monopoly was “everything which is appropriate to facilitate and boost savings, capital accumulation, land acquisition and homestead” (Richter).

A hundred years later, though in a totally different historical environment, the arguments of then Labour Minister Jose Piñera sounded astonishingly similar to those of Richter and Bamberger. Piñera had never spoken of the pension project in mere economic terms, and never made high capital returns his principal point. When publicly announcing the reform on November 6th, 1980, he said it would “drastically widen the margins of individual liberty”, that “an enormous source of state power and discretion” would be eliminated, and that it made “every worker an owner”. Richter and Bamberger would have been delighted.

State-run PAYGO might, in many countries, perform in a mediocre fashion or worse, but in (Western) Europe it seems Bismarck and his heirs have won the battle on moral terms. Friedrich Hayek had once said that “the ultimate decision about what is accepted as right and wrong will be made not by individual human wisdom but by the disappearance of the groups that have adhered to the “wrong” beliefs”. This is exactly what has happened. At least in continental Europe, positions such as those of Eugen Richter and Ludwig Bamberger have simply disappeared. Concerning old-age provision, criticism of PAYGO systems’ vulnerability to demographic change and of their high payroll tax burden may be heard. But Bismarck’s concept of a virtuous, all-benevolent government is still the unchallenged implicit premise behind social policy debates. Practical issues must also play a large role, for example, public fears that funded pensions were “unstable” need to be addressed.

But anyone who tries to win a debate which is primarily of an ideological nature with technical arguments only has lost right from the start. Supporters of private funded solutions must not abandon the realm of moral arguments to their competitors. Like Richter, Bamberger and Piñera, they must emphasize the differences of principle that distinguish a mechanism based on individual property accumulation, secured by enforceable contracts and independent of personal discretion, from a system based on dependence on political bodies, all of whose variables are and can only be decided in the sphere of political power.

It is sometimes said that European systems are run by the government, while the Chilean and similar ones were ‘run by the private sector’. This is a superficial perception. A mechanism of private savings accounts is not really run by anybody; it simply runs and steers itself. Europe needs to be reminded that this is the only form of provision fit for a society of free citizens.



2010 © www.josepinera.org