Abstract
The welfare state and old-age pensions are central pillars in modern societal and economic thought. Doubts about the future sustainability of the national pension system have prompted reforms which must contend with sensitive economic, demographic, social and political issues. This article reviews in depth the role of the state and private actors in pension systems. It suggests wider participation of the private sector in current pension systems coupled with well-considered policy regulation.
Keywords
The modern idea of a ‘welfare state’ constitutes one of the most important achievements of European political thought. Old-age pensions, in turn, are a crucial feature of a modern welfare state and a modern social market economy. Pensions are the main tool to provide citizens, on a large-scale basis, with acceptable welfare levels at retirement.
In the face of this, demographic trends initiated in the past and now showing major effects are casting serious doubts on the future sustainability of national pension systems as they are organised today. Countries worldwide are dealing with increasing budgetary problems in paying out pensions due to population ageing and, especially in Europe, to a historical tendency towards early retirement. Awareness of this phenomenon has triggered a widespread debate in all Member States of the European Union and throughout the world about reforming national pensions systems. Currently, almost every OECD country and EU Member State is debating the pension problem and undertaking its own reforms. 1
However, pension system reform is a ‘sensitive target’, because of the very nature and importance of pensions. There are economic, demographic, social and political aspects that must be considered. Furthermore, from a technical point of view, there is controversy about the empirical magnitude of the pension problem and the underlying economic theory and, consequently, about best practices and policy advice for national governments. It is, therefore, no surprise that the debate on pension reform has grown intensively and now spans a variety of dimensions: from identifying appropriate actuarial formulae for computing benefits at retirement, to the general economic impact of different pension designs, to the more political issues of social consensus and implementation.
Pension reform and the associated debate are also taking place beyond the boundaries of the EU-27 and OECD, a primary example being the debate in China.
A central theme of the current debate, and the theme of this article, is the respective roles of the state and the private sector in organising pension provisions for the future. The following section briefly describes pensions from an economic point of view, and defines what is meant by public and private pension schemes. This is essential for the remainder of the paper and to understand the ongoing debate on pension reforms. The third section discusses the current problem of fiscal sustainability of public pensions. The fourth section discusses how increasing the private sector's participation in pension provision can help to achieve the sustainability of pension systems in view of the current demographic tendencies, and elaborates on the effect of private pensions on savings and growth, on financial markets and on individual protection against various types of risk. This section, takes a cautious position in favour of expanding the contribution of the private sector to current pension provisions, arguing that this can be a successful strategy provided that the transition is coupled with careful policy interventions fostering regulation of financial markets. Advice on such regulatory issues is given in the fifth section. The position of this article is not to call for more regulation; rather, it is to favour the revision of regulations in a direction that accounts for the specific characteristics of pension plans.
What is a pension scheme?
A pension is a contract under which an individual acquires the right to be transferred quotas of future output (the benefits) in exchange for quotas of his current output (the contributions). The contract establishes a point in time–-retirement age–-when this transfer starts.
In principle, an individual could ‘sign a contract with himself’; that is, he could store part of his own current production for future use. However, this is not viable for services and non-storable goods, for example health care and services for the elderly, clothing and food; these are precisely the basic kinds of good and services one is interested in after retirement.
In face of this physical constraint on individual actions, modern societies have coordinated individual pension contracts into a pension system, where individuals do not need to consume their own past production during their retirement time, but instead, consume output produced by the individuals working actively during that time. Every modern pension system constitutes essentially a transfer of output, in each period, from the current young generation–-meaning here as those actively working–-to the current old generation–-those who are retired.
Members of the current old generation can finance their claims on current production in different ways, and this constitutes a fundamental dimension along which pensions systems differ. Relative to the financial mechanism with which pension claims on current production are organised, pension systems fall theoretically into two broad categories.
In a pay-as-you-go system (PAYG), the current young generation contributes mandatorily to the pension system through labour income taxes, whose revenues are partly transferred to the current old generation as pension benefits. In doing so, the young generation acquires the right to receive pension benefits when they grow old. This system mimics explicitly the intergenerational transfer of resources embodied in any pension system.
In a funded pension system, the members of the current young generation mandatorily contribute part of their labour income to a pension fund, which invests in financial assets. When they retire, the fund pays out pension benefits in the form of lifetime annuities, which finance the pensioners’ consumption of goods and services produced by the current young generation. 2
Funded schemes can also be based on voluntary contributions. All the analyses of this paper apply to voluntary contributions as well, while a main issue in the current pension debate is about the opportunity to have mandatory funding.
Hence, in a PAYG system, pensions are paid out of current income, while in a funded system they are paid out of the fund asset accumulation. It is worth mentioning that PAYG and funding are only different financial mechanisms to finance the intergenerational transfer of production from the current working generation to the current retired generation.
PAYG systems are managed by the state, which taxes the workforce and coordinates the intergenerational transfers. Virtually all state pensions are PAYG. On the contrary, funded schemes are generally run by private institutions, like insurance companies, large firms and industry sectors. 3 For this reason, the debate about organising pensions in the future as either more PAYG or more funded becomes of a debate about public versus private pension schemes. 4
Theoretically the state could also organise and entirely manage a funded pension system. The point is, then, to asses whether the system can be better run by the state or by market institutions, This is important, but must be addressed in a future article.
Other than in the way they are financed, i.e., PAYG or funded, pension arrangements differ crucially in the relationship between contributions and benefits they embody. In this respect, a commonly used taxonomy distinguishes among pensions with defined contributions, with defined benefits and with notional defined contributions. Barr and Diamond [3] provide for the basics of this distinction. This article pursues a wider approach and does not need to specify contribution-benefit relationships.
Demographic change and public pension imbalances
This section discusses the problem of the current fiscal sustainability of public pension systems. A PAYG system is fiscally sustainable, or financially balanced, when in each period, the total nominal contributions–-given as the contribution rate times the average wage times the number of workers–-equal the total pension expenses–-given as the average pension times the number of pensioners [1]. 5
The balancing formula is more complex than this, including in particular the possibility for the public sector to lend and borrow plus a time variable The simple version used here is enough for discussion and avoids technical details.
In OECD countries, public spending for retirement pensions–-that is, the right-hand side of the balancing formula–-is expected to rise dramatically in the coming decades. Recent projections for the period 2000-2050 show that public age-related expenditures, as a percentage of GDP, may rise on average by about seven points. In particular, the proportion of GDP represented by public spending for retirement pension and early-retirement programmes is projected to increase by almost 10% in Norway, by 8% in Spain and Korea, by 5% in Germany and by almost 4% in France [4, Graph 3].
For the European Union, ECOFIN [5] projects an increase of the gross (before taxes) public spending in pensions of 2.3 and 2.2 percentage points of GDP, respectively, for the EU-15 and EU-25. This implies that by 2050, the EU-15 will spend almost 13% of its GDP to finance public pensions, while the EU-25 will spend 12.8% [5, Table 3.3]). Compared with 1990 levels of spending on old-age pensions–-generally the greatest source of public pension spending–-this consists of a rise of more than 5 percentage points with respect to GDP for the EU-15 [10, Table 1]).
The main reasons for these increases in public spending are demographic in nature, namely the confirmation of the tendency towards an increase in longevity and a decrease in fertility rates. 6 The demographic forces at work are captured by the projected trend of the old-age dependency ratio: the number of people aged 65 or more over the number of people aged 15-64, the latter being a conventional measure of the consistency of the working population. Eurostat projects an old-age dependency ratio increase from 25.4% in 2008 to 53.5% in 2060 for the EU-27. 7 This means that if in 2008 there are four persons of working age for every person aged 65 or over, by 2060 this figure will drop to two persons of working age for every person aged 65 or over [6].
See ECOFIN [5, Table 3.10]. where the contribution of these demographic factors to the change in public pensions spending relative to GDP is assessed.
In particular, the working population is expected to decrease by 50 million persions, while the elderly are expected to increase by 67 million persons.
The projected increase in longevity, combined with a tendency towards early retirement, is responsible for the expected increase in the expenditure for retirement pensions. At the same time, projected pension contributions are expected to fall because of the decreasing fertility rates. Indeed, lower fertility rates tend to lead to a decrease in the working population and, under reasonable assumptions about productivity, to only a moderate expansion of GDP. As a result, given the current public pension formulae, the system will experience serious fiscal imbalances, with an expected increase in the quota of GDP transferred from the young working generation to the retired old one.
To achieve balance in their public pension systems, most OECD countries have reformed or are in the process to reforming their systems, either in isolation or within the revision of the whole system of social security. There are, however, only three ways to correct a PAYG system under the stipulation that no increases in public deficit are allowed: increasing contributions, decreasing pension benefits or increasing retirement age. Virtually all ongoing public pension reforms are indeed introducing mixtures of these corrections. This is sometimes accomplished using the complexity of the actual pension systems to minimise the political cost of such reforms. Furthermore, besides reforming public systems, many countries are considering introducing or expanding the role of funded pension provisions. 8
See Whiteford and Whitehouse [10] for a description of ongoing pension reforms in OECD countries, and ECOFIN [5] for the same with respect to EU-27.
The World Bank has suggested that reforms should proceed with mixed systems of public and private pensions, along a ‘multi-pillar model’ [11]. At its simplest, a three-pillar version of this model has the following structure: a first pillar formed by a pay-as-you-go system, including some redistribution and a poverty relief scheme; a second pillar mainly based on mandatory membership in a privately or state-managed, fully funded scheme (e.g., an occupational pension scheme); and a third pillar based on voluntary membership in a privately managed, fully funded scheme.
Public versus private pension schemes
This section discusses how increasing the role of funded pension schemes can help to maintain fiscal sustainability in the presence of population ageing. It also elaborates on the effects on private pensions, savings and growth, financial markets and the protection of individuals against various types of risk, making comparisons with a PAYG system when necessary.
Funding allows the system to be kept financially balanced in the face of population ageing. In a fully funded system each individual in a given generation, call it ‘generation A’, builds up her own ‘pile of financial activities’ during her working life. At retirement, the accumulated activities get transformed into a pension annuity whose actual value exactly equals the value of the accumulated assets for each individual. Under population ageing, during the retirement period of ‘generation A’ the working population will decrease. However, every individual in ‘generation A’ has already accumulated his pension fund before retirement. Hence, the pension system has by design enough ‘money’ to pay out A's pensions. This argument is valid for each generation, and thus implies that a funded pension system, contrary to a PAYG, remains financially balanced in the presence of population ageing.
As pointed out in Barr [1] and Barr and Diamond [3], the fact that a funded system stays financially balanced does not mean that it can keep invariant the real levels of pension annuities in presence of ageing. Their argument runs as follows. The decrease of the working population implies–-other things being equal–-a decrease in output. Hence, when ‘generation A's’ pensioners use their annuities to buy good and services, an excess of demand may arise, leading to an inflationary process which lowers the real value of A's pension annuities. 9
This is an equilibrium argument that applies in particular to the case where the accumulated financial activities are monetary, for example, government bonds. A similar argument applies in the case of equities accumulation.
While this argument is correct, it is a somewhat minor concern. First, the funded pension system remains nominally balanced. Most importantly, financial imbalance may cause the system to break down, and this is the main concern which has triggered the pension debate. Second, Barr and Diamond's argument is essentially about economic growth. It is the lack of growth, due to the decrease in the working population, which lowers real pensions. But growth is a general problem that affects both PAYG and funded systems, as well as the economy as a whole. Even if it is not known whether a funded system is always more effectual than a PAYG system in fostering economic growth, both PAYG and funded system are surely fragile in the presence of a lack of growth, but a funded system at least remains financially balanced. These arguments are seen as ones in favour of extending the role of funding in pension provision.
Funding may increase savings
As mentioned in the previous subsection, it is not obvious that savings, and thus economic growth, are higher with funding schemes than with PAYG. 10 Of course, contributing to a pension fund means saving. But, as pointed out in Barr [1], mandatory contributions to a fund may reduce voluntary saving, for example, because a funded account is a stronger parallel to voluntary saving. In this case, the effect on total savings may even be negative. Furthermore, as discussed in Barr and Diamond [3], a shift from PAYG to funding schemes decreases tax revenues due to diminishing mandatory contributions to the PAYG system, and this may lead the government to borrow more, again with ambiguous effects on savings.
The link between savings and growth is itself not trivial. Conditions must be met for savings to transflate into new investments, hence into growth. While this issue is far from the topic of this article, we elaborates on it below.
These arguments are correct and relevant. However, while the second is sound, the first may not be totally convincing. The amount of income that an individual is willing to contribute for pension purposes may be decided independently of how the pension system is financed, whether PAYG or funded. If this is the case or, alternatively, if a relevant number of individuals are not affected by the purely financial aspect of the pension system in their contribution decision, then once the desired amount of pension contributions has been set, a PAYG system transfers these contributions directly to the current old generation, which uses it to consume, while a funded system keeps the contribution in the form of savings, making it eventually available to financial investments (assuming the correct functioning of financial markets).
In conclusion, while the link between funding and savings is not obvious, funding may increase savings, while PAYG generally does not.
Funding ameliorates capital markets
This is a somewhat less controversial point, albeit one that is questionable in some cases. Funding will channel more income into financial activities, and this dimensional effect is generally beneficial: risk can be better spread over a broad market than a narrow one, and a broader financial market implies more efficient resource allocation and hence more growth.
However, a well-functioning financial market needs attentive government regulation. This article elaborates more on this in the sixth section, ‘Towards private pensions’.
Funding, PAYG and risk and uncertainty
As with any financial contract, pension contracts have to deal with several sources of risk. It is crucial, therefore, to evaluate the effects of different pension schemes on the distribution of risk; in particular, who will bear the risk of longevity and earning losses (due to job loss or to different wage dynamics over his working life). In principle, the risk should be mostly transferred to the agent that has the lowest risk aversion: if the private fund, say, one managed by an insurance company, is risk-neutral with respect to the specific risks considered, it is efficient that it bears all risk.
In general the private fund does not bear risk from earning losses, which is–-in the absence of other insurance mechanisms–-fully translated to the pensioner. A pension subscriber receives pension benefits which vary depending on his earnings, contributions, retirement age, life expectancy and the interest rate that the fund management expects to earn over the lifetime of the annuity. Uncertainty over these last two variables leads to ‘annuity risk’ [1]. Instead, longevity risk is borne by the insurance company under a purely defined-contribution scheme.
Economists sometimes make a distinction between uncertainty and risk. The idea is that not all the uncertainty is translated into risk. Individuals may not be aware of some contingencies, or may just not be able to evaluate their likelihood.
There are sources of uncertainty and risk that affect both funded and PAYG systems, such as macroeconomic and demographic shocks and political risks. But there are also risks which affect private pensions but not PAYG systems, the most immediate being the risk of low return on the financial activities in which contributions are invested. 11 There is also a class of risk called here contractual risk, which is typical of funded systems.
Contractual risk
Most contractual risk originates from the fact that the information available to parties at the various stages of the pension contract is asymmetric. For example, the pension subscriber is not perfectly informed about the investing strategies of the fund management, and the fund management has less information than the pension subscriber about his relevant characteristics, such as his life expectancy. The first case represents what is referred to in economics as a ‘moral hazard’ problem. More generally, this is a contractual risk arising from the possibility that one of the parties may execute the contract choosing different actions, some of which may negatively affect the other party. The second case is an ‘adverse selection’ problem. More generally, this is a contractual risk typically emerging at the time the contract is signed because one party has less information on some of the other party's characteristics, which may materially affect the future outcome of the contract.
An instance of these market imperfections can be seen in occupational pension schemes. Moral hazard and adverse selection, for example, seem to explain why employees stick to the pension offered by their employer, even if not mandatory, rather than choosing the scheme offered by a third party. First, in fact, the employer has privileged information on his employee, thus mitigating adverse selection. He is therefore able to offer a pension scheme at lower costs than third parties can. Second, the employee may be biased towards his employer's offer because he knows him and can obtain information on his general activity at a lower cost. 12
Obviously, this last point is correct only if the outside funds are otherwise equal in terms of risk-return characteristics to the one offered by the employer.
Observe that some forms of contractual risks may show up also in a PAYG public pension system, but to a lesser extent than with a private funded system. For example, the adverse selection problem is relevant in a PAYG system as well, but in a less dramatic way due to the greater number of people involved in the same pension scheme. 13 The moral hazard in a PAYG system exists too, but is even less dangerous since in general the features of public plans are established by law.
If all people are enrolled in the same pension scheme, their life expectancy can be estimated looking at the life expectancy of the whole population, which is easily obtained from demographic statistics
Another form of risk coming from asymmetric information in a funded system is the management risk. The return on contributions depends on how contributions are invested. Management risk is the risk associated with bad investing strategies by the fund. The fund management may be incompetent, or may be plagued by conflict of interest. For example, with occupational pension schemes an employer may intentionally use contributions to overinvest in his own equities, which–-being positively correlated with the corporation results–-will tend to increase the risk of earning losses: in the extreme, if the corporation defaults without insurance on the pension scheme, the employee loses his job and his pension (as in the Enron case).
Uncertainty is also more pronounced with private pension contracts. Sticking with mandatory pensions–-management risk–-is, for many individuals, management uncertainty. As the recent Swedish experience showed, it is difficult for employed workers to evaluate a wide menu of different pension products simply because they are not able to evaluate their uncertain characteristics [9]. For pensions, this problem is even more dramatic than it is for other financial assets, because
pension schemes are highly illiquid assets; and
pension subscribers often have no previous experience with this type of financial instrument, and will gain such experience only later in life, possibly at retirement.
Towards private pensions: government regulation
According to the previous section, a funded system, compared to a PAYG system, keeps pension provision financially viable in the face of population ageing and could help increase saving and growth, and can make financial markets more efficient.
These are important arguments in favour of funding. But we have also seen that asymmetric information means private pensions bear an entire class of risks, the contractual risks. This problem requires that a move to funding be assisted by a careful effort to regulate financial markets.
This section discusses financial market regulation and identifies some specific directions of intervention. This article's position is not one that calls for ‘more’ regulation. Rather, it favours a revision of current regulations in a direction that accounts for the peculiar characteristics of pension plans.
To start, consumer protection from contractual risk is important. It is precisely a regulatory system that should avoid a situation in which consumers bear ‘too much risk’, especially of a type they are often unable to value, as opposed as to risk transferred to intermediaries or financial institutions. This is even more important for pension schemes because (1) pensions are highly illiquid assets; and (2) subscribers face a good deal of uncertainty at the time of subscription, and by the time they experience the product they have chosen it is probably too late to switch to another one.
Regulation may mitigate these two common aspects of private pension schemes. The degree of liquidity of a pension scheme is partially due to the regulatory system, and can be increased. An example is the recent experience of mortgage contracts, which in some countries can be renegotiated at reasonable cost. For pension schemes, renegotiation may also include the possibility to switch funds and their providers. Clearly, this would call for a sufficiently high degree of standardisation of pension plans. Regulatory interventions in the direction of higher standardisation may also help to improve individuals’ knowledge of pension products, thereby reducing their uncertainty. The recent Swedish experience with over 3,000 funds taught that too much variety may be undesirable [9].
Nugée and Persaud [8], agreeing on the importance of protecting pension subscribers, suggest that a regulatory system should target consumer protection as opposed to the financial stability and soundness of pension institutions. The heart of their argument is as follows. To protect consumers against risk of provision of private pensions, the system should effectively regulate and supervise the characteristics of the pension fund products and activities. A third party, a public authority or private rating agency, could certify that pension products belong to a certain category, meeting some standards. All other financial products and activities not directly sold to consumers may in principle be left unclassified and unregulated.
On the contrary, current financial regulation in the spirit of Basel II (Revised International Capital Framework) is now largely based on imposing accounting standards and principles on financial institutions. Monitoring often reduces to requiring pension funds to provide ‘fair value’ on their current financial assets and liabilities at pre-specified dates. Yet these static financial evaluations may largely distort the picture of firms’ investment strategies and risk management [7]. Indeed, because pensions are highly illiquid, fund managers should overweight long-term assets, whose return pays a liquidity premium over short or medium term. Yet under the current prevailing accounting standards, this would imply a higher volatility of liabilities, forcing managers to choose between either more liquid assets or a higher solvency reserve position. Since both strategies lower the return on contributions, managers have often tried to increase returns by increasing the proportion of equities in the portfolio. 14 This has precisely the effect to increase that type of sector or entrepreneurial risk which pension funds and subscribers should not be exposed to. It is essential that the employer fund does not bear this entrepreneurial risk; thus, either risk management strategies should avoid this or the pension fund should be insured against entrepreneurial default, as is the case, for example, in the US and UK.
Danish Central Bank data show that the proportion of equiteies in pension fund portfolios grewe from 10% in 1990 to 40%.
A regulation that requires the classification of pension products can protect consumers without distorting markets, and requires a much lighter and thus less costly monitoring system. An independent public authority (or private rating agencies) could analyse and classify those products which pension funds or other intermediaries propose to offer to consumers. This is a fairly simple task with respect to assessing a correct risk-management approach in each and every pension scheme, especially if some product standardisation is introduced. Product regulation should enhance or at least not discourage financial innovation and the entry of new firms. With respect to financial innovation, it is important that the classification/standardisation and supervision process not be too tight and detailed for the firm. This could be achieved, for example, by establishing some minimum standards in terms of plan characteristics (first, in terms of return and risk) for each step of the classification.
This approach to regulation, based on consumer protection, can also cope with systemic risk. Financial institutions belong to an industry with a high level of interconnectedness. Banks, mortgage institutions, pension funds and other financial operators are tightly bound by a thick and complex web of lending and borrowing relationships. These relationships often involve firms of different countries. The failure or default of some of them may strongly destabilise the others, sometimes causing default chains. The effects of a run on a bank during the Great Depression, the Asian financial crises in the 1990s or the current mortgage crisis in the USA are well known.
Systemic risk is even more pronounced if financial operators tend to assume similar risk profiles and risk management strategies. In that case, one person's ill luck becomes everyone's ill luck. By recognising the differences between financial institutions, product-oriented regulation would have the positive effect of sustaining and giving incentive for different risk-management strategies, with the pension funds investing long-term in more illiquid assets, banks bearing more standard forms of credit risk, and so on.
Conclusions
This paper take a cautious position towards a wider participation of the private sector in current pension systems. It argues that this may be a successful strategy provided that the transition is coupled by well-considered policy interventions fostering financial regulation. The position is not one that calls for more regulation, rather one that favours the revision of regulations in a direction that first and foremost accounts for the peculiar characteristics of pension plans. This could be successfully accomplished following a regulatory approach that targets product characteristics and consumer protection; which, we argue, could be more effective and easier to implement than targeting pension funds directly.
Here, we want to point out that whenever a pension system has a strong redistribution objective–-in particular, if referred to intergenerational redistribution–-then there is room for a PAYG component as well, possibly endowed with correction devices. In this respect the public sector has more financing opportunities than any private institution. On the other hand, the rest of the institutions which, according to the multi-pillar model, are part of the pension system should be financially sustainable and targeted for efficiency, and a funded system seems more appropriate in this respect.
Footnotes