POLICY RESEARCH WORKING PAPER 1229 The Structure, Regulation, Comipany pension funds can make important contributions and Performance to retirement income and to o Pension Funds capital market development. But they need to be regulated in Nine Industrial Countries andsupervisedtoavoid fraud; protect the interests of workers, and minimize E. P. Davis restrictions on labor mobiiity. The World Bank Financial Sector Development Department December 1993 POLICY RESEARCH WORKING PAPER 1229 Summary findings IDavis offers an overview of issues relating to the consequences for econormic efficiency. Though countries development of funded pension schemes in industrial differ widely in their regulation of pension futnds, some countries. The analysis applies the economic theory of suggestions for good practice canl still he made. pension regulation to experience with the structure, * Whether pension funds are a cost effective way of regulation, and performance of funds in nine countries providing pensions depends on the real asset returns that - Canada, Denmark, Germany, Japan, Netherlands, can be attained, in relation to the growth of real wages. Sweden, Switzerland, the United Kingdom, and the Ideally, there should be a gap of 2 to 3 percent between United States - seeking to shed light on the finance of theni. Portfolio distributions and fund management are old age security in developing countries and the reform the key determinants of returns to pension funds, subject of pension funds in industrial countries. to the returns available in the market. Prudent The main peints of the analysis are as follows: diversification in domestic and foreign markets and * Pension funds are either defined benefit or defined indexation of much of pension funds' portfolios both contribution. The individual bears more risk with appear to be imnortant. defined contribution plans because the pension benefit * Pension funds affect capital markets in many ways. depends on asset returns. Conceptually, defined benefit They influence market structuee and demand for funds offer better "employee retirement insurance." securities; stimulate innovation, allocative efficiency, an Private defined benefit pensions are generally available market development; and have a positive effect on only through companies and typically include sctne overall saving. They may also have some deleterious restriction of labor mobility. effects, such as irncreases in volatility, "short termism," * Because of some shortcomings of fully or I irgely and weakening of the control exerted by investors and funded plans, especially for income redistributiotn, creditors over firms. governments have chosen to maintain at least basic le.'els * Prospects for pension funds in industrial countries of pay-as-you-go social security. rhe scope of such vary with the maturity of existing funds and the unfunded social security schemes is the key determinant generosity of social security benefits. In countries such a of the scale of private retirement saving. France, Germany, and Italy, growth in coming decades 7 The extent to which pension funds are used as a could be sizable. vehiclz for retirement saving depends on the regulatory * The key recommendations for countries that are just regime. Tax advantages are the most important starting pension funds are for a mix of social security an incentive, but a wide range of other regulatory choices private funds; for separate funding rather than "book also make pension funds more or less attractive to firms reserves;" for defined benefit plans, subject to and employees. And some regulations, such as those appropriate regulation; and for company-based pension affecting the portability of pensions, may have important funds. The paper - a product of the Financial Sector Development Department -was prepared as background material for th forthcoming Policy Research Report on Income Security for Old Age. Copies of this paper are available free from the Worl Bank, 1818 H Street NW, Washington, DC 20433. Please contact Priscilla Infante, room G8-118, extension .37642 (5 pages). December 1993. The Policy Research Working Paper Series disseminates the findings of work in progress to encourage the exchange of ideas about development issues. An objective of tl7e series is to get the findings out quickly' even if the presentations are less than fully polished. 7he papers carry the names of the authors and should be used and cited accordingly. The findings, interpretations, and conclusions are the authors' own and should not be attributed to the World Bank, its Executive Board of Directors, or any of its member countries. Produced by the Policy Research Dissemination Center THE STRUCTURE, REGULATION AND PERFORMANCE OF PENSION FUNDS IN NINE INDUSTRIAL COUNTRIES E P Davis This paper was prepared as background material for the forthcoming Policy Research Report on "Income Security for Old Age". The views expressed are those of the author, who is with the Economics Division of the Bank of England, and not necessarily those of his employer. The author thanks J-P Beguelin, Z Bodie, D Blake, N Collier, S Hepp, E Kroeger-Lohrey, F Lauritzen, W Naef, D Vittas and participants in a seminar at the World Bank for comments and suggestions, K Faulkner and S Friend for assistance with typing and K Woodfine for research assistance. Contents Introduction 1 I General Issues and Deflnitions 2 2 Structure 5 (a) The Pillars of Retirement Provision 5 (b) The Size of Funded Sectors 5 (c) Causes of Differences 6 (d) Structures of Pension Provision In Nine Countries 7 (e) Other Determinants of the Importance of Funding 11 3 Regulation 12 (a) Taxation 12 (b) Integration with Social Security 14 (c) Regulation of Portfolio Distributions 14 (d) Funding Rules 15 (e) Ownership of Surpluses 17 (f) Portability 18 (g) Internal Tran-sfers 19 (h) Insurance 20 (i) Fraud 21 0) Disclosure to Members 22 (k) Structure and Mechanics of Supervision 22 O) Is there a Consensus on Regulatory Practice? 23 4 Performance 25 (a) Benefits and Contributions 25 (b) Portfolio Distributions 26 (c) Returns on the Portfolio 31 (d) Fund Management 33 5 Effects on Capital Markets 36 (a) Jnnovation 36 (b) Market Structure 37 (c) Demand for Capital Market Instruments 37 (d) Volatility and Short Termism 38 (e) Development of Securities Markets 39 6 Conclusions 41 (a) Summary 41 (b) Prospects for Advanced Countries 42 (c) Issues and Recommendations 43 References 46 Notes 51 Appendix I Appendix II I INTRODUCTION This paper provides an overview of the principal Among the key issues are tax treatment, rules on policy issues relating to pension funds, illustrated funding, portfolio regulations, benefit insurance, by data and details of current practice for a protection against insolvency and fraud, vesting, selection of advanced countries, namely the US, ownership of surpluses and the mechanics of the UK, Germany, Japan, Canada, tU,e supervision, Netherlands, Denmark, Sweden, and Switzerland. Given its largely pay-as-you-go system, France Is The fourth section assesses aspects of the given limited coverage. The definition of pension performance of funds. To the extent that the funds employed is of flnancial intermediaries, available data permit, the level and nature of usually sponsored by non-financial companies, benefits paid, contributions and administrative which collect and Invest funds on a pooled basis costs of pension funds to the company are for eventual repayment to members in pensions. compared. The kev influence on returns and hence costs of providing benefits, namely The work is structured as follows; in the flrst portfolio distributions, are compared and related section we provide key deflnitions and an outline to asset returns, capital market structure, the of the general economic issues relating to pension nature of liabilities and regulation. Estimates of funds, many of which are developed and pension fund returns are presented' and the illustrated in the rest of the paper. In the second nature of the fund management process section we give an overview of the different considered. Some qualitative effects on capital structures for old age security, and the role of markets are also discussed, notably effects on pension funds in the countries studied, which is innovation, market structure, supply of funds and both of direct re:evance to the issues addressed the development of securities markets. A final and key background for the rest of the analysis. section offers a summary, takes a view of The third section addresses the main issues in prospects in advanced cA ies, and assesses pension fund regulation (including fiscal issues and makes recommenua.tions for countries treatment), and seeks to assess whether there is seeking to set up pension funds de npo.y any consensus on "good regulatory practice". 2 l. GENERAL ISSUES AND DEFINIONS Pension funds, which car be defined as financial The main fatures of pension funds can be intermediaries, usually sponsored by non-financial analyzed partly by contrasting them with other companies, which collect and invest funds on a types of provision for old age and financial pooled basis for eventual payment to members in institution. Hence unlike pay-as-you-go pension the form of pensions, are among the most funds, where workers' contributions are paid important institutions in certain national financial direct to pensioners, large quantities of funds are markets. For example, in 1988 in the US, accumulated by or on behalf of workers to pay pension funds held 17% of equitihs, in the their own pensions, and there is no Netherlands funds accounted for 40% of personal intergenerational transfer (The relation between sector assets, and in Switzerland their assets were pension funds and social security is discussed in equivalent to 68% of GDP. In contrast, in other Section 2.) Unlike banks, pension funds benefit advanced industrial countries such as France, from regular inflows of funds on a contractual Germany, and Italy, funds are of minor basis and from long term liabilities (i.e., with no importance (reasons for this are assessed below). premature withdrawal of funds', which together Reflecting these patterns, most economic analysis imply little liquidity risk. The main risks are has been performed in countries such as the US, rather those of inaccurate estimates of mortality the Netherlands, and the UK (see for example and lower than expected returns on assets. Bodie (1990a), Turner and Daily (1990), Van Loo Defined benefit pension schemes may also suffer (1988), Blake (1992) and Davis (1988), and their from the influence on liabilities of unexpected bibliographical references). changes in salaries, transfer payments out of the scheme and legal changes (e.g., equal retirement Pension funds are of two main types, namely ages). defined benefit and defined contribution, which differ in the distribution of risk between the Given the nature ot liabilities, pension funds may member and the sponsor (typically a non-financial concentrate portfolios on long term assets yielding company). In the former, the sponsor undertakes the highest returns, co npensating for the to pay members a pension related to career increased risk by pooling across assets whose earnings, such as a predetermined percentage of returns are imperfectly correlated. Pooling is final or average salary, subject to years of facilitated by the size of pension funds which service. Hence members trade wages for lowers management, information and transactions pernsions at the long-term rate of return in the costs and facilitates investment in large indivisible capital marke. jvhile employers undertake to top assets, such as commercial property. Portfolio up the fund to keep it in actuarial balance. This distributions and resulting risks and returns are risk sharing feature is absent from defined discussed in Sections 4 (b) and (c). Pension funds contribution schemes, where contributions are may in turn aid development of capital markets fixed and benefits vary with market returns; all (Section 5) although this may be hindered by the risk is borne by the employee. (In the case of portfolio regulations (Section 3(c)) or the structure a stock market crash just prior to retirement, such and behavior of the fund management sector risks for defined contribution plans may be severe (Section 4(d)). - pensioners in the UK who retired in 1974 often had pensions less than half the value of those Meanwhile, unlike other types of institutional retiring in 1973.) In addition, with defined investors (life insurance, mutual funds) pension benefit schemes there may be a transfer of risk funds in most countries benefit from tax deferral. between young workers who can bear investment Contributions are tax free, as are accumulated risk, and older workers and pensioners. This interest and capital gains; tax is only paid on enables such funds to have a high share of receipt of a pension after retirement. (Tax equity - trading return for risk. Note that both treatment is discussed in Section 3(a).) Hence, types may also have life insurance aspects e.g., for both the sponsoring company and the widows' benefits. employee - or for the individual, in the case of 3 poksonal pensions - pension funds are superior to schemes Is because they provide superior alternatives (for the company, unfunded schlemes, insurance to deflned contributione - although the for the employee, other forms of sav,ng). In implication is also that company-based defined addition, pension funds are generally contractual contribution schemes are superior to individual annuities, meaning that lump sum withdrawals are contracts. The balance in the countries studied precluded even during the period when claims are between defined contribution and defined benefit, payable after retizement. In contrast, for life and its daterminants, are discussed in Sections 2 insurance, early withdrawal is possible (at some and 4(a). cost) and policy loans also entail a degree of liquidity for holders. Members of pension funds Note that the Informdtion and Insurance are willing to accept low liquidity given potential arguments for employer provision suggest why for higher returns (at greater risk) that contractual the market (insurance companies, options annuities permit, supported by benefits of tax markets, etc) does not (and perhaps cano) deferral and implicit insurance of pension levels provide defined benefit schemes.' But the (in defined benefit schemes). Pension funds tend approach also highlights the fact that, particularly also to have much more liberal portfolio in the absence of separate funding, and legal regulation than life insurers, partly dv' to lower separation, pension benefits may be vulnerable to risks to solvency resulting from contractual risk of default of the firm in questici. In annuities, which again enables them to offer high contrast, social security pensions are sublect to returns. political risk that future governments will not honor benefit promises. As regards economic analysis, pension funds generally clearly have a role to play in the Ijf Insurance is not the only way to view pension jy&k pattern of saving, with the function of funds; there is also the tax shelter perspective ensuring that sufficient assets are available to (which suggests tax advantages to companies are provide income after retirement. Biu more the main reason for growth o funds). From a specifically, at a micro level Bodie (1990a) has labor economics perspective, defined benefit suggested that pension funds (particularly defined funds assist the employer by reducing costs of benefit) should be seen as a form of empioy labor turnover (if vesting is imperfect, i.e., early retirement insurance.3 Given risk sharing, leavers do not gain a proportionate share of insurance is provided against an inadequate benefits in relation to contrbutions) and hence replacement rate, social security cuts, longevity, funds can be a source of labor market investment risk and (to some extent) inflation. inflexibility. Even with perfect vesting and Pension funds are seen as insurance subsidiaries indexing, workers tend to lose out by changing of the sponsoring firm. He suggests this defined-benefit funds compared with those approach expiains a number of features of pension remaining in one fund, because part of their funds, notably provision by the employer and the pensions are based on the low salaries that they dominance of defined benefit schemes. Employer earnt early in their careers. Defined contribution provision may occur because they have superior schemes avoid these problems. (Portability is information over earnings, which are of key discussed in Section 3(f).) relevance to the employee's long term financial needs; benefit from economies of scale in The corporate rinanc perspective sees defined processing information, employing competent benefit persion fund liabilities as corporate debt fund managers, etc compared with individuals and fund investments as corporate assets which arranging their own pensions; can implement collateralise the pension obligation. (Section 3(d) enforced saving by deferring wages and salaries; offers a discussion of funding rules.) Given tax ca- overcome many of the agency problems faced deductability, corporations manage pension by individuals in dealing direct with financial funding and investment to maximize benefit to institutions;' and can avoid some of the adverse sharehulders. This perspective also rises the selection problems of private annuity insurance.5 issue of the status of members as stakeholders in Meanwhile, the dominance of defined benefit the firm, given ownership of the surpluses - as 4 well as liability for deflcits - rests with the owners of policy regarding surpluses, see Section 3(e).) of the company. Although the independent status of a fund cOers some protection from predators in lhe paper now go% on to assess the status of a takeover, stripping of surpluses and reduction of funded pension schemes in the structure of expected beneflts has been a controversial issue retirement Income provision in the nine countries (Schleifer and Sur-mers (1988)). (For discussion studied. S 2. STRUCTURE (a) The Plllars of Retirement Provision rates of return and increase in costs per capita.) Funding may adversely influence the exchange Pension funds are convintionally seen as merely rate and the current account if ex ante domestic one part of a system of piovision for old age. investmnent is less than the ncrease in saving. The other so-called "pillars" include compulsory The increase in saving may depress the domestic flat-rate social-security pensionsg (which is rate of return. A trust fund run by the u.ually pay-as-you-go, i.e, workers pay government could be diverted into public pensioners directly); & rnings-related social consumption, or at least be vulnerable to political security, often for those without private pensions influence (private funds avoid this problem). A (again pay-as-you-go); individual saving, transition from pay-as-you-go to funding can be including that via life insurance savings plans and difficult, as one generation has to "pay twice", purchase of residential property; support by the once for existing pensioners via pay-as-you-go, family; and work after retirement. and once for their own pensions via funding. Key macroeconomic and welfare issues arise in Also the problem of competition over domestic this context, particularly from the choice o resources is not entirely removed by funding; (public or private) pension systems between instead it is switched from pensioners seeking a fimding and pay-as-you-go. The issue arises share of labor income (v.a taxation) to claims partly from agi g of the population, which as over the returns on the capital stock. (Vittas discussed below will put increasing strain on pay- (1992) shows their equivalence in a closed as-you-go systemns, as workers and employers economy.) But at least ownership of the capital become increasingly burdened by social security stock may be a more secure basis for retirm.ent contributions, (i.e., there will be an increasing than the willingness of existing workers to pay problem ef competition over domestic resources). pensions as in pay-as-you-go schemes. In Such difficulties will impact on competitiveness, addition, the potential for conflict over use of depending on the situation in other countries. If domestic resources in the case of funded schemes contributions are seen as taxes they will distort can be reduced by international investment in the labor supply decision; this does not occur (developing) countries that do not face with funding. Pay-as-you-go may also discourage demographic problems. Such diversification saving and hence capital formation, while funding would alh. reduce any adverse effects on the increases it, thus raising future output for workers domestic rate of return and the exchange rate. and pensioners. However, given the conflicting risks arising from There are nevertheless some arguments against funde(& and pay-as-you-go schemes, analysts such funding. From a welfare point of view (Pestieau as Vittas (1992) suggest countries are best advised (1991)), funding may be objectionable for to have a mixture of both. intergenerational equity9 (because no transfers are possible between generations,'0 to compensate (b) The Size of Funded Sectors for a changing economic environment) as well as within generations (well paid workers who stay The data in Table 1 show a contrast between the with one firmn benefit most from the fiscal benefits role of pension funds in the Anglo-American offered). Pay-as-you-go schemnes can offer countries (the United Kingdom, the United States, Jmmediate pensions, without waiting for assets to and Canada), the" Netherlands and Switzerland," build up. They remove inflation risk to where they account for a sizable part of personal pensioners by linking future benefits to wages. sector saving and wealth, and those in other They can provide a higher rate of return to each continental European countries such as Germany generation if the sum of wage and *'mployment and France. Japan occupies an intermediate growth exceed the interest rate. (But if they do position, with sizable total assets but srnL'l in not, then there mnay be a corresponding fall in relation to personal wealth, saving or GNP. Table 1 Pension fund assets("), 1988 Stoct? of % of I of Total net I o I of assets (end- personal GNP investment personal ON" 1988) $bn sector Obn sector assets saving UK 475.9 27.2% 57.0% 21.9(c) 71.30) 3.5% Us 1646.7 13.20 33.8% 72 6(c) 49.9% 1.5% Canada 130.9 14.1% 26.7% 11.4Wd! 38.8% 2.4% Japan(b) 134.1 2.1% 4.6% 17 o(d) 19.5% 0.6% oermany(b) 41.1 :4 3-5% 44 (d) 3.9% 0.3% Netherlar1esf) 177.4 39.6% 77.9% :1 6(dl 37.9% 5.1% Swedon(q) 51.2 - 28.4% 3 d'd' ()) 2.1% Denmark(h) 13.1 12.9% 1 2(d) - 1.1% Iwitserland 121. 1 - 68.0% ll j(d) 95.0% 6.2% HemoJ France 27.7 3.1% 3.0% 1 o(C) 1.5% 0.1% Sources National Flow-of-Funds Data Notes (a) The table covers only independent funded schemes, which are the main subject of the paper, and hence excludes pension funds managed by life insurers, which in 1988 had asoets of $100 billion in the United Kingdom, $628 billion in the United States, $80 billion in Japan and $6 billion in Germany. (b) The data exclude unfunded Japunese and German pensio4 reserves held directly on the balance sheet of the sponsoring firm (booking). In 1988 these amounted to $87 billion in Japan and $100 billion in Germany. (a) Flow. (d) Difference of stock (ie may include some revaluations) (e) The large t lancing item in the United Kingdom national accounts means this ratio may be inaccurately measured. (f) Includes both public (ASP) and private funds. Private funds alone were $104 billion. (g) Data for Sweden relate to the ATP scheme, which is a hybrid between social security and funded private schemes (it is nationally coordinated but relies on employers contributlons and employers are represented on the investment boards). There are also private schemes in Sweden (ITP/STP) but they are usually booked or unfunded. (h) 1987. (i) Not meaningful (saving negative). 5a 6 Similar contrasts are apparent over time. The othor types of financial asset. (Although proportion of peisonal sector financial wealth'2 legislation in the United Kingdom outlawing accounted for by poasion fijid assets, and the compulsory membership, as well as the ratio to GDP (see Table 2', has increased in all development of personal pensions in a number of the countries illustrated, although by different countries, may make the situation more fluid.) amounts. Absolute growth hr- also beeni rapid. On the other hand, the nature of the benefits Real growth in Japan over 1980-8 was at an offered may provide an incentEve to work for a avorago of 17% (UK 13.3%, US 8.8%, Cianada particular firm, making it attractive for that firm 6.4%, Netherlands 7.5%). to of'er a particular type of scheme. For employees, pensions have often been a subject for Savings basad life insurance policies, and pension collective bargaining (particularly in the United funds managed by life insurers, are of eourse States, Denmark, and the Netherlands). And as alternative ways to pension funds of fir ncing noted, private annuity markets suffer from retdiument. The combined size of life insurance imperfections, encouraging employees collectively and pension fund sectors has also grown, albeit to press for vension funds to be set up. The more often nmore slowly than pension funds alone generous the beneflit offered, and the wider the (Table 3). The principal change In the ordering Is ;overaU, the more assets pension funds will in Japan, where the size of the life insurance require. Finally, taxation and regulatory sector is almost eight times that of pension funds proyisions, as discussed in Section 3, make it (run by trust banks). In most other countries the more or less attractive for the firm to offer a size of the jife sector is commensurate with the pension fund. For employees, too, high marginal size of its pension funds. tax ra.es may increase the attraction of tax deferral via pension funds. This section now goes on to outline the causes of these differences by refeisnce to structural But the most crucial point is that private funded features of the funds themselves and the main schemes cannot usefully be viewed in isolation; alternative pillar, namely social security. Note the principal altematiy to a private pension fund that a complete assessment of the causes of the is the state social security pension scheme. Not differences in size of funded sectors must also surprisingly, the growth of private schemes can incorporate the arguments presented in sections 3 be related to the scale of social security piension and 4, which respectively address regulation and provision, which impose limits on private sector performance; these underly the structural schemes. Note that social security is invariably a differences between funded schemes that are compulsory, indexed, defined benefit, and usually outlined here, as well as the scale of their use as unfunded pension scheme. compared with other types of private saving (the 'third pillar"). On the other hand, as noted, the age structure of the population will determine likely future strains (c) Causes of Differences cii a social security system. As shown in Table 4, rapid aging of the population, with a rising What accounts for the differences in the proportion of, retirees, is projected for all importance of funded ors in the provision of advanced countries, but especially those in pensions? Since pension funds comprise financial Continental Europe and Japan (see also Hagemann assets, it is natural to begin with portfolio and Nicoletti (1989)). This results largely from considerations such as risk and return. However, declining birth rates, but also greater longevity the majority of pension fund members are and a decline in the amplitude of migration. affiliated as a consequence of their employment, Where social security is relatively generous (see and such fund membership is often compulsory, Table 5), maintenance of promises may lead to although setting up of a fund is not compulsory vastly increased contributions from the workforce, for the firm, except in Switzerland and France. resulting in a loss of competitiveness due to Therefore rates of return on pension funds do not higher wages and/or a marked reduction in attract investors in the same direct way as do personal income. For example, Mitra (1991) Table 21 Pension fund assets (as a p.raentage of ODP) 1970 1971 1980 1985 1990 Ut. 17 15 23 47 55 Us 17 20 24 29 35 Germany 2 2 2 3 3 Japan 0 1 2 4 5 Canada 1 13 17 23 28 Netherlands 29 36 46 68 77 Sweden 22 2. 30 29 28 Iwitierland 38 01 51 59 69 Denmark 5 5 7 12 15 lourc:t Nat!e,nal llow-of-Fun4s data. Table 3i Life insurance and pension fund assets (as a percentage of GDP) 1970 1975 1980 1985 1990 UK 43 37 46 83 97 uS 37 37 42 49 59 Ge' ny 10 11 14 19 22 Japan 8 10 13 20 41 Canada 31 28 31 39 46(3) Netherlands 45 51 63 86 107 Sweden 42 48 51 55 63 Switzerland 51 1is 70 82 n/a Denmark 14 14 19 31 n/a Memo Items5 France 6 7 7 9 13(1) Italy n/a 4 3 6 12(2) (1) 1988 (2) 1987 (3) 1989 6a Table 4: Percentage of Population Over 65 1990 2020 Percent 2050 Percent L%ange Change As a percentage of population 15-65 UK 23.1 25.6 10.8% 30.4 18.8% Germany 22.5 33.2 47.6% 42.3 27.4% Netherlands 18.5 28.9 56.2% 38.1 31.8% Sweden 27.4 33.0 20.4% 35.8 8.5% Denmark 22.7 30.5 35.6% 39.8 30.5% Swltzerland 25.0 48.1 92.4% 46.0 -5.6% France 21.0 30.5 45.2% 37.8 23.9% US 18.7 25.0 33.7% 31.8 27.2% Japan 16.6 33.7 103.0% 37.6 11.6% Canada 16.8 29.0 72.6% 36.4 25.5% Source: Hagemann and Nicoletti (1989) Table 5: Indicators of the scope of social s*curlty pensions Payments/GDP (1985)(1) Social securLty replacee nt rate for single worker (1980)(2) US 8.2% 44% UK 7.5% 31% Germany 13.5% 49% Japan 5.2% 54% Canada 6.4% 34% Netherlands 11.8% 44% Sweden 13.0% 68% Denmark 9.8% 29% Switzerland 8.5% 37% Memo: France 14.4% 66% (1) Source: Mitra (1991) (2) Source: Aldrich (1982) 6b 7 suggests that contribution rates for social security This section discusses the balance between social pens.ons in Germany might rise from a current security and funded pensions, together with key 14% of labor costs to 23% in 2010 and 30% in structural features of the systems which determine 2050. French government calculations suggest a the balance between them. These features are rise from a current 19% to 31-42% in 2040. In' summarized in Table 6. Japan the contribution rate would be 30% in 2020 under unchanged policies. In theJlK (Blake (1992)), 70% of workers have a funded pension, of whom 50% are in company Political problems would be likely to follow such schemes. Schemes are quite long-etablished; the increases, as well as the workforce and industry current level of coverage of company schemes "voting with its feet' to shift to other countries was reached in 1967. Defined benefit plans, with lower contributions. Elements of this are often with provisions for a degree of indexation, already apparent in Germany, where firms are cover all public sector and the majority of private tending to locate new factories in countries with sector beneficiaries. Defined contribution plans lower social costs. Note that taking the strain via declined in popularity during the mid-1970s, an increased public deficits instead of taxation will era of high inflation and low real rates of return only postpone the problem until the bonds need to to investment. Defined beneflt plans are be repaid with taxpayers' money. A rational obviously vulnerable to deficits during periods rf private sector in the sense of Barro (1974), which securities market weakness, such as the 1970s, anticipates perfectly the future taxes to pay off and firms had to make large "topping-up" bonds and immediately adjusts its expenditure payments in the late 1970s. More recently (since accordingly, would not even differentiate the two 1981), asset growth has reflected the strength of cases. capital markets, and with in addition widespread reductions in membership due to redundancy Governments are seeking to limit social security (which reduces projected pension obligations), commitments, and stimulate private saving for many schemes became overfunded, with firms retirement, in the light of these potential burdens. taking contribution holidays. And the advent of They may also seek partially to fund social personal pensions has accompanied a resurgence security. (Related policies are to increase the of defined contribution plans. The development labor force participation rate, notably of older or of social security has been favorable to private even "retired' people, to encourage immigration, schemes; employees with company pensions may to raise the retirement age,"3 and to seek to "contract out" of all but the most basic state promote fertility.) In the light of these polic-es, scheme, and the government, concerned over and associated expectations of further action, in future state pension obligations, is offering many countries, individuals now anticipate incentives to individuals without a company promises will be scaled down in the light of the scheme to take a personal defined contribution burden of such schemes on future wage earners pension instead of an earnings-related state md/or government borrowing. This in turn is pension. It is also reducing the maximum ,timulating precautionary saving via institutions benefits from the latter. indeed, private pensions can be seen as a form of private sector insurance against the political risks In the United States (Turner and Beller (1989)), of a government run system). Section 3 probes coverage is lower, at 40%. Most primary private more deeply the further question of why increased funded pension coverage is again in defined precautionary saving should occur via pension benefit schemes (which account for two thirds of funds, rather than private voluntary saving of pension assets). However, a large number of other types. workers also have supplementary defined contribution plans. From 1975 to 1985 US Structures of Pension Provision In Nine workers covered by defined benefit plans rose ountries from 27.2 to 29.0 million but fell from 39% to 30% of the workforce, while participants in 'hat are the structures of the pension systems? defined contribution plans rose from 11.2 to Table Gas Features of funded penuion systess UK US Germany Nature of benefiu for Largely defined benefit Primary cover largely Largely defined benefit average member based on final alary. defined benefit based with flat rmteLbenefit on fnl salary. bued on yea of Supplementary defined service. contribution plans widespread. Taution of mnded Contnbutions and asset Contributions and uset Employers' schem. returns tax free returns tax fre. contributions taxed as Benefri tuxed, exept Benefit taxed. wages; employee tax free lump sum. contributions and aset returns tax free. Benefits taxed at low rate. Social securiy Low replawomnt ratio. I w replacement tio. High replacement ratio. Scheme members can ontract out of earnings related sochl scurity. Regulation of porolios Prudent man oDncept; Prudent man concept; Guidelines; maximum S* self investrnet 10% Umit on self 20% equity, 5% Umit; concentration investment for defined property, 4% foreign; limit for defined benefit plans. 10% self investment contribution plam. limit. Regulation of funding Maximum 5% overfund Maximum 50% Funding obligatory for (seo Section 3(d) of IBO or PBO. overfund of ABO. pension funds Funding only obligatory Higher insurance (Pensionskassen), albeit for contracted out part premia if underfunded. only up to PBO. of social security. Option of booking (tax exempt- pensions taxed at normal rate). Maturity of funds Matur. Mature. Immature. Coverage of workforce 50% (company 4S% 42% (approx) schemes) 20% (p'rsonal pensions) Insurauce of bonefits No (although stato Yr (special guarantee Yes (via insuranco guarantee payment of corportion). supervisors). Booked minimum pension if benefits insured by fund default). Pension Guarantee Association. PortAbility features Vesting in 2 years. Vesting in 5 ye". No Vesting in 10 years. Indexation of acrued indexation of accrued Indexation of accrued benefits. Trnsfer benefits. Lump sum benefits. must be made to other disribution permitted pension fuds. on truufer. Indexation Discretionary (to date) Full inuexation rare Mandatory. but total or parial (5% of schemes). indexation common in Discretion cost-of- pratice (75%). IivLg increases oommon. 7a Table *bt Features of funded pension systems Jaran wanada Netherlands Nature of benefits for Largely defined benefit Largely defined benefit Almost exclusively aveoge member based on years of baed either on final defined benefit baned service and career salary or flat rate on fil ulary. earning or final basic benefit. salary. Often taken as a lump sum. Taxation of funded Contributions tax free. Contributions and aset Contributions and aset schemes. Tax on asset retums. returns tax free. returns tax free. Benefit taxed, exoept Benefits taxed. Benefits taxed. tax free lump :.m. Social seurity High replacement ratio. Low replacement ratio. Lo- -ent rdo. Scheme members can contract out of earnings related social security. Regulation of portfolios Guidelines; maxirnum Prudent man (since Prudc it .- -; 5% se(f 30% equity, 20% 1987); tax on foreign investm ... lizri except property, 30% foreign, assets above 10%; 7% for ABP (sc. su&). 10'% one company. limit on real etate. Minimum 50% bonds. Regulation of funding Funding optional. Tax Funding obligatory. Funding obligatory for. (see Section 3(d) exempt up to ABO Maximum 5% overfund IBO or PBO. only. (Book reserves of PBO. tax exempt up to 40% of liabilities). Maturity of funds Immature. Mature. Mature. Coverage of workforce 37% (funded plans 41% 83% (approx) only) Insurance of benefits Yes (under wage No (but social security Contnbutions insured payment law). Mutual provides backup). for one year. guarantee scheme for EPFs introduced 1988. Portability features Vesting graded between Vesting after 2 years. Vesting in one year. 5 and 30 years for Little indexation of Accrued benefits voluntary leavers. Low accrued benefits. indexed. transfer values for Tranferability within voluntary early leavers. extensive penion circuits with same conditions. Indexation Rare except for part Provisions rae (6% of Indexation almost replacing social private schemes); some universal) albeit not security. discretionary increaes. mandatory). 7b Table Get Features of funded pension systeas Denmdrk Sweden (ATP) Switzerland Nature of benefits for Largely defined Defined benefit based Majority of schemes averago member contribution. on best income years. (60%) defined contribution but with targets of 60% replacement rato which contributions adjusted). 40% defined benefit. Taxation of funded Contributions tax Contributions tax free. Contributions and asset schemes. deductable. Fund may Tax on asset returns returns tax free, be taxed. Benefits (1991) benefits taxod at benetts taxod. taxed, including 40% low rate. of lump sum. Social security High replacement ratio. Low replacement ratio; Low replacement rtio; only for basic needs. designed tn be supplemented by mandatory private scheme. Regulation of portfolios Real estate, investment Majority to be in listed 30% limit on domestic trusts, shares linited to bonds, debentures and shares. 50% domestiu 40%. 60% in domestic retroverse loans to real estate. 20% debt. No self contributors. foreign currency assets investment. Only 10% foreign shares. srr.aU proportion' can be invested internationally. Regulation of funding Irrelevant as defined Contribution rate Funding compulsary for (see Section 3 (d)) contribution. adjusted S-yearly to PBO or IBO. ensure. IBO is funded. Maturity of funds Mature. Mature (pre-BVG immature (post-BVG). Coverage of workforce 30% (company funds) 90% (compulsory) 90% (compulsory to (approx) 20% (persona! workers and pensions) employers). Insurance of benefits State backup as national Yes; Government schf ie. Safety Fund. Small funds backed by insurance companies. Portability features !mmediate access to Vesting immediate - Immediate access to own contributions, 5 national scheme and minimum contributions; years total vesting. transferability perfect. imperfect vesting for Tr-..1sfer values can be employers' excas negotiated. contributions, with graded vesting between 5-30 years of service. Indexation No. Yes. Indexing not compubory but almost universal in prumice. 7c 33.2 million (14% to 33%). Advantages of qualified pension plans (TQPPs), authorized defined contribution plans for the employer 1962, are similar to Anglo-American funded include lower regulatory and administrative costs pension plans, and are available to firms with, (including avoidance of PBGC insurance premia), or more employees. In 1989 they covered 28 as they need not meet the actuarial funding of the private sector workforce and held assets of standards required of defined benefit funds; shift $76 billion. 90% of benefits are taken as a '--- of risk to employees, as noted in Section 1, sum. Employee pension funds (EPFs) (19 although this should be offset by higher unlike TQPPs, enable the private plan to replace compensation; and self investment being the earnings related component of social securi permitted for over 10% of assets. (and hence the firm can contract out of earning related social security contributions), and are o But in fact Kruse (1991), using US micro data, available to large Ftrms with 500 or more suggests that rather than a positive shift by employees. Benefits are in the form of an employers, with termnination of a defined benefit annuity equal to the social security pension plus plan, the relative shift io defined contribution the excess (which has to be at least a further relates largely to slower employment growth for 30%) - often taken as a lump sum. These cover firms offering defined benefit plans (although 26% of the workforce and had assets of there was some supplementing of defined benefit $143 billion in 1989. Both schemes' "defined by defined contribution). Effects of relative costs benefits" usually relate to final "basic" salary, on shifts towards defined contribution were also which may not keep pace with total remuneration, not large. Finally, greater economic instability in given the importance of bonuses and allowances. an industry leads firms introducing new pension These plans coexist with traditional unfunded plans to choose defined contribution, perhaps due retirement bonuses, which benefit from a 40% tax to lower risk. deduction for accruing liabilities, payable when they are earmarked through an accounting entry Social security is the US is again supportive of in the books of the firm. private schemes; the replacement rate is low (though finds can take full account of social In contrast to the US and UK, social welfare security in paying pen-ions, so as to ensure a promises in Japan were historically relativel fixed replacement ratio for all levels of income). generous, with a prospective "replacement ratio A recent reform will make social security (average pension as a proportion of averag pensions a smaller proportion of earnings, earnings) of over 50 per cent (Table 5), although beginning in the year 2000, and will increase the a reform of 1985 will gradually reduce publi age at which full benefits are payable. It also pensions as a proportion of average earnings. A introduced a degree of prefunding for social in the US, some assets, amounting to 50% of security; funds are accumulated in a trust fund GDP at present, are accumulated by the state in and invested in government bonds. This should advance of benefit commitments; this can help in principle reduce any tendency for social allay demographic concerns. Such social security security provisions to reduce national saving benefit commitments are likely to constrain the (while increasing the risk that it will be diverted growth of pension funds. However, social by the government to unproductive uses). security in Japan is not payable until 60, while However, as pointed out by Bodie and Merton retirement is often at 55, so a private pension can (1992), it is not clear that government's bridge this gap."' In addition, as noted willingness to repay bonds (or at least, not to companies can opt out of part of social security devalue them by a bout of inflation) should be contributions by paying an equivalent pension. any more relia*le- than the promise to pay pensions, unless the funds are used for productive The German private pension system comprises capital investment, with revenues hypothecated to four main types of scheme (Deutsche Bundesbank pay pensions. (1984)). The largest are unfunded schemes, "direct commitments" (Direktzusagen) on the In Jaa (Murakani (1990), Clark (1991)), tax balance sheets of large firms, which are usually 9 mutually insured to cover the risk of as an initial response to demographic concerns. bankruptcy."S In 1990 these were 60% of (See Schmahl (1992a).) pension liabilities, valued at DM 181 billion. Another common form of company scheme is In the Netherlands, 'supplementary" pension 'direct insurance" (Direktversicherung) (10%), funds have developed over a long period, often as whereby an enterprise concludes a contract with a a result of collective bargaining, to cover virtually life insurer on behalf of its employees. the entire labor force (83%) - despite not being Employees then have a direct claim on the life compulsory for employers"7 - and were codified insurer. Risk and administrative expenses are in the Pension and Savings Fund Act of 1953 (see shifted to the life insurer, but the funds are of no LutJens (1990), Zweekhorst (1990)). 90% of direct use to the firm. An enterprise may also pension plans are defined benefit (usually paying commission a legally independent "pension fund' 70% of final salary, in combination with the basic (Pensionskasse) (1990; 20%; DM 61 billion) or social security pension), and 90% of pensioners "provident fund""' (Untersttitzungskasse) (1990; receive inflation protection. Private pena.-n 10%; DM 29 billion) to handle its pension provision in the Netherlands falls into three scheme, operating as a mutual insurance categories; industry funds covering multiple association. Pension funds are closest to practice employers (40% .f the workforce); and elsewhere. Provident funds face no limit on individual company funds (19%); insurance investment; all can be loaned back to the contracts (3%). There is also the pension fund sponsoring company, and there is no legal right to for public servants (ABP) (28%). Industry funds benefits. However, since 1974 only part of may be made compulsory by collective agreement transfers to provident funds h3ve been tax- for all employers and employee organizations. deductible for firms as an operating expense (all Corresponding to the development of private may be deducted for pension funds) and pensions, social security only offers a minimal employees' legal rights to benefits have been basic benefit related to the minimum wage. strengthened, so provident funds have declined. In Canad funds are again largely defined benefit. A recent development is "special security funds' Private 'trusteed" schemes, which cover 40% of (Kapitalanlagegesellschaften), a form of the labor force, co-exist with a flat rate non- investment company whereby highly-liquid firms contributory state pension scheme (OAS), a having direct commitments can invest part of their negative income tax (GIS) for those over 65 on pension provisions in the capital markets. This low incomes and a contributory earnings-related overcomes the concentration of risk inherent in public pension (CPP/QPP). The last is partly booking the liability on the firm's balance sheet. funded. Given the attraction of exemption from capital gains tax and turnover tax these have grown In Sweden, the_main funded pension scheme is a rapidly; inflows were DM 19 billion and a;sets compulsory, publicly directed "National DM 116 billion in 1990, although only a part of Supplementary Pension Scheme" (ATP Scheme), these were counterparts to pension liabilities. set up in 1960, which complements a basic, flat rate, social security scheme. It covers 90% of the The development of German private pensions workforce. The aim is to accumulate significant needs to be put in perspective, as it accounts for a quantities of funds to provide future benefits, thus relatively small proportion of personal saving and offering an occupational pension that is indexed wealth, even if unfunded schemes are included and equal to a sizeable proportion (60%) of the (Table 1). This is largely because Germany has a best years of earnings. The fund is administered relatively generous, mandatory and wholly pay-as- independent of the government in a series of sub you-go state social security scheme (Table 5). funds, which invest monies from different sectors Private schemes are supplementary, and need far of the economy (public sector, large firms, small fewer assets to cover their more limited firms/self employed) in a variety of both public commitments than elsewhere. However, the and private financial assets (Section 4). There are retirement age has recently (1989) been increased also certain smaller private schemes in Sweden, 10 one for white collar workers (the ITP system) and private pension schemes, which already cover one for blue collar (the STP system). The ITP 85% of the workforce. Afterl'institution of BV system is funded either through book reserves, this rose to 90% (it excludes the unemployec through insurance contracts or through contracts some part time and temporary employees, an, with a special pension company, while the STP those under 18). Unlike the public ATP scheme scheme is unfunded. However, we focus in the in Sweden, fund management is not centralized, paper on the ATP scheme, as the major funded but arranged by the individual employer. scheme invested directly in the capital markets, while bearing in mind - and using for comparative In France and Ialy, the generosity of the sta purposes - its public sector basis. scheme (supplemented in France by "hybrih private/public industry-wide pay-as-you-g4 The jDjsh funded pension schemes are private, schemes) has been such as to almost completel' largely defined contribution plans run by private crowd out funded private pension plans (s companies for their staff, although some Metais (1991)). For this reason, these countries multinationals offer defined benefit schemes. are not covered in detail in the current analysis. There are also natianw; .- sectoral and Recent proposals in France to increase the professional pension funds, which are classed as importance of private pension schemes (for mutual insurance companies. Retirement assets background, see Commissariat Generale du are accumulated in banks and life insurers as well Plan (1991)) face difficulties given the short-run as pension funds; the latter account for only 28% fiscal implications of tax-free pension of the total. The attraction of private pensions to contributions. However, in Italy a 1991 reform blue collar workers is reduced by the generosity did seek to raise the retirement pension age from of the public pension system, which currently 60 to 65, impose higher contributions and offers a replacement rate for married couples of longer contribution period (20 to 35 years). 66% (OECD (1988)). (Note that the data in Table 5 are for 1980 and single employees.) To summarize, the influence on the development High composite marginal income tax rates on of private schemes of the scale of social security. supplementary pensions are also a disincentive to offset in some cases by demographic concerns, pension saving. can be discerned in each country; for example the Swedish public and Swiss private national The SAiss pension system (Hepp (1990)) consists funded systems are designed to provide the bulk of the state social security scheme (AHV/IV), the of retirement benefits beyond a basic flat rate compulsory occupational pension schemes pension, and are accordingly both compulsory and (BVG/LPP) and individual saving. The formation comprehensive. In a more free-market context of the BVG/ILPP schemes stems, as in Sweden, the forces encouraging funding are also at work in from recognition that the state pay-as-you-go the Netherlands, the United Kingdom, the United scheme would impose a rising burden on future States and to a lesser extent Canada and generations, as well as desire to increase the Denmark; state pensions are not comprehensive, proportion of final salary provided in pensions and thus development of funded schemes is (i.e., to fill the gap between state pensions and encouraged. Meanwhile, in Germany and Japan the retirement income considered socially relatively generous social security promises, as desirable). However, the funds are more clearly well as tax incentives to "booking" have - at least private-sector than the Swedish system. The until recently - accompanied smaller funded BVG requires companies basically to set up a schemes, while in France and Italy they have deflneu contribution plan, which, together with crowded them out completely. An illustration of social security, offers a defined benefit target these relative patterns is the ratio of social (90% of retirement income for the low paid, 60% security pensions to national income, which in at average earnings, 25% for top earners). Many 1985 was 12.5% in Continental Europe, 7.5% in individual funds offer defined benefits, which may the UK, 8.2% in the US and 5.2% in Japan (see target a higher replacement ratio. When instituted Table 5). The relatively high UK and US levels, in 1985, the scheme was grafted onto existing which contrast with the low replacement ratios, 11 relate to the current age structure (Table 4) with a whether they have a long-run ratio of contributing relatively high proportion of pensioners. to benefiting members. Immaturity helps explain the growth of schemes in the Anglo-American (e) Other Determinants of the Importance of countries, the Netherlands, Sweden, and Funding Switzerland over the last twenty-five years. Now, some of these schemes are maturing, and the Personal pensions, which ar invariabl3 defined- growth of their assets will slow (to around the contribution, have grown in importance in recert growth rate of real wages), although changing years, the main aims being to provide the tax regulations, such as those for indexation and incentives of pension schemes to those not in retirement ages, as well as broadening of company schemes, to enable company schemes to coverage following moves to compulsion, may be supplemented, and/or to offer greater add to this. (Commentators suggest that recent portability than is available from company changes in UK regulations could boost liabilities schemes. In some countries, boosting national by £40-50 billion.) As discussed in Section 4(b), saving was also a motive, although evidence as to maturity may have an important effect on its success is mixed. (See Venti and Wise (1987) investment, as income from assets becomes and Gravelle (1991) for opposing views on the relatively more important than capital growth. US.) Individual retirement accounts (IRAs) were Maturity for an individual scheme will depend on introduced in the United States in 1974 for its history and development, and demographic workers without company pensions; they offer factoms. Thus, 'aging of the population' in many the same tax benefits as pension funds and grew countries is leading to growth in pension funds. more rapidly after 1982 when all workers and their spouses became eligible (15 million plans As an exarnple of maturity, outflows in the United were open in 1985). Similar provisions, States exceeded inflows by $lbn in 1989 and introduced in the early 1970s, cover 3 million $6bn in 1990 (growth of assets also depeids on workers in Canada (1987). More recently asset returns, of course). Also the number of 4.5 million have taken "personal pensions" in the beneficiaries rose 41 % between 1980 and 1986. United Kingdom, generally opting out of the United Kingdom net inflows were 19% of assets social security earnings related scheme. in 1980 and 4% in 1990. By contrast, schemes in Regulations state that UK personal pensions must Germany and Japan are less mature, so future be indexed up to 3%, and 25% of the value can growth will continue to be strong. For example, be extracted as a tax free lump sum. France and in Japan in 1988 on y 9% of the population over Switzerland have introduced similar provisions. 65 received a pension from a funded scheme. On balance, personal pensions seem to have complemented rather than substituted for other Coverage is obviously also important (i.e., the types of private provision. Andrews (1990) proportion of employees covered by pension suggests that countries such as Japan find personal plans, which as shown in Table 6, varies between pensions unnecessary, due to low labor mobility 90% in Sweden, Switzerland, and the Netheriands and a high savings rate. to around 40% in the US, Germany, Japan, and Canada). However, this is a consequence of A further factor influencing the size of pension factors discussed above and in Section 3, rather funds is the maturity of the schemes, i.e., than a separate cause of growth in itself. 12 3. REGULATION This section assesses the main issues in pension assets in long-term institutions in countries such fund regulation, comparing and contrasting the as the United Kingdom and the United States as a adopted solutions in the nine countries studied. It proportion of personal portfolios has a counterpart is suggested that whereas social security is the key in a continual reduction in direct personal equity determinant of total precautionary saving for holdings as a proportion of financial assets. This retirement, it is the fiscal and regulatory clearly partly results from the fact that direct environment that influences the use made of equity holdings generally suffer from double pension funds as a vehicle for such saving. An taxation (purchases of securities are made from attempt is made to come to a view regarding taxed income, and both dividends and capital "good regulatory practice". gains are also taxed)."' However, in the longer term, this reduction may also result from an (a) Taxation equalization of the income and wealth distribution, where only the wealthy could One of the main determinants of the scale of economically maintain equity portfolios with benefits and advantages of pension funds as a adequate risk diversification, although mutual means of saving is exemption of contributions funds overcome this problem. As a means of from taxatioq. As discussed in Johnson (1992), retirement provision, equity holdings also have pensions may be taxed at three points, when the disadvantage of greater capital and income money is contributed, when investment income is uncertainty than institutional iavestment and earned and when retirement benefits are paid to (particularly) defined benefit pension funds. scheme members. In general, taxing contributions only and benefits only are equivalent In other countries such as Denmark, (except to the extent that progressive taxation may supplementation of the state retirement schemes be lower on lower post-retirement income, and occurs via insurance schemes and bank investment deferment itself means pretax rather than post tax products as well as through pension funds (given income is available for investment and a more even tax treatment). accumulation). These are expenditure tax regimes, where the post tax rate of return equals Reasons for taxing pensionr¢ relatively leniently the pre tax rate, and the consumption/saving include, first, the need to assist people to save choice is not distorted; consumption is taxed at enough to maintain post retirement living the same rate now and in the future. standards; second, a desire to encourage people to save and thus cut the cost to the state of means- In contrast, regimes where investment income is tested social security benefits; and third, to raise taxed as well as contributions or benefits are the general level of saving. comprehensive income tax regimes (they tax income equally regardless of source). This The first is the most important, and is largely reduces the incentive to save by driving post-tax paternalistic. It suggests that people are generally rates of return below-the pre-tax rate. Note that myopic, and/or that there is a form of moral if the distinction between- nominal and real returns hazard, in that they assume they will be cared for is not made by the fiscal authorities (i;e., nominal by the state even if they do not save. That people returris are taxed) a comprehensive income tax do not save sufficiently is confirmed by US also induces a growing distortion dependent on studies such as Diamond (1977), and recent the rate of inflation. In general, pension funds evidence in New Zealand shows removal of tax are given expenditure tax treatment, while other exemption can cut retirement saving sharply. Of forms of saving are not. There is thus a course, compulsion (as in Sweden and distortion between types of saving, encouraging Switzerland') is an alternative way of ensuring accumulation' via pension funds. adequate saving, but tax exemption mitigates the associated element of coercion. However, despite Corresponding to this distortion, the growth of this argument, Munnell (1992) argues for taxation 13 of pension fund income in the US on equity In Germany, employer contributions to grounds, as with coverage of less than 50%, the independent pension and provident funds (and bulk of benefits to tax deferral, which amount to direct insurance) ire treated as current income of over $50 billion per year,20 go to richer employees and are subject to wage tax - hence people.21 This is particularly the case for deferred taxation is absent - although pensions are Individual Retirement Accounts (Munnell (1984)). taxed lightly compared with earned income, partly to compensate. This provision makes "direct The argument of encouraging saving and reducing commitments" (i.e., pension liabilities held on the social security is only applicable when state books of the sponsoring firm), which are fully schemes are means tested and/or opting out is tax-deductible, more attractive. They are possible, as in the UK and Japan. consequently the dominant form of private pension obligation, accounting for 60% of pension The evidence for the third effect, i.e., raising the liabilities, compared with 30% for funded level of saving, is positive but minor (see Section independent pension funds (Pensionskassen) and 4). Johnson (1991) concluded that these provident funds. arguments for special treatment of pension funds are less well founded than those for the general In Denmark, there is a special variable tax expenditure-tax treatment of saving, all of which (currently) 44% on pension asset returns, which is could contribute to retirement income (a counter imposed when real returns exceed 3.5 %. This argument is presumably that other forms of saving thus avoids the comprehensive income tax's may be decumulated at will, whereas pension difficulties with inflation (as outlined above), but funds are unique in being contractual annuities, as does impose some deviation of pre and post tax defined in Section 1). returns. Equities are exempt. The reason for the tax was concern that high real returnis could lead The United Kingdom is an example of pension payments to exceed earnings. expenditure tax treatment of pensions, where Meanwhile, taxes on receipts of supplementary employees' and employers' contributions and all pensions are reportedly so high as to constitute a returns on investments are free of tax and disincentive to pension saving. Sweden imposed employers' pension contributions, unlike wages, a major reform in 1991 (Munnell (1992)), to tax are not subject to national insurance contributions. all annual earnings on pension funds, to offset (A pay-as-you-go scheme, in contrast, would not losses in revenue due to tax deferral and improve gain tax privileges nor be eligible to contract out equity with other forms of saving. The rate is of earnings related social security.) However, an 15%, half the rate of tax on other forms of anomaly, which is contrary to expenditure tax saving. Taxation of benefits is relatively low; treatment (as wel' as the idea of pension funds as contributions are tax exempt. "contractua! annuities" (Section 1)) is that up to one and a half times an employee's salary (up to In the case of company pensions, the attraction of £150,000) may be taken out at retirement as a schemes to employers is important, since tax-free lump sum. Recently nominal limits have provision is only compulsory to firms in Sweden been imposed on tax-free contributions, and other and Switzerland. "Direct commitments" in forms of saving such as equities and deposits have Germany, in effect, offer tax-deductible "free been accorded (limited) expenditure tax treatment. capital" to the firm,' tinough in principle the The tax treatment of pension funds is broadly liabilities arising from pension claims should be similar in the United States, Canada, Japan, and reflected in the share price. In Japan a taxation the Netherlands. However, in Japan (Clark change in 1980 encouraged companies to replace (1991)) other forms of saving also enjoy tax unfunded by funded pensions or bonuses, by privileges, pension funds' asset returns are subject reducing from 50% to 40% the amount of tax to a special 1% corporate tax, and unfunded free book reserves that could be san against liabilities are partly tax deductible, (which could pension obligations. Many schemes remain help explain the slow growth of funded schemes unfunded however. In the Anglo-American in Japan). countries and the Netherlands the tax exemption 14 of funded schemes makes them the cheapest way also often limits on self investment, to prote for firms to provide retirement benefits to against the associated concentration of rih employees. Unfunded private pensions - which regarding insolvency of the sponsor. Apart from acccunt for virtually all private pensions in the self investment control, the degree to which France, and which are themselves compulsory - such regulations actually contribute to security may appear advantageous to companies when open to doubt, since pension funds, unlike population and the economy are growing, interest insurance companies, face the risk of increasing rates are low and employment is high, but in liabilities as well as the risk of holding assets, and more adverse circumstances may prove more hence need to trade volatility with return. risky to the firm, workers and pensioners. In Moreover, appropriate diversification of assets effect, they may face similar demographic and can eliminate any idiosyncratic risk from holdin, flnancial problems to state social security without an individual security (such as an equity), thu the ability to raise taxes. These problems also minimizing the increasc in risk - and if nationa arise for German or Japanese 'book reserves" if cycles and markets are imperfectly correlated actual investment does not follow the booking of international investment wi;l actually reduce provisions, and/or the investment is unprofitable. systematic risk (see Sections 4 (b) and (c)). Clearly, such regulations may affect th( (b) Integration with Social Security attractiveness to companies of funding pensions - and the generosity of provision - if it constrains Certain regulatory issues are raised by the managers in their choice of risk and return (i.e., treatment of the relation between private penSions forcing them to hold low yielding assets and nd social securitv. As noted, in Switzerland, the increasing their risks by limiting their possibilities schemes aim to dovetail so as to offer a declining of diversification).24 replacement ratio, the higher up the income scale the retiree is, thus ensuring maintenance of living This is not, however, the case in all countries.1' standards. In the US, by contrast, pension funds For example, United States pension funds are are allowed to aim for a fixed replacement ratio subject to a "prudent man rule" which requires (including social security) across the board, and the managers to carry out sensible portfolio hence low income earners may not receive a diversification; there are no limits on portfolio pension at all, despite the firm having contributed distributions other than a 10% limit on self on their behalf. This system is strongly criticized investment for defined benefit funds. United by Munnell (1984) as an abuse of tax privilege Kingdom pension fiinds are subject to trust law and social injustice. In the UK a compromise is and again follow the "prudent man" concept; reached, whereby pension funds may substitute they are not constrained by regulation in their for earnings related social security, but may not portfolio distribution except for limits on self- take flat rate social security into account, which investment (5%) and concentration. Dutch ensures a falling replacement ratio over the private funds face no restrictions,' except for a earnings scale, other things being equal. 5% limit on self investment, see Van Loo (1988). (an contrast, the public service fund (ABP) faces (c) Regulation of Portfolio Distributions strict limits, being able to invest only 5% abroad, and 15% in shares or real estate). Similar Quantitative regulation of portfolio distributions is prudent man rules are implicit in current EC imposed in a number of countries, with the proposals for a Pension Fund Directive, stressing ostensible aim of protecting pension fund security (consistent asset/liability matching, beneficiaries, or benefit insurers, although diversification and limited self investment), motives such as ensuring a steady demand for liquidity and profitability. government bonds may also play a part.21 Limits are often imposed on holdings of assets Other countries impose portfolio limitations, with relatively volatile returns, such as equities though the degree to which they bind varies. For and property, as well as foreign assets, even if example, Japanese funds face ceilings on holdings their mean return is relatively high. There are of certain assets (such as 30% for foreign assets 15 and for equities), which Tamura (1992) suggests (d) Funding Rules "(Inappropriately) imitate regulations devised for trust banking and life insurers". German pension Regulation of the funding of benefits is a key funds, besides a 10% self investment limit, aspect of the regulatory framework for defined remain subject to the same panoply of regulation benefit pension funds. Note that by definition, a as life insurers (4% limit on foreign asset defined contribution plan is always funded, holdirigs, 20% limit on equities, 5% on property). whereas with defined benefit plans there is a It Is arguable that these are particularly distinction between the pension plan (setting out inappropriate for pension funds given the indexed contractual rights to the parties) and the fund (a nature of their liabilities (Section (d)), though they pool of assets to provide collateral for the could be justified by the need to protect the promised benefits). When the fund is worth less insurance funds (Section (h)). They may be than the present value of promised benefits there contrary to the EC Capital Movements Directive, is underfunding, when the opposite, overfunding. depending whether they are judged to be Calculation of funding requires a number of "reasonable prudential restrictions". Resolution actuarial assumptions, In particular the assumed of this question is being sought in the Pension return on assets, projected future wage growth Funds Directive, as discussed in the (for flnal salary schemes) and future inflation (if conclusion.:n Note that by offering tax there is indexing of pensions). privileges to "booking", Germany and Japan effectively impose no limits on self investment of Minimum funding lim!ts seek to protect-ecurity book reserves (although the Germans do insist on of benefits against default risk bv the company, insurance of such reserves). given unfunded benefits are liabilities on the books of the firm, and therefore risk is Swiss limits are similar, if slightly less restrictive concentrated and pensioners (or pension insurers - than the Germans'; a 30% linit on shares, 50% see below) may have no better claim in case of for real estate and 20% on foreign assets. bankruptcy than any other creditor. Funding Scandinavian limits are in some ways even offers a diversified and hence less risky tighter, in that minima are also specified. The alternative backup for the benefit promise, as well Swedish funds have historically been obliged to as offering the possibility of unplanned benefit holl the majority of their assets in domestic listed increases if the plan is in surplus. Extra bonds, debentures and retroverse loans to protection against creditors of a bankrupt firm is contributors (although recent deregulations have afforded when the pension fund is an independent permitted limited investment in property, equities trust (as in the Anglo-American countries), or a and foreign assets, which some private schemes mutual insurance company (as in some have reportedly taken advantage of); Danish funds Continental European countries) and, as in most have to hold 60% in domestic debt instruments, countries, when self investmen. is banned or although since 1990 they have been allowed to severely restricted (see above). However, hold 20% in foreign assets. Some countries have funding does not increase perseral saving or switched to prudent man rules; Canadian funds wealth in an economic sense - it only affects the were strictly regulated till 1987 (when the prudent distribution of the cost of insuring those benefits. man concept was introduced) and have till There are usually also upper limits on funding, to recently faced limits on the share of external prevent abuse of tax privileges (overfunding). assets as tax regulations limited foreign Bodie (1990b) suggests that the three main nvestment to 10% of the portfolio. A tax of 1% reasons why firms fund, besides regulations per of excess foreign holdings was imposed for every se, are the tax incentives, provision of financial month the limit was exceeded. In 1990, it was slack (when there is a surplus) that can be used in announced that the limit would be raised to 20% case of difficulty, and because pension benefit over 1990 95. There is also a 7% limit on real insurance may not cover the highest-paid estate. employees. In the United States an important influence was i6 the Employee Retirement Income Security Act important for risk. (ERISA) of 1974, which provided for minimum standards of vesting and increased funding This "wind-up" definition of liabilities, the requirements, both of which increased the burden "solvency" level at which the firm can meet all its to firms of running a pension scheme. It also current obligations, is known as the accumulated introduced the Pension Benefit Guarantee benefit obligation (ABO). Indexation up to Corporation (PBGC) to guarantee (up to a limit) retirement, as is normal in a final salary scheme, benefits of funds in default, funded by gives the projected benerit obligation (PBO) contributions from all defined benefit plans; the which is not guaranteed except in the United funding requirement can be seen partly as a Kingdom although it is common in the protection for PBGC. (This has not prevented Netherlands (80% of members are covered). heavy financial claims on the PBGC, following Taking account of future obligations instead of several cases of default of underfunded schemes, purely focussing on current liabilities is likely to as discuissed further below.) Following ERISA, permit smoother levels of contributions as the the growth in pension funds slowed. Some firms fund matures, which may be better for the terminated their schemes, and the number of new financial stability of the sponsor. The indexed defined benefit plans initiated dropped. Some benefit obligation (IBO) assumes indexation a.'er firms swit(;.; to defined contribution plans; and retirement, which is not generally guaranteed in overall coverage cx ,ced to grow. Japan, the US or Canada but is in Switzerland, the Netherlands, Germany, and Sweden, and More recent chank. - wo .I.n'ted States regulations will be soon in the UK. (See Bodie (1991) for a have clarified funding, ruies by defining pensior further discussion of these concepts.) fund liabilities as the present value of pension benefit owed to emplo).ces under the benefit In the US, the accounting standard FASB 87 formula absent any riroiections of salary, focuses on the PBO, in contrast to the minimum discounted at a nominal rate of interest. funding regulations as described above. In Implicitly, these are the obligations of the fund if addition, overfunding in the US is limited to it were wound up immediately. Current estimates 150% of the ABO or the PBO, whichever is the suggest that 76% of pension funds are overfunded lower. These limits may have different effects. on this basis, with an average overfund of 74%. The 150% of ABO limit implies a rise in interent If pension assets fall below this level, the rates could prevent further funding, leaving the unfunded liability must be reported in the firm's scheme underfunded when interest rates fall. This balance sheet, and since they are senior debt, they is not the case for a PBO definition taking act as a major problem for the firm in raising projected rises in benefits into account, as long as funds. However, a surplus cannot be included on interest rates rise with expected inflation. the balance sheet (although it can be implicitly recouped via a reduction in contributions, see In Japan, as noted in Section 2, the traditional Section (e)). In this definition, indexing up to means of provision of retirement benefits was via retirement is not compulsory but only an implicit pay-as-you-go, with a special reserve account on promise, despite the fact most US schemes are the balance sheet as benefits accrue. The TQPFs actually final salary. This has an important and EPFs, as described above, must be funded influence on portfolio distributions, discussed at only up to the ABO, and there is reportedly very greater length in Section 4, since underfunding on little overfunding, partly because contributions this basis can be avoided by holding bonds; above the ABO are taxed. equities are only suitable for overfunded schemes. As discussed below, regulations now seek to In Germany, various laws or court decisions akin reduce the moral hazard of deliberate to ERISA have enforced minimum standards of underfunding by charging higher PBGC insurance funding for pension funds (while leaving open, as premia to underfunded schemes; but they do not in Japan, the choice of an unfunded book reserve take account of the asset composition of system) and what amounts to inflation indexing of underfunded schemes, which may be more pensions. However, although this implies funding 17 the IBO, it appears that provisions for indexation If it is too high, funding may be inadequate; if are taxed - only the PBO is tax &Tee. These too low, there may be overfunding and provisions were felt to be particularly corresponding abuse of tax privileges. In the burdensome, despite the relatively low level of Netherlands, where funding is compulsory, the German inflation, and, along with the decline in government sets a maximum real interest rate profitability of firms, helped blunt the growth rate assumption of 4%, and an assumption for wage of private pension schemes in the 1970s and early growth. Since in practice Dutch funds have been 1980s. (See Deutsche Bundesbank (1984).) able to earn over this level, surpluses estimated at 30% were present by 1990. A special levy of In the United Kingdom, the reform of the state 40% is to be enacted on such surpluses in excess scheme In 1978 had an important influence on of 15% of liabilities, to offset the implied tax private schemes (by setting a "guaranteed evasion. In the US, the accounting standards minimum pension" (GMP)) and enforced a degree FASB 87 and 88 have imposed common of funding sufficient to cover the GMP. standards. In Japan contributions are set However, funding above this level is not legally assuming a 5.5% nominal rate of return on fund required - although trustees are bound by their assets. In the UK and Canada the govenranent duty of care to ensure funding is in place - nor is accepts the (varying) judgement of the actuaries, any standard method of calculating funding and generally also allow for an assumption of imposed, or a requirement to include deficits in wage growth. company balance sheets. There is also no svstem to guarantee non-GMP pension benefits in the Finally, since many Danish funds (as well as a United Kingdom - partly for this reason proportion of funds in Switzerland and the Anglo- regulations can be less strict than elsewhere, and American countries) are defined contribution, the managers can offer a high return by taking a issue of funding does not arise. However, the higher level of risk. issue of limiting tax privilege does arise, and is dealt with via contribution limits or taxation of A plethora of more recent changes have limited returns. overfunding to 5% of projected obligations, (in practice, either the PBO or the IBO), including (e) Ownership of Surpluses discretionary provisions giving five years to remove surpluse-s; enforced a degree of indexation Ownership of surpluses in defined benefit pension (up to 5%) of pensions up to retirement for early funds is a key issue in a number of ccuntries, leavers (in contrast to the United States, Japanm particularly because predator firms may seek to and Canada); may make a degree of indexation strip surpluses after taking over another firm, & retirement compulsory;' have outlawed although also, as noted above, because the firm compulsory membership; limited tax-free may seek to recoup the funds for its own use. On contributions and benefits; enforced transferability the one hand, this may be seen both as an abuse of assets between schemes and may enforce equal of tax privileges and (more contestably) as seizing pension ages. (For a discussion of related issues assets held for the benefit of members. On the in the United Kingdom, see Blake (1992).) A other, it can be argued that if the fund is only a decline of the company pension fund sector is backup for the firms' promise of pensions, and if predicted, but there is little evidence of this to the firm is equally responsible for making good date. Few employees have left company any deficit, then the surplus should belong to the schemes, although there has been a sharp rise in firm. It is important to note that the funding personal pensions. And few companies have rules outlined above define the surplus. Note in closed their schemes, even though some have addition that such issues only arise for defined switched to defined contribution or made them benefit funds; in defined contribution funds there less generous for new entrants. is no surplus to strip. As noted, the interest rate assumed to be earnt on In the US, a 1987 law states that the employer assets is a key aspect of the funding arithmetic. owns all surplus assets so long as certain 18 standards are complied with. This, following the In Japan the surplus may neither be stripped nor second line of argument above, is seen as used to increase benefits, but used to operate economically reasonable since funds are purely a "welfare facilities". This puts the fund under means to collateralize a (separate) benefit pressure to smooth its income to ensure such promise. In other words, the employee has rights payments continue - which may entail inefficient to a pension, bnt not to the means of financing investment. those rights. However, there are limits to such ownership, as under ERISA, firms cannot use (f) Portability pension assets as collateral for loans. In the 1980s, many funds with surpluses were yesting, treatment of transfers between schemes terminated and the surplus taken by the sponsor and of prior service credits, particularly for (asset reversion). It can be argued that such defined benefit plans, have a key role to play in behavior implied breach of implicit contracts labor mobility, which in turn may be important between employer and employee. Later, for economic efficiency. Indeed, Lazear and substantial tax penalties were introduced to Moore (1988) estimate that labor turnover in the discourage this, although there is nothing to stop US would be twice as high in the absence of firms absorbing surpluses more gradually, by pension funds. This is because of the losses in taking contribution holidays. pension benefits that may be incurred by early leavers compared with those staying in one job In the UK, the surplus is again held to belong to (US calculations suggest that these may be as the company, which can be recovered by direct much as 50%, see Munnell (1984)). There are withdrawal (subject to a 40% tax) or by a obviously also problems of equity in such contribution holiday. However, court judgments patterns. Women may be particulariy vulnerable have severely restricted ability of predators to to such losses, as they change jobs more extract surpluses from takeover target's funds via frequently and spend fewer years in one job. winding-up or spin-off termination of schemes. The 1990 Social Security Act states that when a Solutions include shorter vesting periods, which plan is terminated, it shall be assumed to provide ensure that benefits are nonforfeitable on for indexation up to 5% inflation, thus reducing retirement; transfers permitted with full the potential surplus to be extracted. Moreover, allowance for benefits accrued; or service credits there is increasing support for arguments on the (in the case of final-salary based schemes) employee's side, namely that pension rights are indexed till retirement. Note that these problems not gratuities but part of a remuneration package do not arise with defined contribution, nor does earned by service. This point of view has been the last arise for career-average based defined supported by recent rulings of the European Court benefit plans; hence portability is an argument in that for the purposes of equal treatment pensions their favor; in contrast even with full indexation are to be considered as deferred pay (Goode to prices of accrued benefits in final-salary plans, (1992)). The logical conclusion would be to the early leaver loses out, because his real wage outlaw even contribution holidays and make would probably have been higher at retirement. employers much more restrained in funding. The arguments for portability, though strong, In the Netherlands, where the pension fund is an should not be overstated. Whereas the suggested executive body independent of the sponsoring reforms would make pension plans more attractive firm, usually in the legal form of a financial to employees, they may reduce the ability of institution, or in Switzerland, where it is a employers to use pension plans to manage their foundation with joint representation of employer workforces, and hence reduce their attractiveness and employee representatives on the board, to them. Frovision, when voluntary for the firm, ownership lies with that body itself. This means may thus decline. Low labor mobility is not the company cannot lay claim to the assets, always inefficient; higher labor turnover may although surpluses can be returned by reduced have adverse effects on the incentives for firms to contributions. train their labor forces, given the "market failure" 19 that employees may leave once trained, thus benefits are, indexed (up to 5%) prior to wasting the employer's investment. Countries retirement," employees have a right to a cash with "lifetime employment' such as Germany and transfer to another pension scheme31 in line wAth Japan have of course been conspicuously accrued benefits and restricted transfer circuits successful economically - although pinsion exist (e.g., in the public sector). However, arrangements which discourage turnover are difficulties may arise outside such circuits from probably best seen as a consequence or support the non standardization of the valuation methods for the system, rather than a cause. Nevertheless, for liabilities. In the US, Japan, and Canada past could be suggested that a small aniount of benefits are not indexed; however, there is a esistance to labor mobility arising from pension transfer circuit in Japan enabling workers in provision, so long as it does not lead to unfair Employee Pension Funds to shift their contracted deprivation of pensions, may not be entirely out social security benefits (which are indexed) ndesirable. only between emrployers. In the US, service transfers are available as a lump sum, which Vesting standards in the US under ERISA give poses the risk that tax advantaged pension assets three alternatives; however, the most common Is will be used for other purposes. In Denmark, to demand that companies offer 100% vesting differing medical examination requirements after 10 years of service (the alternatives are 25% between schemes are reported to give rise to vesting after 5 years, rising to 100% after 15 transfer difficulties. years or 50% vesting when age and service add to 45, increasing to 100% five years later). Recent Transnational moves of employment pose legislation will reduce vesting to 5 years. There particular problems for pension funds, given the are no common vesting standards in the EC. differing tax treatment that may make transfer They vary from 10 years in Germany and 5 years impossible. The EC have found this insoluble so Denmark to 2 years in the UK and one year in far, despite the premium put on international e Netherlands. In Sweden the ATP scheme is a labor mobility. uational one, so the issue of vesting does not arise 'though it does for the ITP/STP). The most (g) Internal Transters restrictive countries are Switzerland and Japan, both of which appear to assume 'lifetime Difficulties of early leavers, who implicitly employment". In the former, vesting is graded subsidize those remaining till retirement, are not .ween S and 30 years of service (for payments the only case of potentially inequitable internal excess of the legal minimum, which is vested transfer within defined benefit funds. As noted mediately), while in the latier, vesting takes 15- by Riley (1992), in countries such as the UK, years, with early leavers typically penalized final salary schemes give incentives for managers Jhough vesting is quite short for involuntary to award themselves large salary increases in 'heir irement, and for lump-sum distributions as last year of employment, thus benefiting opposed to annuities). Particularly in Japan, particularly at the expense of workers forced into restrictive conditions are seen as socially early retirement (given the expense to the pension desirable, to support "lifetime employment". fund), early leavers and those workers (such as manual A'orkers) whose earnings peak in mid regards service transfers, this is a career. More generally, if contribution rates are straightforward matter in countries such as the based on expected average increases in salaries, Netherlands, where benefits are generally indexed contribution rates may fall short of costs for those to and beyond retirement and tran'sfers occur whose salaries rise faster than the average, and hrough portability clearinghouses called transfer vice versa for slow climbers. Finally, given that circuits. A requirement to join the clearinghouse the rate at v hich benefits are accrued rises as the s that the fund be ind,xed and pensions based on worker nears retirement, there are strong inal salary. In Sweden, there are again no incentives for firms to retire workers early, which 3roblems for the ATP in this context, as the may not be economically efficient. Understanding em is a national one. In the UK, too, past of these issues may be hindered by the complex 20 rules of a defined benefit plan. they fall below a certain minimum funding level. It is hence essential that the insurer have access A related equity problem, and implicit form of the assets, the assets have a defined market value, transfer, was that US funds were traditionally and that there are agreed standards for only for managers, despite the use of income determining minimum funding levels. Analogous from the firm as a whole to contribute to their to bank capital, it is also desirable that there be pensions (i.e., not merely reflecting their own cushion of over funding to protect the guaran^ productivity), and benefit of tax privileges (i.e., and frequent auditing. A system relying on basically schemes were a means of tax-avoidance monitoring might not be efficient with illiqu for managers). This was clamped down on by fund assets, as their wide bid-ask spread imposes ERISA, which insisted that all full time workers costs either on the sponsor or the guarrntee over 21 should be eligible to allow tax agency. A second approach is restrict1in_A deductibility. But as noted in Section (b), asset choic of pension funds to ensure an upper inequity may still arise from differing treatment of bound on the risk of the assets servsng as social security pensions. collateral for the promised benefits, for example, by insisting on immunization of assets equal to (h) Insurance the guaranteed benefits, see Section 5(a). A third is setting the premium rate for the guarantee in As noted, insuranc of defined benefit pensions linewith thrisk, which depends in turn on the against default risk for the sponsoring firm is a variance of the value of the collateral and the tim feature of most of the countries studied. Note between audits (which allow the fund to chang that insurance of benefits of defined contribution its risk exposure adversely.2 plans is unnecessary, as there is no fixed pension right to guarantee (although investment rules may The US example (Bodie (1992), Bodie and still be useful to protect members from risk Merton (1992)), where the Pension Benefit concentration, and insurance may be needed to Guarantee Corporation (PBGC) was set up as a protect members against fraud, etc). Also compulsory insurance scheme to guarantee basic funding of defined benefit obligations - or at least retirement benefits, shows the difficulties that assurance of seniority of claims against other arise when such controls are not properly applied creditors in the case of bankruptcy - are the first PBGC premia have traditionally been non-risk line of protection of members against default risk. related, thus encouraging risk taking; minimum Insurance provides a second line of defence. funding rules have proved ineffective, and indeed till recently plan sponsors could freely transfer But any system of guarantees, including deposit some of their unfunded peision liabilities to the insurance as well as pension insurance, faces the PBGC by voluntarily terminating an underfunded difficulty that it may create incentive structures plan (subject to a provision allowing PBGC to leading honest recipients to undertake excessively take 30% of the employers' net worth to make up risky investments, which in turn give the risk of for underfundinge); given lack of control over large shortfall losses to the insurer. In other pension fund management, firms in financial words, losses may not arise merely from fraud or distress have faced particularly strong incentives incompetwnce but the incentive structure itself, to take risks and reduce funding; courts have What is needed are means to control risk, which ruled that the PBGC has no better claim on assets could (Bodie and Merton (1992)) include an of bankrupt firm with an underfunded pension appropriate mixture of monitoring, asset plan than other creditors; and fragmentation of restrictions and risk-based guarantee premia. We regulatory authority and conflicts of interest consider this a useful and flexible framework for among government departments (as discussed in analysis of the regulation of guarantees. Section (k) below) weaken monitoring. Finally the PBGC was set up to serve goals other than In the case of pension funds, controls could, first, purely protecting pension benefits, namely include monitoring of the market value of pension revitalization of depressed industries by assuming assets, with the right to seize and liquidate them if part of the burden of pension benefits, and 21 preservation of defined benefit plans against the defined benefit company plans, while imposing trend to defined contribution.'3 These further greater risk on the beneficiaries in cases such as dilute the effectiveness of its control mechanisms. Maxwell, as discussed below. However, the absence of insurance in the UK need not exclude As a result of these difficulties, plans that are discretionary assistance by the government on a terminated are often vastly underfunded (typically case by case basis, which may create less moral 60%), having been only 20% underfunded five hazard than a guarantee scheme (there are strong years before. Sponsoring firms eitzer minimize paral!els with the issue of deposit insurance vs pension contributions directly or encourage early lender of last resort for banks, see Davis (1992)). retirement of workers whose pensions are not In addition, defined contribution schemes run by funded. Accordingly, the PBGC has a deficit insurance companies are covered by (mutual) astimated in late-1991 to be over $2.5 billion," insurance compensation arrangements, covering and is paying an average of $2352 per year to 90% of the investment. 325,000 retirees in 1,700 failed plans. Meanwhile Smalhout (1992) suggests that (;) Fraud companies such as Chrysler have unfunded iabilities of $4.4 billion, a quarter of schemes are Protection against fraud has come to particular underfunded (to a total of $40 billion) and the prominence in the UK, given the Robert Maxwell worst 50 companies account for $21.5 billion in case. Large quantities of his companies' pension unfunded liabilities. These data suggest a fund assets were lent to private companies owned )otential liability on a 'Savings and Loan' scale by Maxwell against poor security, or were see Davis (1992) and his references for an invested directly in them. When the private utline of the S and L crisis and its relationship to companies became insolvent, the assets were lost. eposit insurance). The fraud was partly concealed from fund trustees by the fund manager or stock custodian - both Other countries having guarantee schemes, such again controlled by Maxwell - but was partly as Germany, have tended to impose extremely legitimate self investment carried out with the severe asset restrictions on funds to protect the knowledge of the (pliant) trustees. In other insurance fund against loss, while simultaneously words, it partly revealed the inadequacy of legal imposing higher costs on plan sponsors than provisions, as well as vulnerability of pension would be necessary in the absence of guarantees. funds to fraud. The case has cast doubt on the The Netherlands offers partial insurance. If the use of trust law as applied in the UK, as the employer is unable to pay contributions, due to means of redress - civil action against trustees by bankruptcy or any other reason, the Industrial members once things go wrong - were seen as Insurance Board will pay contributions for up to a inadequate. This is especially as members lack year. There is also a form of insurance for the prudential standards against which to monitor the employee which is absent in Anglo-American fund and trustees - and have no regulatory body untries, whereby if a worker over 40 becomes to do so on their behalf - and may find it difficult employed, the Pension Insurance Advance to interpret performance measurement data. nancing Fund will pay supplementary pension Also, except in cases of theft and fraud, there is tributions as long as the employee has the usually an indemnity cluse in case of court action t to wage related benefits under the against trustees for breach of fiduciary rules. employment Act. In Japan, participation in the This-leaves the employer to resolve the problem - )ension guaranty programme for EPFs is and it may be insolvent. In contrast, in the US voluntary, with lower guarantees resulting from fiduciaries in such cases may face heavy personal non payment. But all firms reportedly do liabilities. contribute, perhaps as a consequence of social consensus. Independent custodians,36 less leverage by the employer over the trustees, better independent the UK, the need for portfolio regulations is actuarial information for trustees, more employee iated by absence of a guarantee scheme for trustees, as well as limits on self investment and .22 more frequent checks on a higher standard of and investment standards as well as dealing with minimum funding, are among other proposed cases of fraud, while the Internal Revenue Service remedies. Independence of custodians, both from sets maximum funding rules to prevent abuse of trustees and from fund managers, is already the tax advantages. Then the Pension Benefit rule in the US. Some in the UK hive argued for Guarantee Corporation collects insurance premia an insurance scheme similar to PBGC in the US. and pays benefits but has few enforcement The discussion above suggests moral hazard is a powers. So, for example, the tax authoritie strong counterargument against general insurance would prefer minimal funding to prevent loss o of benefits, but this need not rule out insurance tax revenue, while the insurers would seei aga!nst fraud. maximum funding to prevent large insurance claims. Moreover, the tax authorities can grant (J) Disclosure to Members contribution waivers to firms in financial distress, which leads to underfunding of pension plans, Standards of information for members has come against the irterests of the PBGC. Meanwhil to prominence recently in the UK, and the fund trustees are responsible for ensuring funding Maxwell case (above) is likely to bring it further is in place for beneficiaries and have to! to the fore.37 Under ERISA in the US, pension demonstrate in an audited annual report of incom funds must provide each plan participant with a and assets filed with the IRS that they h- summary of the Annual Report outlining the plan managed the fund prudently. Also a master and its administration, information on the right to custodian has to be appointed to oversee receive pension benefit, and the status of fulfillment of ERISA requirements, kee individual pension benefits. In the UK under the appropriate records, and provide security agains 1986 Pension Schemes Regulation, trustees are prohibited transfers. required to disclose trust deeds and rules on request; annual reports must be provided free of In the UK statutory pension fund regulation is charge, covering information such as the names of again administered by different bodies, namely trustees, Lztuaries and fund managers, number of the Occupational Pensions Board on behalf of the beneficiaries, contributions, increase in benefits to Department of Social Security and the Pension current pensioners,- distribution of assets, an Schemes Office for the Inland Revenue.3' As in actuarial certificate saying to what extent the the US, the tax authorities are concerned to avoid scheme is financially viable, presenting results of overfunding, but the Pension Board only checks performancemeasurenient of fund managers and on a three yearly basis whether assets are how they are remunerated (see Section 4 (d)). sufficient to pay the minimal state-guaranteed Every three years a more detailed valuation report pension (GMP).3 Otherwise, as noted, there must be included, giving a view of long term are no minimum funding rules. The duty to viability. It is particularly crucial that members check funding is in place belongs to trustees, as in receive such information in defined contribution the US, (they are supposed, under common law, plans. In Switzerland, similar to the UK, audited to "act in the best interests of the annual accounts and an individual benefit beneficiaries"') but the wider bounds offered by statement must be made available to members. In the funding rules give more responsibility to them Japan Tamura (1992) reports that disclosure is to stand up to employers in insisting a scheme be vestigial; members only receive occasional funded. There may be difficulties where trustees circulars. are not independent of the employer, which may be the case through a variety of channels, siace (k) Structure and Mechanics of Supervision employers as well as employees and pensioners are beneficiaries of the trust." (Noble (1992)). Effectiveness of pension fund regulation is This was the weakness that partly enabled the influenced by regulatory structures and procedures Maxwell fraud to occur (he was able to persuade which in several countries are somewhat the trustees to agree to imprudent but legal self- unwieldy. For example, in the US the investment), and also can lead trustees to accept Department of Labor oversees minimum funding too readily the case for removing surpluses via 23 contribution holidays, etc. Also there may be (I) Is There a Consensus on Regulatory conflicts of interest between scheme members and Pract:ce? employer, or pensioners and working members,"2 that trustees may find it difficult to It will be apparent that there is no overall resolve. international consensus on good regulatory practice. (See the summary in Table 6.) On the In Canada, apart from federal taxation provisions, one hand, there appears to be reasonable regulaticn is carried out at the provincial rather agreement on tax provisions and ownership of than the national level, and hence pension law can surpluses. For example, most countries accept differ between provinces (in practice, Ontario the arguments for the expenditure tax treatment of tends to be the leader). This can ereate particular pensions, although there are moves in some cases problems for employees moving between jobs in to level the playing field by granting si,nilar different provinces - which in turn foreshadows treatment to other forms of saving, or even to possible future difficulties in the EC. In impose comprehensive income taxation on pension Switzerland, too, regulation is generally carried funds. Again, it is generally accepted that surplus out at cantonal level, except for "large' assets belong to companies, although their access companies, but the federal authorities are tending to them is generally restricted, given the potential to oversee and harmonize cantonal supervision. tax abuse. On the other hand, there are strong divisions on portfolio regulations (prudent man vs In most Continental European countries such as portfolio restrictions); on funding (unfunded vs the Netherlands, regulation is carried out by a ABO vs PBO vs IBO, as well as regulatory rules single statutory authority, the Insurance vs trustee responsibility); on insurance; and on Supervisory Board. Pension funds are legally vesting and service t.ansfers (between countries obliged to provide the Board with detailed insisting on rapid vesting and those assuming information annually on the benefit payments and lifetime employment). Issues of fraud and investments of the fund. It ensures that the information disclosure have come to the fore only commitmnents of the pension funds are sufficiently in some countries. Another important aspect on covered by their assets. It also involves itself in which there is no consensus (covered in Section more general structural issues. If the Board finds 4(a)) is regulation of the indexation of benefits. procedures or regulations unsatisfactory, it can apply social pressure by making a public There are no obvious right answers to many of complaint. In practice, this is rarely necessary. these issues. Historical development clearly plays In some countries such as Germany, the a major role. When reform is feasible, the supervisors also check that portfolio regulations "correct" approach depends crucially on the are complied with and require a five-year business tradeoff desired between costs to the company and plan. In Denmark, there are three yearly associated effects on competitiveness on the one actuarial reports. hand; and the interests of the recipients, the perceived importance of labor mobility and the It will be noted from this description that the need to avoid insurance losses on the other. mechanics of supervision generally entail reliance However, some a priori suggestions can be made on annual reports and accounts prepared by (amplifying comments made above). auditors and full actuarial reports at longer intervals. However, the Netherlands is unusual in For example, it is notable that most countries with that the authorities conduct on-the-spot inspections strong portfolio regulations offer lower returns of all funds every 10 years. In the US, the than those with prudent man (Section 4(c)) albeit Department of Labor runs computer checks to also with lower volatility. Only in the case of identify plans needing further investigation (or self investment would modern portfolio theory investigations may be triggered by complaints by agree with the need for quantitative portfolio members). 250 investigators are employed. regulation (although its avoidance may be implicit in a prudent man rule). Funding rules tailored to the nature of the benefits (as in the Netherlands), 24 such as the FBO or IBO in the case of final salary less important if there are defined contribution schemes wouid seem to offer greater security to and not defined benefit funds. These need to be members than the alternatives of no funding rules, weighed against the superior employee retirement only covering state pensions, only the accrued insurance and various benefits to employers (such obligation or relying on the fallible independence as lower labor turnover and ability to take of trustees. It also ensures smoother funding contribution holidays) offered by deflned benefit patterns for the sponsor as the fund matures. plans. Insurance against fraud would seem to increase security without the effects of moral hazard (or Meanwhile regulatory structures and procedures need for tough restrictions) implicit in overall appear to have developed piecemeal in a number guarantees. Such overall guarantees may be of countries. It could be suggested that the Dutch inferior to discretionary bailouts of failed plans, have a reasonable model (one supervisor, annual reserved for extreme cases. When insurance of checks on the adequacy of funding, overview of benefits is chosen, a mixture of controls on risk plan rules, on-site inspections, etc.) for others to taking, as recommended by Bodie and Merton follow. Finally, given the long term nature of (1992) would seem to be justified. pension schemes, there is mucl. to be said for continuity of the regulatory framework' It is also worth noting that many of the issues Retrospective changes in regulation affecting (vesting, transfers, funding, ownership of liabilities are particularly undesirable, given their surpluses, guarantees of benefits) are absent or likely impact on corporate finances. 25 4. PERFORMANCE This section reviews the relative levels of benefits mid career. provided by the funds, followed by an examination of their portfolio behavior and its In the United States, where defined benefit underlying determinants. Effects of the latter on schemes are again final salary, there are often overall risks and returns, and the influence of discretionary pension increases to compensate for fund management on fund behavior and costs, are inflation after retirement, although explicit also assessed. indexation for inflation is less common. Indeed, as pointed out by Bodie and Merton (1992), even (a) Beneflts and Contributions the indexing of pensions pi2r to retirement only holds to the extent that the employee continues to Comprehensive data on benefits paid are not work for the same employer; his wage keeps available, and would in any case be distorted by pace with general wage inflation; and the factors such as the inclusiveness of the statistics employer continues with the same plan. Early and the degree to which pension funds cover the leavers' accrued benefits are not indexed. Unlike different sectors of the income distribution. For the other countries, preretirement cashouts from a example, Dailey and Turner (1990) show that pension plan must not necessarily be invested in average retirement benefits in the US in 1989 a-zother pension scheme, which raises the issue of were $6359, but as noted by Munnell (1992) they potential misuse of the tax advantages for non only beneflt a 'relatively privileged subset of the retirement expenditures. population". In Switzerland average benefits are $6236 and Canada $5100, but the latter may be In Canada even discretionary increases of benefits boosted by inclusion of public sector schemes. to allow for inflation are relatively rare (93% of France ($3203) and Japan ($2304) appear low, the private sector participants are in plans with no former due to coverage of low paid workers and formal inflation protection); a fixed income the latter because only annuities and not lump related to final salary is promised in retirement. sums were captured. We suggest it may be better In Germany most pension funds promise an to gauge the nature of benefits offered to the plan amount dependent largely on duration of participant more directly. employment; final salary schemes are less common than in the other countries. However, For example, in the UJnited Kingdom the nature indexation is mandatory. In Japan benefits tend of benefits has changed since the 1960s. Final to relate to years of service and final basic salary, salary based defined benefit plans, 75% of whose but the ratio to the latter tends to be less than in members benefit from guaranteed indexation, the Anglo-American countries; (such benefits are cover all public sector and the majority of private often taken as a lump sum). Only the part of sector beneficiaries. Indeed, indexation of pensions replacing social security is indexed. In b- nefits up to an inflation rate of 5% may soon Denmark, where funds are in any case defined become mandatory.43 The typical replacement contribution, there is little explicit indexation. ratio after 40 years is 50-66%. Lump sum withdrawals at retirement are permitted, up to Inflation indexation of pensions is of course a key 150% of final salary. Meanwhile in the policy topic in its own right. The move from Netherlands 90% of pension plans are defined career average to final salary pension plans in benefit (usually paying 70% of final salary), and some countries can be seen as an attempt to 90% of members receive inflation protection. correct for effects of inflation prior to retirement Pensions in Sweden and (in practice) in (leaving open difficulties for early leavers, and Switzerland are again indexed. Swedish pensions the issue of indexing after retirement). However, are based on best years of income (suitably in Japan the dependence of pensions on basic indexed) and not final salaries, which may offer salary and not full remuneration may mean that superior equity between managers and manual pre-retirement indexation is imperfect. And as workers, since the latters' earnings may peak in noted above, post retirement benefits are rarely 26 fully indexed in the US, Canada, Japan, and condition of the fund. EPFs are more flexible - Denmark. Of course, social security pensions are contributions are set to obtain the promised invariably inflation indexed. benefit given an assumed nominal return of 5.5%. Bodie (1991) suggests that automatic indexation The distribution of contributions between may be avoided by employers and not pressed for employer and employee varies widely, although by employees in countries such as the US because its economic implications need not be significant of lack of an asset providing an inflation hedge (employers can reduce salaries to offset their (unlike index linked gilts in the UK); because via contributions). The proportion paid by the social security, real estate investment, etc, employer is 100% in Japan, 89% in Germany, individuals already have enough inflation 87% in the US, 70-75% in the UK, Canada, and protection, and providing it would increase costs the Netherlands, and 58% in Switzerland. unacceptable for young workers; or due to money illusion. Bodie finds the third explanation most Administrative costs of pension funds are only plausible. However, Blake (1992) argues that if available for a selection of countries, ai;. are not real wages and hence contributions rise at 2-3% directly internationally comparable. However, per year, and fund managers can obtain real some patterns do emerge from US data (Turner returns of 2%, indexation to prices should be and Beller (1989)), namely that costs are higher easily attainable. Vittas (1992) disputes this for small funds than large, and defined benefit calculation and suggests that real returns need to over defined contribution. For funds with assets ed real earnings growth by 2-3% for of Sl million in 1985, costs were 2% of assets for indexation to be possible at reasonable cost. defined benefit, and 1.4% for defined Section 4(c) suggests that real returns of 2% in contribution. For plans with assets of $150 excess of real wage growth are attainable in most million, tha costs were 0.7% and 0.2%. countries. Anecdotal evidence for the largest funds in the UK suggests figures as low as 0.1 %. Evidence Given the burden on employers as outlined, from several countries suggest that the costs of policymakers in most countries have tended personal pensions are much higher than for historically to avoid legal provisions enforcing company plans, given economies of scale, indexation, even where, as in Switzerland and the advertising, commission costs, etc. Netherlands, de facto indexation tends to apply. However, there are signs that this is changing, as The features of pension funds outlined in Sections in the UK laws will shortly enforce indexation for 2-4(a) are summarized in Table 6. p to 5% inflation (they already insist on pre- retirement indexation of accrued benefits), and as (b) Portfolio Distributions noted, indexation is mandatory in Germany and weden. The portfolio distribution and the corresponding return on the assets held are the key determinant Contribution rates are generally limited by tax law of the cost to the company of providing a given to around 15% of salary, except in Denmark, pension benefit' (although obviously the where they are unlimited,' balanced by the real prevailing nature of benefits in a given country, interest rate tax on the funds' yields. For as outlined above, also influence the overall cost). example, in the UK, total contributions are This section discusses portfolio distributions per limited to 17.5% of the employee's salary, and se; the next assesses their implications for e maximum employee contribution is 15% of performance. salary. In Sweden, contributions are 13%. In countries such as Germany, where private pension Changes in portfolio distributions of pension schemes have limited "supplementary" objectives, funds over the period 1970-90 are shown in detail ontributions are typically much lower, around in Tables 8 to 20 and summarized in Table 21. It 1.5% of salary. In Japan, contributions to TQPPs should be noted that the data generall) exclude are limited to 3.2% of salary regardless of the pension funds administered by life insutance 27 companies. The data for the Netherlands exclude significantly higher than in other countries. the public pension scheme (ABP), which invests Germany also has the lowest and least volatile virtually all its funds in loans to the government inflation rate. Meanwhile, international and local authorities, or government guaranteed diversification' in equities also offers sizeable loans to private firms. The data are from national real returns, at generally lower risk than sole flow-of-funds tables and are not always at market focus on domestic shares, despite exchange rate value (e.g., United States bonds and Canadian risk. equities are at book value) and may exclude certain assets (e.g., United States property). To In principle, the portfolio share of ligiud asets maintain comparability, asset holdings combine can U. small because withdrawals are predictable domestic and foreign assets. Hence equities in (the "contractual annuity' aspect noted in Section Table 14, for example, are both domestic and 1). German, Japanese, Dutch, Swedish, and foreign. (In most cases, foreign asset data was Danish funds have accordingly always held less obtained from separate sources.) Finally, in than 4% of assets in this form. The higher levels recent years the data may be partly misleading, that have ofien been observed at various times in given increased use of derivatives. A suitably other countries (Table 8) are therefore likely to hedged equity may have the characteristics of a reflect high market returns on liquid assets bond (see the discussion in Section 5(a)) - relative to other assets. This was particularly true although ownership of the company clearly for the United Kingdom and the United States in remains with the equity holder. 1974 when the equity market fell sharply. The United Kingdom has returned to roughly its pre- As background, estimates of real total returns and 1974 level of short-term assets, while Canada and their standard deviations for 1967-90 are shown in the United States have built them up considerably. Table 7. The table was constructed using annual This has largely resulted from the accumulation of average data on summary or market indices of market paper, though deposits have grown interest rates, yields and asset prices drawn somewhat (Table9). largely from the BIS macroeconomic database. No allowance is made for taxation or transaction These increases coincided with deregulation and costs, which would affect actual returns to expansion of short-term markets (Stigum (1990)). investors. Owing to lack of data, a number of Swiss funds have always held a high proportion of bond price indices were estimated from changes liquid assets, which has latterly expanded to 12%, in yields. This is of course only a sample over a largely in the form of short term money market relatively short period and does not necessarily instruments (Table 104'), due to the shape of the ndicate long run expected returns. For example term structure. the United States real equity yield is thought to be over 8% higher than the risk free rate. Bonds (Table 11) constitute over two thirds of (Reference: Ibbotson and Sinquefield (1990)). pension fund assets in Sweden and Denmark, largely due to pcitfolio regulations and the nature Among the notable features of the data for of the domestic financial markets. As shown in domestic assets are that the highest return - and Table 6, 60% of Danish assets must be invested the highest risks - are generally offered by in domestic debt instruments, while the majority equities, followed by property. Both are generally of Swedish assets are to be in listed bonds and in excess not only of inflation, but also - crucially debentures (and retroverse loans). In the United for final salary plans - the growth rate of average States, where regulations make it optimal to hold earnings. Bonds in most countries offer a much a large proportion of bonds despite their weakness lower real return, and generally a highly volatile as an inflation hedge', bonds still form around return (note that the calculations are based on 40% of pension funds' portfolios. Levels similar annual holding period returns, ie including capital to the US are maintained in Canada and Japan, gains and losses arising from changes in interest while being only 30%e in Germany and rates). The main exceptions are Germany and Switzerland. In contrast, the bond share has enmark, where real returns on bonds have been fallen sharply in the United Kingdom, from 50% 15bk 7: Chwaaedes of rea NM reOwas, 1 -9 Men (_adad devun* of read bIaIIhogdpeg od rem (danel cawncy) Per nt W WIled Gummy JaPan Canab Nub.h Swam Dk Swilnerhu hWMI SW. K}gim Fm Laons 3.5 (2.9) 1.4 (5.0) 5.3 (1.9) 0.9 (4.3) 4.0 (3.7) 3.8 (3.6) 3A (3.1) 6.1 (3.6) 2.5 (2.0) 2.6 (3.2) Morwaaes 2.0 (13.4) 2.0 (5.2) 4.7 (1.4) 3.0 (4.9) 2.4 (12.3) 4.3 (2.6) 2.6 (3A) S.J (3.7) 1.3 (.3) 3.7 (2A) Eluides 4.7 (14.4) 3.1 (13.9) 9.5 (20.3) 10.9 (19.4) 4.5 (16.5) 7.9 (23.2) 3.4 (3.3) 7. (27.5) 6.2 a2.3 ) 9.4 (26.9) Bouds -0.6 (14.4) 0.8 (11.1) 2.7 (14.9) 0.2 (12.3) 0.0 (12.1) 1.0 (13.1) -0.9 (3.S) 3.4 (16.1) -2.2 (17.6) 1.0 (13.1) Shmt4erm a*m 2.0 (.5) 1.7 (4.9) 3.1 (2.1) -0.5 (4.6) 2.5 (3.3) 1.6 (4.0) 1.3 n3.) 1.6 (1.3) 1.2 (2.2) 24 (3.4) Propeny 3.4 (6.4) 6.7 (11.4) 4.5 (2.9) 7.2 (6.3) 4.6 (6.2) 4.6 (15.0) - Foreign bods 1.5 (15.2) -0.3 (16.0) 3.2 (12.3) 1.5 (14.9) -1.1 (125) .0.S (11.7) 0.0 (13.3) -1.7 (12.4 ) -1.6 (14.0) O. (13.2) Foreip eqiices 9.1 (17.1) 6.5 (16.4) 10.4 (143) 7.8 (19.6) 6.6 (14.9) 6.4 (14.4) 7.3 (14.5) 6.1 (14.5) 6.1 (16.3) 7.2 - (13.5) Memoanhm kemn: Inflatio (CPIu 6.0 (3.0) 8.9 (5.3) 3.5 (Z.1) 5.5 (5.3) 6.4 (3.0) 4.9 (3.1) 7.7 (3.0) 7.7 (3.2) 4.0 (2.5) 7.1 (4.1) Redempion yield an gaomerm bonds 2.6 (3.1) 1.9 (4.3) 3.9 (1.1) 1.0 (4.4) 2.9 (3.0) 3.2 (2.7) 2.3 (2.3) 5.3 (2.4) 0.9 (1.3) 3.3 (2.3) Real eaning growth 0.2 (2.1) 2.6 (2.5) 4.0 (3.1) 4.2 (4.2) 1.7 (2.8) 2.4 (3.2) 1.5 (35) 2.3 (3.6) 1.9 (2.1) 27a Table So Short term assets (as a percentage of assets) 1970 1975 1980 1985 1990* UK 4 8 5 4 7 Us 1 7 8 10 9 Germany 3 3 2 1 2 Japan 2 1 2 4 3 Canada 5 6 9 10 11 Netherlands 3 3 2 2 3 Sweden 0 0 0 1 3 Switxerland 7 6 6 7 12 Denmark 3 3 2 1 1 * 1989 for Canada 1987 for Denmark fable 9: Market paper (am a percentage of assets) 1970 1975 1980 1985 1990* UK 2 5 3 1 1 uS 0 3 3 2 3 Germany - - - - - Japan - - - _ _ Canada 2 2 5 6 10 Netherlands 2 1 1 1 1 Sweden 0 0 0 1 3 Switzerland 3 2 4 6 10 Denmark - - - - * 1989 for Canada 1987 for Denmark 2 pb Table 10: Deposits (as a percsitage of assets) 1970 1975 1980 1985 1990* UK 2 3 2 3 6 us 1 4 5 a 6 Germany 3 3 2 1 2 Japan - - - - - Canada 3 4 4 4 1 Netherlands 1 2 1 1 3 Sweden 0 0 0 0 0 Switzerland 4 4 3 1 1 Denmark - - - - - * 1989 for Canada 1987 for Denmark Table 11t Bonds (as a percentage of assets) 1970 1975 1980 1985 1990* UK 32 24 24 20 14 uS 45 42 41 40 36 Germany 19 18 24 32 25 Japan 12 34 51 49 47 Canada 53 50 50 49 47 Netherlands 15 13 10 19 23 Sweden 76 76 74 77 84 Switzerland 25 24 28 31 29 Denmark 72 72 63 67 67 * 1989 for Canada 1987 for Denmark 27c 28 of gross assets in 1966 to 14% in 1990. bonds casts some doubt on their efficacy as a means to protect against future risks to social This may reflect different liabilities; in countries security, given the bonds are to be repaid by the such as Canada, only nominal returns are taxpayer in the same way as they would have to promised after retirement, while in the United finance future social security burdens. Kingdom a degree of inflation protection both before and after retirement is expected. Similar Except in Germany, where the bank bond market prornises are made by the Swedish supplementary remains buoyant, as well as Sweden and national scheme, despite which the bond share is Denmark, where a large proportion of bonds are extremely high, suggesting an inefficient portfolio issued by credit institutions for housing finance, allocation. The fall in the UK bond share also private bond holdings of pension funds have reflects alternative means of diversification; after tended to decline (Table 13). Nevertheless, in the abolition of exchange controls UK funds sold United States the share remains over 20%. The bonds to buy foreign assets. A decline has also share of US funds in total corporate bonds been observed in the Netherlands, from 20% in outstanding has also fallen. The general decline 1966 to 10% in 1980, although it has recovered partly reflects availability, but also a shift into since, with the increase in public debt issue. Van public bonds (which are more liquid) and equities Loo (1988) relates this to higher returns and (which offer higher returns). Notably in the UK longer maturity (and thus better matching to and US, pension funds have taken advantage of liabilities) by private placement loans, while the regulations permitting equity holding and have recent recovery in bond holding corresponds to a thus been able to profit fromi patterns of relative narrowing of the yield differential. Patterns of returns which have favored equities over bonds bond holding may also relate to asset returns (see (Table 7). Table 7); where (Dartly owing to low and stable inflation), real returns on bonds and other flxed Since in many countries pension funas may offer interest assets are relatively high in Germany, real returns (either in the sense of indexation to Denmark, and the Netherlands while in other wages before retirement, or in some cases countries boads have performed poorly. Swiss indexation after retirement), they consider it is bonds have done particularly badly, as have those sensible to invest in 'real" assets such as equity in Sweden, where bonds have a high portfolio and real estate.'9 share. Much of the past growth of Japanese funds' bond holdings may reflect the high share As shown in Table 14, the share of eguities in of public bonds purchased, under government most countries has grown significantly over the pressure, a practice that has now been abandoned. period shown, albeit at levels in 1990 varying from 19% in Sweden to 63% in the UK. As The share of goverment bonds in pension funds' noted, German funds are limited to a maximum of portfolios has grown significantly since the mid- 20% by regulation, Swiss and Japanese to 30% - 1970s in all of the countries studied except the hence at 18% and 27% respectively in 1990, the United Kingdom where there was a contraction in German and Japanese ceilings are almost binding. the supply of public debt in the late 1980s (Table An exception to the patterns of growth has been 12). The decline in the UK occurred despite the the United States, where levels in 1990 were introduction of index linked bonds, which should only slightly above those in 1970. In the in principle be an attractive means of pension Netherlands shareholding remains low - 20% - fund financing (depending on the real yield despite absence of portfolio restrictions. This relative to growth of average earnings). The may relate to the narrowness of the domestic increases in other countries parallel the size of equity n.trket and risk aversion of pension fund government deficits and corresponding ex ante trustees. Proportions in the United Kingdom, the real returns on such bonds (although, as shown in United States, and Canada were strongly affected Table 7, such returns have not always been by price instability in the mid-1970s whereas the realized er oost). Investment of a fifth of the 1987 crash had little effect on equity proportions. Swedish quasi-public funds' assets in government Reflecting portfolio regulations, the equity share Table 12t Governoent bonds (as a percentage of assets) 1970 1975 1960 1985 1990* UK 16 18 22 18 11 Us 7 9 14 22 20 Germany 9 6 13 20 17 Japan. 11 16 15 13 5 Canada 38 34 40 42 39 Netherlands 10 7 5 13 14 Sweden 12 17 24 30 22 Switzerland - - - - - Denmark 11 6 4 14 11 1989 for Canada 0 Government-guaranteed bonds only. 1987 for Denmark Table 133 Private bonds (as a percentage of assets) 1970 1975 1980 1985 1990* UK 14 6 2 2 3 Us 38 33 26 19 16 Germany 10 13 11 12 8 Japan - - - - _ Canada 1S 17 12 8 8 Netherlands 3 4 3 3 4 8weden 64 59 50 47 63 Switzerland - - - - - Denmark 61 , 66 59 52 56 * 1989 for Canada 1987 for Denmark 28a ¢ abl 14: BqpaitL*s (as a percentage of assets) 1970 1975 1960 1985 1990* UK 49 30 52 62 63 us 45 42 41 43 46 Germany 4 5 9 12 is Japan 6 10 9 16 27 Canada 22 25 21 28 29 Netherlands 11 11 5 11 20 Sweden 0 0 0 0 1 Iwltzorland 3 5 9 12 16 Denmark 0 0 3 6 7 * 1989 for Canada 1987 for Denmark !able 15: Mortgages (as a perceutage of assets) 1970 1975 1980 1985 1990* UK - - - - US 6 3 2 2 2 oermany 19 22 15 12 9 Japan 0 4 11 2 1 Canada 11 12 11 6 4 Notherlands 8 6 6 4 4 Sweden - - - - Swltzerland 15 13 10 9 8 Denmark 6 2 3 8 6 * 1989 for Canada 1987 for Denmark 28b 29 in countries such as Sweden and Denmark is is maturity of the fund, as the need to pa exceptionally low, despite the Danish tax on real pensions puts a greater focus on incom returns (Section 3(a)), which encourages generation, i.e., bonds, as opposed to capit substitution of equities for bonds. growth, i.e., equities. This may be an importan factor in the future in the UK and US. Fu.nding regulations can influence the equity s!iare, for example in the. United States where a Pension funds in all countries show a declining drop in market values can cause underfunding share of mortgages in recent years (Table 15); in which has to be reflected in the employer's profit Canada and the Netherlands weakness in the and loss account. As dis, issed below, this housing market has stimulated this trend. encourages holding of bonds and/or forms of However, note that Swedish and Danish funds hedging. have considerable exposure to housing markets via mortgage related bonds, and loans to housing Accounting conventions can also have an effect on credit institutions. Together with mortgages, equity holdings. In Japan, equities are held at these amounted to no less than 57% of Swedish book value, and a fixed return on the fund (based funds' assets in 1990, while Danish funds in 1987 on interest and capital gains) is targeted for every had 63% of assets in mortgages or mortgage year This gives adverse incentives to sell well association bonds. These imply an enormous performing equities as general share prices fall exposure to potential effects of recession and and retain those showing price declines (Tamura falling house prices. They may also imply a (1992)). In Germany and Switzerland, Hepp draining of resources from private industry (as (1992) suggests that application of strict contributors) as well as a diversion of personal accounting principles, which are more appropriate sector saving, depending on the post tax interest to banks than pension funds, restrains equity rates payable by mortgage borrowers. holdings by funded schemes independently of the portfolio regulations (evidenced, particularly in Lans face greater liquidity risk than bonds, while Switzerland, by the fact that funds' equity having the advantage of being tailored preJisely to holdings are far below the ceilings permitted). the needs of borrower and investor (longer These conventions, for example, insist on positive maturities, etc.). They constitute a large net worth of the fund at all times, carry equities proportion of Dutch and German pension funds' on the book at the lower of book value and assets (Table 16), reflecting the structure of market value and calculate returns net of financial markets as well as returns. Loans by unrealized capital gains. However, Lusser (1989) German funds are largely to banks and companies suggests that Swiss funds are also inhibited from (including the sponsoring company); Dutch fuznds equity investment by lack of expertise, lack of lend predominantly to the public sector. Swedish market transparency and limits on transferability and Swiss funds, that used to relv heavily on of shares. loans, now only do so to a limited extent. In Sweden the decline (both in "retroverse" loans to In contrast, the UK accounting standard permits participating companies and promissory note long-run smoothing and focusses on dividends loans) is related to the increased efficiency of the rather than market values, and hence enables domestic capital market in intermediating funds. funds to accept the volatility of equity returns. In Japan, the share of loans has again fallen The concern of some commentators in the UK is sharply, although these medium-term floating-rate rather whether equity holdings are too high given yen loans to firms were consistently the most the risks; however, note that 18% of the 63% profitable investment in Japan in the 1970s. It equity share in 1990w was actually in foreign can be argued that this highlights a general point, equities, thus reducing risk somewhat. No other that protection of fund managers from external country has anything comparable to this portfolio competition (as was the case in Japan till recently) share of equities. may lead to a sub-optimal investment strategy from the point of view of plan beneficiaries. (see A further factor that may influence equity holding also Section (d)). Table 16: Loans (as a percentag- of assets) 1970 1975 1980 1985 1990* UK 0 0 0 1 0 us - - - - Germany 31 33 37 36 36 Japan 52 30 22 15 13 Canada 0 0 1 0 0 Netherlands 46 52 63 52 39 Sweden 22 24 26 22 10 Switzerland 33 31 27 21 14 Denmark 1 1 4 1 1 * 1989 for Canada 1987 for Denmark Table 17: Property (as a percentage of assets) 1970 1975 1980 1985 1990* UK 10 15 18 10 9 Us - - - - - Germany 12 12 9 7 6 Japan 27 21 6 3 2 Canada 1 1 2 3 3 Netherland3 16 15 14 11 11 Sweden 0 0 0 0 1 Switzerland 16 20 18 18 17 Denmark - - _ _ * 1989 for Canada 1987 for Denmark 29a 30 The same comment appeared for a long time also. investing in LDCs, and the risk of heightened to apply to declining investment by Japanese volatility arising from short term herd-like shifts pension funds in propey (including equipment of funds between markets, see Section 5(d). For and real estate trusts) (Table 17), which has fallen further discussion, see Davis (1991). from almost 30% of the portfolio in 1970 to 2% now, although the current property crash now Of course, international investment poses casts doubt on this judgement. Property holdings additional risks compared with domestic in Germany, the Netherlands, and the United investment. Exchange rate risk means that the Kingdom, (where much of the accumulation returns from foreign assets may be more variable followed weakness of the equity markets in the than for domestic instruments, especially in the mid-1970s) have also declined recently. Once short term. Transfer risk may affect the ability to UK equity returns recovered and exchange repatriate returns. Settlement risk in some controls were abolished, property investment securities markets may be large, with a high declined owing to its lack of liquidity and lower proportion of delayed and failing transactions. returns than the alternative of foreign equities. Liquidity risk that transactions may move the As in Japan, in the light of the property crash in market against the fund may be significant In the UK in the late 80s, this strategy proved narrow markets. There may also be restrictions sensible. Dutch holdings were made less on investment given concerns over foreign control attractive by a tightening of rent and tenure and disruptive capital inflows." controls. Canadian holdings are small, and restricted to 7%. The principal exception to the Table 18 shows that foreign asset holdings have picture are the Swiss funds, which retain around a grown sharply over the 1980s in the United fifth of their assets in property. As noted by Kingdom and Japan. In both countries, this Schmahl (1992b), this focus may drive up the pattern followed abolition of exchange controls, at price of lauid and does. not contribute to capital a time when the econoii,'s were generating formation. Lusser (1989) also criticizes this current account surpluses and verseas investment approach, and suggests funds will face decreasing returns looked attractive. In Japan, restrictions returns on (domestic) property in the future, as on overseas investment were also progressively the population declines. eased over the 1980s. There is a contrast, however, in that UK foreign assets are virtually n principle, international diversification can offer all equities, whereas Japanese funds invest heavily a better risk/return tradeoff to fund managers, by in foreign bonds, see Table 19. Meanwhile reducing the systematic risk of investing in Dutch funds have long held a significant domestic markets arising from the cycle or long proportion of assets abroad, partly due to the term shifts in the profit share. It will be of large v)lume of pension fund assets compared particular importance in small markets with a low with domestic security and real estate markets. umber of liquid stocks - where domestic Growth was much less marked in the other nvestment would hence imply a high degree of countries (Table 18); in Germany and Canada ndustry risk. Alternatively it can be seen as a this is partly for regulatory reasons.52 Data for means of hedging against risks of imported Sweden and Denmark are not available, but their inflation (which vary with the openness of the foreign asset holdings are believed to be economy). It will also allow investment in extremely small. industries not present in the domestic economy. In a longer-term context, intemational investment The characteristics of pension funds' portfolios, in countries with a relat' 'ely young population which result from the asset selection discussed ay be essential to prevent battles over resources above, are shown in Table 21. The exceptionally between workers and pensioners in countries with high level of real assets for UK pension funds, an aging population. As a by-product, and the low levels in Scandinavia, are particularly international diversification should also improve notable. Reflecting their heavy investment in efficiency of global capital markets, subject to loans, German and Dutch funds have relatively ,itations funds impose on themselves on low levels of marketable securities. These Table 18: Foreign assets (as a percentage of assets) 1970 1975 1980 1985 1990* UK 2 5 9 15 18 Us 0 0 1 2 4 Germany 0 0 0 1 1 Japan 0 0 1 5 7 Canada - 3 4 5 6 Netherlands 7 8 4 9 15 Sweden - - - _ _ Switzerland - - - 3 5 Denmark - - - - - * 1989 for Canada 1987 for Denmark Table 19: Foreign assets of pension funds *nd-1988 Foreign assets Percent of Foreign bonds Foreign equities as (Sbn) total assets as percent of percent of foreign assets foreign ass-ts UK 53.8 13.9% 6% 94% Us 62.8 4.0% 14% 86% Germany 0.2 0.4% 93% 7% Japan 65.2 7.1% (50%) (50%), Canada 6.9 5.3% 7% 93% Netherlands 15.0 14.4% 41% 59% Switzerland 9.0 4% 70%+ 30%+ Memo: France 1.2 4.0% 15% 85% *. Percent of securities holdings only + Estimated Data for Sweden and Denmark not available. 30a Tabl- 20: Other assets (as a percentage of agsets) 1970 1975 1980 1985 1990* UK 0 0 2 4 6 Us 3 5 8 6 2 Germany 11 6 3 1 1 Japan - - - - Canada 8 1 2 2 2 Netherlands - - - - _ Sweden 3 0 0 0 0 Switzerland I 1 1 1 1 Denmark 23 25 28 24 25 * 1989 for Canada 1987 for Denmark 30b Table 21: Characteriutics of pension funds' portfolios United United Canada Japan Germany lItherlands Sweden Denmark Switzerland Kingdom States an a proportion of total ausetu"' Marketable 1970 0.85 0.90 0.77 0.21 0.23 0.28 0.76 0.72 0.31 securities b) 1980 0.79 0.86 0.73 0.64 0.34 0.15 0.74 0.66 0.41 1990 0.78 0.85 0.86 0.74 0.43 0.44 0.B8 0.74 0.55 Real assetsc 1970 0.61 0.45 0.23 0.37 0.17 0.28 0.00 0.00 0.19 1980 0.70 0.41 0.20 0.16 0.18 0.19 0.00 0.03 0.27 1990 0.72 0.46 0.32 0.29 0.24 0.31 0.02 0.07 0.33 Capital-uncertain 1970 0.93 0.90 0.76 0.51 0.36 0.42 0.76 0.72 0.44 assets(d 1980 0.94 0.82 0.70 0.7 0.42 0.29 0.74 0.66 0.55 1990 0.96 0.82 0.79 0.76 0.49 0.54 0.86 0.74 0.62 Long- term Fixed-interest- 1970 0.32 0.51 0.65 0.14 0.69 0.61 0.98 0.76 0.58 bearing assets"' 1980 0.24 0.43 0.64 0.54 0.76 0.72 1.00 0.70 0.55 1990 0.14 0.38 0.51 0.47 0.70 0.66 0.94 0.74 0.43 (a) Categories overlap, so they do not add up to unity. (b) Equities, bonds and market paper. (c) Equities and property. (d) Equities, property and bonds. (e) Bonds, mortgages (for Canada, the United States, the Netherlands, Denmark, Sweden and Germany), other loans (for Germany, Denmark, Sweden, Switzerland and the Netherlands). 30c 31 observations apart, for the United Kingdom, the measure for an ongoing portfolio, since they take United States and Canada, the table reveals a full account of losses or gains due to interest rate comparative lack of change in the characteristics changes (aithough other assumptions regarding of pension funds' assets, which may in turn be holding periods could also be made). The related to unchanging aims and regulation. The estimates suggest that pension funds in the United main shifts have been a move from fixed interest Kingdom obtained the highest real return over the to real assets by United Kingdom pension funds period 1967-90, Sweden, Switzerland, Canada, and into marketable and capital uncertain assets and the United States the lowest. These results by Canadian funds. This observation suggests must be reflected either in funding costs fo, that many of the portfolio shifts discussed above sponsoring firms or the level of benefits offered did not imply changes in objectives, but rather an and it is notable that UK funds tend to offer adjustment to market conditions within an superior benefits to North American funds (Tab' unchanged set of goals in terms of real return, 6). The result of course partly reflects risk an marketability, etc. Portfolios in Sweden and the share of equity and property, the Unit^ Denmark are also stable - but in this case perhaps Kingdom having the highest standard deviation - due to the tightness of portfolio regulation and the returns (together with Denmark), and by far the conservatism of fund management. Portfolios in highest share of real assets (Table 21). But as Germany, the Netherlands, and Japan have been noted in Section 1, compared with other financial somewhat more fluid; one cause of this, notably institutions, pension funds are well placed to in Japan, was the increased issue of government accept a degree of volatility. This is particularly bonds, with a concomitant shift out of property the case for defined benefit funds where there can and loans. be risk sharing between younger and older members"3 Meanwhile, Swedish, Swiss, United (c) Returns on the Portfolio States, and Canadian funds held high proportions of bonds, which performed poorly over this The patterns of portfolio distributions (Tables 8- period. 20) and risks and returns on assets (Table 7) can be used to derive estimates of the returns and Interestingly, portfolios in Germany and the risks on portfolios, and hence the cost to the firm Netherlands had a high real return and low of providing a given level of pensi6n benefits. volatility, despite their focus on bonds and loans. The method is simply to weight the annual real This relates to relatively high returns on fixed rate rate of return on each asset by the relevant instruments in those countries (Table 7). The portfolio share, thus giving on aggregation a high returns may appear to justify the series of annual portfolio rates of return. conservative asset distribution of German and Transactions and management costs are ignored; Dutch funds. Growing integration of capital actual returns would be lower if these were markets, however, should mean this asymmetric significant (see Section (d)). The average and performance is unlikely to be repeated, and hence standard deviation of these series are given in portfolio regulations locking German funds into Table 22. (The data in Table 22 cover the period this type of distribution remain difficult to justify. 1967-90 and average returns, especially for the Moreover, Table 25 (see below) shows that real US and Canada, are affected by the high level of returns for German and Dutch funds could have stockmarket prices in 1966. Appendix II shows been boosted significantly by an increased share real rates of return over the period 1970-90. It of equities. Investment in international equities can be seen that for the US average real returns would ensure that the associated increase in risk amounted to 4% over this period, twice the level was mitigated, reported for the longer period 1967-99.) Several observations can be made regarding these Annual holding period returns on marketable results. The publically sponsored Swedish fund fixed - rate instruments are used, as in Table 7, does poorly, despite the structure of independent instead of redemption yields. In our view, the fund boards. A further test of ownership effects. holding period returns are the more relevant splitting local government and private funds in the Table 22: Pension fund returns (1967-90) Mean (Standard deviation) of annual real total returns (domestic currrency) Percent us UK Germany Japan Canada Netherlands Sweden Demnark Switzerland Using holding period retunms on bonds (all countries) and fixed rate mongages (US & Canada) 2.2 (11.9) 5.8 (12.5) 5.1 (4.4) 4.0 (9.4) 1.6 (9.8) 4.0 (6.0) 0.2 (7.6) 3.6 (12.7) 1.5 (6.4) Average earmings 0.2 (2.1) 2.6 (2.5) 4.0 (3.1) 4.2 (4.2) 1.7 (2.8) 2.4 (3.2) 1.5 (3.5) 2.8 (3.6) 1.9 (2.! Pordolio return less average earnings 2.0 3.2 1.1 -0.2 -0.1 1.6 -1.3 0.8 -0.4 Govemunentbonds 0.6 (14.4) 0.8 (I1.)) 2.7 (14.9) 0.2 (12.8) 0.0 (12.1) 1.0 (13.1) -0.9 (8.5) 3.4 (16.1) -2.2 (17.6) Market paper 2.0 (2.5) 1.7 (4.9) 3.1 (2.1) -0.5 (4.6) 2.5 (3.3) 1.6 (4.0) 1.3 (3.5) 1.6 (1.8) 1.2 (2.2) Equities 4.7 (14.4) 8.1 (18.9) 9.5 (20.3) 10.9 (19.4) 4.5 (16.5) 7.9 (28.2) 8.4 (23.3) 7.0 (27.5) 6.2 (22.3) Memo: using redemption yields on fixed rate instrunents 3.9 (7.6) 6.3 (10.7) 5.5 (3.0) 2.9 (5.7) 4.1 (5.0) 4.3 (5.5) 2.8 (2.9) 5.8 (3.0) 2.2 (2.a) 31a 32 UK and US, is shown in Appendix. Public funds Canada, Japan and Switzerland and more again achieve lower returns. The Swedish and significantly so in Sweden. A negative Swiss are also both compulsory, thus in principle differential implies a need for higher contributions reducing competitive pressures. The Japanese, or for regular top ups in order to meet specified Swiss, and Germans have generally had little targets.5' It was noted above that this may relate competition in fund management (see Section 4(d) to inefficient asset alloc,ations. below). But as shown by the Germans, good economic performance - or international Table 23 shows the results of illustrati diversification - can overcome a number of calculations on the relation between costs handicaps. providing pensions, average earnings and real returns (based on Vittas (1992)), as applied to the Comparison of the risks and returns on pension results of the sample shown in Table 22. The fund portfolios with Table 6 shows the benefits of Table shows the implied replacement rate th diversification in terms of lower standard would be attainable given the real returns and deviations on the portfolio than individual assets. growth rates of wages shown in Table 22, However, the returns cannot be directly assuming indexed pensions, a 10% contribution compared, as pension fund returns are free of tax, rate, 40 years in service and 20 years retirement. while assets held directly would not be. It should The table illustrates clearly the benefits of also be noted that Table 22 only shows an higher return relative to real earnings; assumin estimate of returns to funds from "passive" pensions are indexed to prices, the UK funds can holdings of the relevant index for each asset, obtain a replacement ratio of 66%, the Swedes weighted by portfolio share. Appropriate stock only 16%. Clearly, given their lower returns, selection could in principle give a higher return - pension funds in Sweden as well as Switzerland, although as discussed below, in practice active Canada and Japan would require highe asset management often lowers retums, given contribution rates in order to achieve simila transactions costs. replacement rates to those of the UK funds (Table 2 in Appendix II shows a much more Comparison, in Table 22, of the results with risk favorable combination of real returns and earnings free yields suggests that the funds generally growth for the period 1971-90 in most countries. outperformed government bonds, albeit only This reflects the better performance of stock narrowly in Denmark. However, in Canada and market in the 1980s and the deceleration in Sweden the portfolio return is below that on earnings growth. It should, however, be stressed market paper (it is open to doubt whether the that what really matters is average returns and markets were big enough to absorb pension funds' earnings growth over 40 or even 60 years.) size, of course). Returns are generally below those on equities, but at a benefit of much lower Table 24 shows the real returns on pension fund risk. portfolios over five-year subperiods, thus offering an additional indccation of the risk of pension The most crucial test is ability of a fund to funds. As above, the patterns are influenced both outperform real average earnings, given that by portfolio distributions and the differing returns defined benefit schemes are basically indexed to on domestic financial markets. And as noted, the them. This is hence the key indicator of the costs degree to which the latter differentials may of the scheme to the sponsor. For the period continue with open and globalized financial 1967-90, the headroom is sizeable (over 2% pa) markets is open to doubt. Subject to these in only two countries, the US and the UK, caveats, the table shows that German funds have although the US result relies on the estimate of earned a positive real return throughout, whereas almost zero real wage growth (net of fringe in other countries returns were negative in 1971- benefits) over the past 25 years. It remains 5. In Denmark, Canada, and the US, returns positive in the Netherlands, Germany and were also negative before 1970, and in Sweden, Denmark. The differential between returns and Japan, Canada, and the US in 1976-80. In the earnings growth is negative, though barely so in 1980s a catchup occurred, although returns on Table 23, Implied replacement ratem with indexed pensions Percent Indexation of Indexation of penslions to prices p-nmions to wagem United States 37.6 36.9 United Kingdom 66.4 52.6 Germany 40.0 27.6 Japan 28.4 18.9 Canada 23.1 19.4 Netherlands 40.5 32.2 Sweden 16.1 13.7 Switzerland 21.6 17.8 Denmark 33.2 25.3 Note: This table uses the simple model developed by Vittas (1992) to calculate the implied replacement rates if the pension schemes were run as defined contribution plans and the real rates of return and real rates of earnings growth were those of the period 1967-90. The model is based on a 10I contribution rate, 40 years of active service and 20 years of retirement life. 32a Table 24a Real pension fund returns in sub periods (usiLj holding period returns on bonds) 1966-70 1971-75 1976-80 1981-5 1986-90 Urnol average UK 4.2 -2.8 4.9 12.4 10.1 5.8 (11.5) (19.4) (5.2) (7.3) (12.7) (12.5) US -S.4 -0.8 -1.9 8.1 11.2 2.2 (6.5) (13.8) (6.9) (13.0) (12.2) (11.9) Gormany 5.0 3.3 3.3 7.7 6.3 5.1 (3.3) (2.7) (4.4) (4.9) (5.9) (4.4) Japan 0.1 -0.5 -1.2 10.9 13.8 4.0 (5.3) (10.9) (5.3) (2.1) (7.8) (9.4) Canada -3.3 -1.2 -1.0 6.1 7.9 1.6 (1.4) (11.7) (4.0) (15.1) (6.7) (9.8) Netherlands 1.7. -1.4 2.02 10.5 6.3 4.0 (3.3) (5.5) (3.3) (4.0) (5.4) (6.0) Denmark -1.9 -1.3 0.8 17.7 -1.8 3.6 (8.7) (12.7) (4.4) (14.6) (10.3) (12.7) Sweden 1.2 -3.5 -5.3 3.9 4.7 0.2 (8.2) (6.7) (5.6) (4.9) (9,3) (7.6) Switzerland 0.8 -0.5 4.0 3.0 -0.2 1.5 (0.0) (6.3) (8.0) (5.4) (7.2) (6.4) 32b Table 25: Mean (standard deviation) of real total returns on diversified portfolios Domestic' Domestic & MeNo: Col 2 Per cent international2 less Average Earnings United States 1.4 (12.1) 1.9 (11.4) 1.7 United Kingdom 4.9 (16.4) 4.5 (14.9) 1.9 Germany 6.3 (11.2) 5.8 (10.5) 1.8 Japan 5.7 (14.2) 4.7 (12.5) 0.7 Canada 2.4 (11.9) 3.0 (11.1) 1.3 France 3.7 (16.3) 3.8 (14.6) n/a Netherlands 4.5 (17.4) 4.1 (15.5) 1.1 Sweden 3.8 (13.5) 3.7 (12.1) 2.2 Switzerland 2.0 (16.5) 2.0 (14.5) 0.1 Denmark 5.2 (18.9) 4.5 (16.4) 1.7 1 S0o domestic equity, 50% domestic bonds. 2 40* domestic equity, 40% domestic bonds, lo% foreign equity, 10% foreign bonds. Note: International diversification should cause a convergence in ret -ns by increasing below average returns and lowering those above average. This would occur if purchasing power parity held all the time. In fact, large deviations between exchange rates and purchasing power parities may prevent this convergence from materializing, at least in the short to medium run until corrections in exchange rates take place. 32c 33 Swedish and Swiss funds are consistently below year of the sample. In most cases this may not be the Germahs-, despite similar portfolios, largely so, as some form of active portfolio management because Ger.man bonds have returned a positive is widely adopted. This warrants a discussion of real yield. Of course, it can be argued that a five the economic issues in fund management, costs, year horizon is irrelevant to the time scale of and the implied effects on pension fund returns. pension funds (given a working career that lasts 3040 years). However, it may be of relevance Fund management is a service entailing when strict funding and accounting rules are management of an investment portfolio on behalf applied. of a client - in this case a pension fund. Management may be carried out internally, that Finally, Table 25 shows the returns on artificial is, by employees of the fund, or externally" by portfolios of holding 50% equity and 50% bonds, a financial institution such as a bank or insurance -and of international diveasification. This shows company. Such delegation raises principal-agent what could occur if portfdlio. restrictions did not problems, as unless the manager is perfectly exist. As noted,' equity hdlding$ are generally monitored and/or a foolproof contract drawn up, below this (Table 14). Compared with Table 22, he may act in his own interests (e.g., in the results confirm that returns may be boosted by generating excessive conmmission income) and raising the share of equity, at some cost in terms contrary to those of the fund. One can suggest a of risk. 'Meanwhile foreign investmnent always priori that such monitoring will be costlier when reduces risk, though in some cases there is a trade managers lack reputation or relationships, which off with returris -(this generally relates to the otherwise constitute assets that would be changes in the exthange rate over time. With a depreciated by adverse behavior. Also internal structurMl depreciation, for example, as in the managers should be less susceptible than external, UK, returns . on foreign assets- are boosted, given a greater degree of control that can be comnpare Tabie 7.). Several of the countries which imposed via employment contracts, etc. fell below a satisfactory return on assets relative to average earnings (such as Canada, Denmark, Various features of fund management in countries and Sweden) would have found pension provision such as the UK can be seen as ways to reduce less costly if they had followed such a rule. principal-agent problems. For example, managers German funds would also have boosted their are offered short (3-year) mandates, with frequent headroom considerably. performance evaluation; fees related to the value of funds at year-end and/or performance related 'Sunrnarizing Sections (b) and (c), support is fees. given to a prudent man rule, backed by flexible accounting and funding standards (perhaps The level of management fees (excluding focussing on income rather than market value) to transactions costs) chargeo by fund managers back holding of high-return but volatile assets. depends on the competitive structure of the Except for a fund with mostly retired members, a market; for example in the competitive UK shift into deficit in one year should not interrupt market a fund would pay no more than 22 basis payment of pensions. These policies in turn points on £100 million. In the US, fees are should help minimize the cost to the sponsor of higher at around 40 basis points. This difference providing a given level of benefits. Since foreign may relate to greater ability to cross subsidize investment is shown invariably to reduce risk, from retail business in the UK and /or higher risk albeit often with a slight reduction in return, of loss of mandate in the US, which necessitates limits on such holding are suggested to be higher fees to break even. Fees in countries such particularly counter productive. as Switzerland and Germany, with relatively uncompetitive fund management sectors, are far (d) Fund Management higher - 100 basis points or more. In Japan, several structural features ensure low levels of The returns calculated above all assume that funds competition in fund management. Until recently, hold the market index of the relevant asset in each only trust banks and life insurers could manage 34 funds. In house rnanagement is restricted to asset categories. This is in-ine with the efficient bonds. New entrants, only recently permnitted, markets hypothesis, which suggests that, given can only invest new inflows. Accordingly, trust prices already incorporate all available banks charge 60-180 basis points, while life information, there is no net benefit from spending insurers charge 2-5% of the inflow.56 extra cash to try to beat the index. Nevertheless, as noted by Grossman and Stiglitz (1980) and However, even more than administrative costs, Cornell and Roll (1981), the efficient markets the crucial influence on pension funds' costs is of hypothesis does not rule out small abnormal course the efficacy of asset management. Here returns as an incentive to acquire informaticn, but again, the countries with uncompetitive fund those acquiring costly information should have management sectors may lose out; for example in only average net returns after the costs of Japan the asset return target is fixed (8%) and acquiring information are taken into account. In there is little incentive to exceed it. Similarly in practice, as shown below, active managers Switzerland there are few rewards for exceeding a underperforn. low return (and considerable costs, given the accounting system, in holding volatile assets that Data for the UJK are shown in Table 26. These could put the fund below actuarial balance). show that even in the home market, funds tend to underperform the index, but underperformance is Where fund management is a competitive sector, particularly severe in foreign markets. This in as in !he Anglo-American countries, portfolio turn justifies an indexed approach to national managetnent is typically a two stage process, with stocks, where the fund manager's skill is traditionally a strategic decision regarding employed in picking undervalued markets rather allocation to different assets and national markets than stocks, and employing stock index futures to being followed by a lower level decision over the gain rapid exposure to such markets. As precise assets to be held within these categories. discussed below and in Howell and Cozzini The latter may include passive indexation of the (1991), this is increasingly the approach adopted market. (However, more recently, there is by large international investors in countries such evidence that fund mangers are picking core as the UK. Table 27 shows that average returns holdings of stocks at a strategic level, and picking are lowest for external managers, which is in turn national markets at a tactical level, with the use of inversely correlated with turnover. These are tock index futures (see below, also Davis (1991) consistent with the principal-agent problems set nd Howell and Cozzini (1991).) out above; managers least under control have higher turnover and lower returns. The traditional strategic choice, typically taken jointly by managers and trustees, is illustrated by Similarly, Lakonishok et al (1992) show that most the data in Sections (b) and (c). The results US investment management is again active, and suggest that there is a tradeoff, as would be that fund managers consistently und ¢rperform the expected, between return and risk and market, for example the equity proportion of US considerable benefits from diversification. This funds (excluding the management fee) in turn points to the need for appropriate underperforms the S&P 500 index by an average measures of risk-adjusted returns and of 1.3% pa over 1983-9, or 2.6% if returns are identification of sources of portfolio performance value weighted. If managers overperform in in order to evaluate fund managers' performance some periods this is virtually never sustained. (Blake (1990), Tamura (1992)). The authors suggest that the persistant use of active management despite such evidence is As regards the traditional short term asset related to agency problems. In particular, they allocation decision, which tests ability of active suggest that these may arise within the management ("stock picking") to outperform management structure of the sponsor; corporate indices inclusive of fees and transactions costs, treasurers seek to bolster their own positions visa the evidence is almost uniformly contrary to the vis their managers, and hence seek fund managers efficacy of active management of funds within that can offer good excuses for poor performance, Table 26: UK Pension fundst Long term returns on equity relative to benchark indices 1981/89 1981 1982 1983 1984 1985 1986 1987 1988 198sverage USA -3.5 -4.2 -4.1 -8.3 -2.7 -3.0 -3.7 -3.4 -0.5 -3.7 Japan 8.5 6.9 9.3 -15.4 -8.7 -1.1 -13.6 -8.9 5.3 -2.0 Cont Europe -6.8 5.7 0.8 -5.3 -4.6 -4.0 -3.0 -0.4 1.7 -1.8 World -3.8 -3.6 1.9 -8.5 -1.9 -2.6 -10.1 -5.8 6.7 -3.1 UK 0.5 1.2 -0.5 -1.6 -0.5 -1.2 -0.8 -1.0 -0.1 -0.4 (1) Prior to 1987, local indices for US and Japan, MSCI for Europe. After 1987, FT indices. Source: WM Table 27: UK Pension funds: Performance and Turnover by Management Method, 1986-90 Nominal Annual Activity Cl) returns (') Internal 11.1 64% Part internal/external 10.9 118% 2 or more manager. 10.6 119% Financial conglomerates 10.8 106% Life company managed 11.2 96% Life company segregated 10.4 117% Independent managers 10.6 118% (1) Activity is the element of turnover in excess of net investment of new money, as percent of assets. 34a 35 clw stories about portfolio strategies and other generally. As discussed below, widespread servies unrelated to performance. They avoid indexation may also weaken incentives to monitor Ixiexat{on, as this would reduce their own day to corporate management. Meanwhile, to the extent day responsibilities, as well as internal asset indexation is not used, a reliable measure of fund management, as this would give them too great a managers' performance as a means of control is reonsibility for errors. The authors suggest seen as essential. Performance measurement these aency costs are additional to the difficulties services - whose data generated the results (u noted above) which arise between a (rational reported above - are well developed in the Anglo- profit-m&ximzig) sponsor and the fund manager, American countries but vestigial in Continental W that a sift to defined contribution plans Europe and Japan. Indexation does not, of would help overcome the difficulties. course, remove the need for selection of aset categories in the light of liabilities and expected The main implication of these analyses is that returns, as well as choice of national markets. To IdenIo, as dicussed below, will be optimal the extent these activities remain profitable, they for mot pension fund assets, with the caveat that suggest that there is greater efficiency within thn te beefits arising from active management are between markets. likely to incrase with the number of funds adopting such predictable passive strategies, and witb the inefficiency of the market more 36 5. EFFECTS ON CAPITAL MARKETS This section discusses the impact of pension funds Bodie (1989) suggests that fixed duration on innovation, market structure, demand for securities (and associated strategies) have little capital market instruments, volatility and the role in terms of household utility maximization, overall development of capital markets. as they are unable to hedge against the inflation risk to future consumption. Hence an individual - (a) Innovation or equivalently a defined contribution pension plan given the distribution of risk to the employee The impact of pension funds on the development - would not seek such instruments but instead of capital markets varies from country to country. would just diversify, seeking to maximize return For example, as regards innoyation, in the United for a given risk. The only difference would be States ERISA codified the legal status of defined that in a tax free pension plan, there is an benefit corporate pension funds and imposed incentive to focus on the least tax-advantaged minimum funding requirements, sharply securities such as corporate bonds (subject to increasing demand for hedging by pension inflation risk). Consistent with this analysis, funds." This has stimulated the development of Berkowitz, Logue (1986) found that returns on immunization strategies (to match assets to defined benefit plan were below other US liabilities) based on long-term bonds. The diversified portfolios over 1968-83, where the incentive to immunize in defined benefit schemes shortfall in returns was identified as the arises from the asymmetry of treatment of pension 'insurance premium". (Although as noted by deficits and surpluses (Section 3 (d)), which Lakonishok et al (1992) in Section 4(d), the implies that the corporate guarantee of the shortfall could also be due to agency problems.) accumulated benefit obligation (ABO) is a put option on the investments of the pension fund Meanwhile United States funds have been in the with an exercise price equal to the present value vanguard of developing passive indexation of the ABO. To minimize the cost of the option, strategies (which appear justified in the light of there is an incentive to immunize the liability via persistent underperformance by active fund an investment strategy of duration matching. managers, as discussed above). In 1990 41 % of US funds employed such strategies, with index The requirement of a fixed duration for funds accounting for a quarter of total US funds' investment instruments has stimulated innovations assets. tailored to funds' needs such as zero coupon bonds, collateralized mortgage obligations and In the United Kingdom, the contribution of guaranteed income contracts (offered by life pension funds to innovation is less clear cut. insurers); immunization strategies also spurred Many trust deeds used to prevent funds from development of markets for index options and using derivatives, though these restrictions have futures. For example, pension funds wri:ing call been relaxed more recently. Taxation was also a options on equities can be seen as converting discentive until the late 1980s (use of derivatives them into short-term fixed-income securities for was counted as "trading" and taxed). There also matching purposes. Another strategy is holding appears to be a more general difference in assets in excess of the legal minimum in equities, attitudes between United Kingdom and United as long as their proportion is reduced when the States managers to innovation. (See Davis market value of pension assets approaches the (1988).) This may be related mainly to the less ABO. This strategy is known as portfolio asymmetric and more flexible accounting insurance or contingent immunization, and has treatment of funds in the UK, where there is no stimulated development of index options and sudden cutoff point where liabilities must enter futures markets (as well as being blamed for the balance sheet. Also minimum funding market volatility, for example at the time of the standards only apply to a subset of pensions (the 1987 crash). GMP). Thus the option/guarantee effect described above for the US does not apply 37 particularly strongly, and funds have so far been position-taking market makers ready and able to happy to hold an overwhelming proportion of facilitate large trades, will hence tend to be unhedged equities. However, Blake (1992) attractive to pension funds. Because liquidity is a suggests that as funds mature and raise their form of economy of scale, it tends to make it holdings of bonds in order to reduce the risk of difficult for other markets or financial centres to not meeting liabilities when they fall due, compete with such markets, even with similar mmunization strategies will come to the fore. technology. London's SEAQ International is a classic example; it currently carries out 50% of However, one area in which UK funds have French and Italian equity trading and 30% of already been particularly active is use of German, for example. Its relative liquidity is derivatives in international investment. As reflected in transaction sizes - $275,000 compared iscussed in Davis (1991), stock index futures are with $25,000 in Paris and $50,000 in Frankfurt. seen as particularly useful in tactical asset Similarly, growth of pension funds in the US has allocation, facilitating rapid shifts between led to development of off-exchange 'block different national markets, which would later be trading". The growth of such exchanges may translated into stocks. Derivatives might also be entail a tiering of markets, with order-driven and used for long term strategic movements into heavily regulated domestic markets retained for markets or stocks, if they enable such shifts could retail investors and for small company stocks. occur without moving the market against the Liquidity may be aided by reduction in fund. This will be the case if the derivatives conunissions, that institutions are well-placed to arkets are more liquid than the underlying (as, press for. Increases in liquidity should in turn be for example, in Japan, where in mid-1991, beneficial more generally to the efficiency of outstanding futures contracts represented three capital markets, and lead to a reduction in the times the daily number of shares traded on the cost of capital. stock- market). Also temporary adjustments in posure could be obtained by purchase and sale (c) Demand for Capital Market Instruments index futures without any transaction in the derlying (overlay strategies), thus avoiding Institutional investors can influence the demand disturbance of long-term portfolios, see Cheetham for capital market instruments in several ways; 1990). Such strategies also facilitate by influencing the rest of the personal sector's ;unbundling" of fund management into currency, portfolio distribution between bank deposits and market and industry exposure. Finally, pension securities, by the institutions' own portfolio funds might invest cashflow awaiting long term choices, and by increasing the total supply of nvestment, in derivatives, as it ensures the saving. manager is always invested and will not miss an urn. In the Anglo-American countries, econometric results (Davis (1988)) suggest that the growth of Market Structure institutions has been accompanied by a shift by persons from securities to deposits, not matched -roader effect of the development of pension in Germany and Japan.' Hence securities are unds and other institutional investors is on increasingly held in the Anglo-American countries -1 market structure. Their key demand is by large, informed, risk-averse investors facing dity, that is, ability to transact in large size low transactions costs. Such a capital market out moving the price against them, and at low should sensitively reflect information on firms' ransactions costs. They are unconcerned by the performance. irmness of investor protection regulation, as they iave sufficient countervailing power to protect This is confirmed by econometric analysis (Davis their own interests against market makers and (op cit)) of the portfolio distributions of pension other financial institutions. Specialized wholesale funds, which show they are strongly influenced markets which focus transactions and increase by relative asset returns, particularly where there dity, usually centered on well-capitalised are few regulations governing portfolio 38 distributions and low transactions costs, as in the the UK in the 1980s. United Kingdom and the United States. Adjustment to a change in such returns is (d) Volatility and Short Termism generally rapid. This implies an efficient allocation of funds. These results do not all hold A further qualitative question is whether where transactions costs are high and regulations institutionalization increases capital mut are strict - e.g., in Germany, Japan, and Canada. volatiljty. Some commentators in the Unit In these countries adjustment to a change in States blamed fund managers' portfolio insuran returns is somewhat slower and allocation of strategies for causing volatility at the time of the funds less efficient. The results also contrast with 1987 Crash, although this is disputed. Regul those for households and companies (Davis performance checks against the market (as (1986)) where adjustment to changes in returns frequently as monthly in the United States, b tends to be slow, due to higher transactions costs less in the United Kingdom) may induce and poorer information. "herding" among funds to avoid performing significantly worse than the median fund. The Most of the literature suggests that Japanese also appear to suffer from such herding, institutionalization has a signiflcant but not major despite a less competitive environment for effect on total personal saving, increased saving managers. via institutions being partly offset by declining discretionary saving, (see Feldstein (1978), Interviews with fund managers suggests this ma Munnell (1986) and the review in Smith (1990).) be an important cause of volatility not only i although some recent studies, such as Hubbard domestic but also in international markets. S- (1986) suggest a much larger effect. In theory, Davis (1988) (1991). As noted in Howell an while the scale of benefits of a pension system Cozzini (1991), the rise of global asset allocatio may have an effect on personal saving, funding as as a tool of fund management, and the such should not. (As noted, funding is rather a development of markets such as those for stock transfer of securities from the sponsoring firm to index futures have stimulated and facilitated the market, which collateralises the liabilities, massive increases in short term cross borde reduces risk of non-payment (because of equity flows. One equity transaction in three in diversification) and gives scope for voluntary Europe now involves a foreign transactor; and increase in pensions when returns are high.) The trading in stock index futures often far exceeds effect that does occur may result from liquidity that in the underlying. Although the investors constraints on some individuals (especially the desire to reduce risk by adopting such strategies, young), who are unable to borrow in order to the focus of funds on a small number of leveraged offset obligatory saving via pension funds early in instruments often leads to destabilization of the life cycle. Pension saving may be partly markets and sharp swings in asset prices. Nor offset at a national level by tax subsidies to need the behavior be confined to equity markets. private saving, especially if they are financed by Besides the fact that equity flows themselves have public dissaving. However, a switch from social a direct effect on the exchange rate, evidence in security to funding would probably have a major mid-1992 suggests that fund managers switched to effect on saving, given the former has been cash in the light of relative returns, and were at shown significantly to depress saving in a number least partly responsible for the prevailing of countries.' exchange rate tensions (a forex manager in a bank has around $20 million at his disposal; a fund Bernheim and Shoven (1988) show that pension manager can have billions). Indeed Howell and funds may change the volatility and relation Cozzini (op cit) suggest that international between saving and real interest rates. Data from regulatory bodies need to tighten supervision of the US show that a rise in real rates may reduce international securities flows, to prevent saving if it makes more schemes fully funded deleterious effects on real economies. (target saving) and reduces the need for contributions. There is also evidence for this in Regular performance evaluation is also said to 39 underpin the short-termist hypothesis, that others (Germany, Italy) do not. There is a willingness of funds to sell shares in takeover question of which Comes firt? battles (to maintain performance) discourages long erm investment or r&d. Conclusive evidence is Although pension funds could develop on the scant, but there is widespread agreement that basis of loans or property investment, their other ways besides takeovers of exerting greatest comparative advantage is in the capital corporate control should be more widely used by market. Loans require monitoring so the institutions, such as appointment of non-executive customer relationship may give banks a directors to represent shareholders' interests, or comparative advantage here. Trading and risk irect involvement of pension funds in pooling are more efficiently undertaken in the management.10 The development of portfolio capital markets where transactions costs are indexation has important implications in this lower, although these need not be domestic context, since the longer term relationships, close markets if there are no exchange controls and monitoring of company performance and large funds can invest in developed capital markets shareholdings needed for these alternatives to elsewhere. Moreover, if one of the spurs to takeover to operate will not be present. development of protection in retirement is income equalization6" (as well as rising averagc related point is that pension funds and other incomes), this may with a well-developed capital stitutional investors may not invest in small market simultanecusly provide the means for rms, given illiquidity of their shares, difficulty development of funded schemes (reduction of and costs of researching firms without track personal equity holdings by the wealthy) which is records and limits on the proportion of a firm's absent in a system dominated by banks. States equity that may be held. They may also lack the might be more likely to opt for a generous social business expertise to supply risk taking venture security scheme in the latter case. capital, given the need for close monitoring of the lients of such finance. It can be argued that Unlike pay-as-you-go social security schemes, e problems would be overcome if, given a where there can be an immediate transfer of greater degree of tax neutrality between types of income to those who have not contributed (who saving, more funds were directed through banks are old at the outset), in finded private schemes rather than institutional investors. On the other the assets are built up while they are maturing, hand, it may be best to avoid the associated and this stimulates investment and *he tendency of banks to monopolize the financial development of securities markets. This effect is markets. of course offset if others reduce securities holders or saving differentially in the case of funded and ese potential drawbacks, it should be noted, social security pensions. The discussion above is ult from the nature of capital markets and from also relevant here, for example in that it suggests institutionalization generally, rather than from funds may increase market efficiency. ,ension funds per se (th6ough they are clearly a component of institutionalization). Any Given their focus on real returns, pension funds iedies should therefore be a feature of general should be particularly beneficial to development nomic and financial policy and not pension of equity markets. Development of equity ,olicy - policies bearing solely on pension funds markets in turn is seen as beneficial in providing taxing short term capital gains) -would risk capital for growing enterprises,' as well as dvantage them without solving the problem offsetting the potential fragility and/or dependence life insurers, mutual funds etc would be on bank finance which, stems from high unaffected). debt/equity ratios (see Davis (1992)). Development of Securities Markets The analysis of this paper suggests that focus on equities by funds can be stimulated by flexible ountries with large pension fund sectors tend to accounting rules, encouragement of indexed iave well-developed securities markets, while benefits and institution of a prudent man rule 40 instead of portfblio regulations. Absence of the causality may be reversed; a dominant insurance of benefits .n help to avoid the need pension fund sector will ensure satisfactory for strict rules on portfolios and on provisioning. treatment for shareholders is maintained. And Some would suggest that there is also a need for indeed, even Swiss funds have become more guarantees of shareholder rights (e.g., equal active shareholders in recent years. However, treatment in takeovers, rights of pre emption over experience in countries such as the Netherlands new share Issues, equal voting rights) In order for (Van Loo (1988)) suggests there is no one to one pension funds to hold equities willingly. Lusser relation between pension funds and equities, even (1989) suggests that such problems help to inhibit with appropriate regulation as outlined, if funds Swiss funds from equity investment. In practice adopt a very risk averse strategy. 41 6. CONCLUSIONS In this final section we offer, first, a broad attractiveness of pension fund saving relative to summary of the content of the paper; second, we other forms of private provision for retirement. focus on prospects for pension fund growth in Taxation is an obvious example of regulation that industrial countries and, third, we assess issues of can stimulate pension fund growth, but it is also potential relevance for countries setting up shown how provisions such as portfolio pension funds de novo. The author's views on regulations, funding rules, ownership of the various policy choices are noted in this last surpluses, portability, insurance and protection section; the reader should, however, bear in against fraud can influence the attractiveness of mind that in most cases such judgements are a pension funds either to the sponsoring firm or the question of choosing an appropriate point on a members. There is little international consensus tradeoff between alternative benefits or costs. on the appropriate scope or even role of many of these regulations. There are tendencies for (a) Summary countries to group together on regulatory issues, as they do on the structural features noted above. Section I discussed the economics of pension For example, the Dutch/Anglo-American group funds, distinguishing defined benefit and defined tends to have "prudent man" asset management contribution plans and outlining the features of regulations and short vesting. In contrast the pension funds largely by contrast to other types of Germans, Swiss, and Japanese have portfolio financial institutions. It was suggested that regulations and longer vesting periods. (See regarding pension funds as offering employee Table 6.) Features such as the structure and retirement insurance was a fruitful way to assess mechanics of supervision are also shown to vary them economically, although other approaches widely between countries. Some a priori such as the tax shelter, labor economics and suggestions regarding best practice were made. corporate finance approaches could also offer insights. Section 4 reviewed the performance of funds. The relative level of benefits offered, and their In Section 2 the features of pension funds in the indexation against inflation, vary widely and not main industrial countries were outlined in the only in response to the relative generosity of light of Section 1, and their place in the patterii of social security. The resultant levels of retirement provision clarified. It is noted that the contributions also vary. The bulk of the section argumeats for funding are not unidirectional, and is devoted to analysis of portfolio distributions, as case can in some circumstances be made for comparative data for 10 asset types is considered. pay-as-you-go. International comparison (as The influences on distributions, as well as risk detailed in Tables I and 6) shows four broad and return, are shown to include the nature of groups of countries, one with long-established liabilities, regulation, accounting standards and funded defined benefit schemes (Netherlands, UK, the supply of certain financial instruments. US, Canada), one with nationally-directed or Estimates are made of returns on the portfolio; it provided compulsory funded schemes (Sweden, is shown that a focus on real assets such as Switzerland), a third with relatively small funded equities end property has generally boosted sectors, but significant levels of unfunded returns, but that variations between countries in corporate pension liabilities (Germany, Japan), real returns on debt instruments have sometimes and an exception, Denmark, with significant more than offset this. Hence the German funds nded defined contribution schemes. The key come second only to the UK, because real returns determinant of the growth of private retiremeni on bonds and loans were never negative in the saving via institutions such as pension funds is the 70s. Such an analysis ignores the role of scope of social security. transactions costs and the portfolio management process, which it is suggested may pose problems Section 3 outl.ned the regulation of pension funds, to pension funds due to high annual charges and suggests it is a crucial determinant of the and/or poor investment performance for active 42 managers. stimulus to capital markets - would probably b significantly greater. It is notable that in th The influence of pension funds on the capital Anglo-American countries, where social security markets was assessed in Section 5. To a degree is less comprehensive, the ratio of personal that varies between countries, they are shown to financial wealth to GDP is around 2, whereas in have stimulated innovation, promoted liquid France and Germany it is below 1.5. If French market structures, boosted the demand for capital financial wealth reached the same level as the UK market instruments (by increasing saving) as well in relation to GDP, as well as pension funds as making demand more sensitive to return and attaining the same share of personal wealth, the risk, and aided the broader development of capital stock of pension assets would be over $40 markets. However, they have also prompted billion. To a degree depending on portfolio some concern over their contribution to capital regulations and the investment climate, this should market volatility at both a domestic and in turn boost demand for equities (as discussed international level, and to short-termism of above). nonfinancial companies. Current proposals for EC reform are also of (b) Prospects for Advanced Countries relevance (see Commission of the European Communities (1991), Kollias (1992)). The EC Growth prospects for pension funds differ sharply proposes legislation to liberalize funded retirement between the countries studied. In Sweden and provision, although the process is still at a Switzerland, there is little prospect for growth, consultative stage. A draft Directive has been with 90% coverage and schemer largely mature. drawn up on funded pension schemes which In the Anglo-American countries and the addresses the following issues: first, the freedom Netherlands' most company funds are mature to offer services across borders (in other words, and therefore any significant growth is likely to administration and fund management can be stem from broadening of the coverage of private conducted in another member state); and second, pensions across the labor force. The success of the liberalization of investment throughout the personal pensions in countries such as the United Community (altnough this freedoin should already Kingdom indicate considerable scope for this. In exist, especially perhaps for personal retirement Denmark, Japan, and Germany, immaturity of provisions, under the Capital Liberalization company schemes indicates further growth is Directive). Meanwhile, discussions continue on a likely. third proposal contained in a recent consultative paper, namely the freedom for pension schemes to But more generally, in many countries (notably in operate across national boundaries on the basis of continental Europe) future demographic pressures home state authorization and for individuals to on pay-as-you go social security are likely to lead join schemes in other member states. This is seen governments to seek to stimulate growth of as the .nost difficult issue, particularly due to private pensions as a substitute for social security. need for countries to agree on funding standards, If such countries were to develop schemes as well as fiscal differences; but it is also the equivalent to those in the United Kingdom, the most important for labor mobility and the sums involved would be sizable. For example, if completion of the single market. A first step may French pension funds were to reach the size of be to cover only migrant and "frontier" workers, their UK counterparts in terms of shares of that is, those living in one state and working in personal sector assets, they would total $235 another. Agreement on these three issues could billion. Similar calculations for Germany give clearly facilitate development of pension funds in $400 billion in assets, which compares with the continental European countries currently $355 billion market capitalization of the German dependent on pay-as-you-go schemes. stock market. In practice, personal sector financial wealth would probably be boosted by a A subsidiary objective in a number of European switch from pay-as-you-go to funding, so the countries, which growth of pension funds may increase in value of funds - and consequent assist, is development of equity markets. 43 Following the calculations above, if funded structure - subject to certain side effects sectors developed in France and Germany on a (volatility, short-termism) (Section 5 (a), (b) and par with those in the UK, and equity proportions (d)). These effects may be stimulated by certain were similar to US funds, the increase in demand regulations such as those for funding (Section 3 for equities would be $106 billion and $184 (d)) but also blunted by factors such as portfolio billion, respectively. (Note, however, that in regulations and the structure and behavior of the global terms these might be partly offset by the fund management sector (Sections 3 (c) and 4 maturity of UK and US funds, which may induce (d)). These are strong arguments in favor of a relative switch into bonds from equities by such developing private funded schemes. The funds.) benefits to the capital market would be absent in the case of book reserves, and hence external (c) Issues and Recommendations funding is seen as more desirable. The side effects, ir considered suMciently undesirable, As outlined by Vittas and Skully (1991), there are should be dealt with In the context of economic a number of questions to be faced by countries policy more generally. They are general seeking to set up or develop pension funds. features of equity markets and/or institutional These include the role of contractual savings investors rather than pension funds per se. institutions in retirement income provision; their impact on saving and capital markets; their The implications of development of pension funds effects on economic efficiency and social equity; for economic efficiency include their effect on the role of government in promoting them; the labor niobility (Section 3 (f)) and distortionary case for preferential fiscal treatment; the need for effects of their taxation (Section 3 (a)), as well as compulsion; and the appropriate regulatory the above-mentioned consequences for the capital framework. This paper has sought to address market. Social equity is affected by the rules on these questions by reference to the adopted internal transfers and equity of treatment (Section solutions in the major industrial countries. We 3 (g)), coverage (Section 2 (c)), rules in relation conclude by highlighting some of the main issues to tax privileges (Section 3 (a)) and the safety net brought out under each of these headings, and of social security 'Section 2 (b)), as well as the seek to come to a judgement on some of the degree of choice and disclosure (Section 30)) and key questions. the scale and indexation of benefits offered (Section 4 (a)). Appropriate regulatory design, The primary role of pension funds is a as outlined in Section 3(l) is needed to supplementary one in each of the countries minimize these difriculties. We consider that studied; there are no cases where they provide indexation, at least up to a certain level the only form of old age support, although the (subject to a "prudent man" asset management height of the social security safety net varies rule being in operation) and rules facilitating a widely, and this in turn has a crucial effect on the degree of portability are particularly development of funded schemes (Section 2). We desirable. Arguments against perfect onsider such a mixture sensible, given the portability (e.g., reduced Incentives of connlicting arguments for funding as opposed employers to train workers) should not be to pay-as-you-go disregarded, however. The impact on saving and long term financial The role of government in promoting pension resources is to boost the former, albeit not in a funds has been shown to be a crucial one. In one-to-one manner, while pension funds also particular, the level of state benefits' and the stimulate the development of securities and equity ability of employees to opt out of the state scheme markets under certain conditions (Sections 4 (b), and personal pensions (Section 2); changes in (c) and 5 (e)). As regards the development and taxation of pensions and alternative assets (Section functioning of capital markets, pension funds may 3(a)), legislation on the nature of benefits be beneficial by promoting inn' vation, liquidity (Section 4 (a)) and legislation on provisioning and efficiency, while also influencing the market (Section 3 (d)) all have a crucial role to play in 44 making the setting up of funds attractive to firms that individuals are myopic is taken seriously; and (assuming their establishment remains voluntary). the evidence seems quite strong. It could be More indirectly, the provision of a stable macro argued that if fund are compulsory, then relative environment and steady economic growth, via its tax advantages are not needed, and all forms of influence on the returns on capital market saving should ideally receive expenditure tax instruments (Section 4 (c)) will also influence the treatment. Compulsion will also have an effe cost of providing funded pensions. The on the corporate sector, since it will impose an government also faces certain key choices in unavoidable burden on companies, which in turn influencing the development of pension funds, of could affect international competitiveness of the which the most crucial is perhaps the deflned economy. These effects would make measures to benefit/defined contribution choice. The choice minimize costs, such as a prudent man rule of comes down to the balance between the (Section 3 (c)) and competitive fund management economic advantages of defined benefit noted in (Section 4 (d)), all the more urgent. We feel that Section I (superior insurance) and the practical compulsion In social security is sufficient; an difficulties with defined benefit that were efcient company pension sector, with discussed in Section 3 (ownership of surpluses, appropriate tax incentives, should be sufficient transfers, etc). A choice must also be made to attract employers and employees. between book reserves and separate funding. In our view, defined benefit plans retain an Some of the regilatory preconditions for advantage over defined contribution, given development of pension funds, which are covered their superior "employee retirement in Section 3, have already been noted in this insurance", especially if regulations are set to summary. They require a balance between cost to overcome the key problems. Only companies the sponsor, economic efficiency, equity and have proved able to offer defined benefit benefit security. Funding rules, adequate plans. And more generally, for both types, institutional stnictures (independent trustees etc.), company based schemes are superior to effective regulatory structures, rules on treatment personal pensions given lower transactions costs of surpluses, portability, vesting, indexation and and avoidance of market failures. Finally, protection against fraud are clearly essential. separate funding is felt superior to book More contestable are the need for quantitative reserves. not only because of the effects on the portfolio restrictions (Section 3(c)) and benefit capital market but due to the concentration of insurance (Section 3(h)). Apart from self risk in book reserves. investment limits, the former may reduce returns and increase risk, thus increasing costs unduly The case for preferential fiscal treatment is relative to a prudent man rule, while the latter outlined in Section 3 (a), and as shown, most of may entail either incentives to boost risk or the countries studied have found it persuasive. require stringent and costly portfolio restriction. The myopia argument was suggested as the most to protect the insurer. We suggest that the crucial, although reducing the burden of social following rules provide an appropriate balance; security and increasing saving may also play a a degree of mandatory indexation of pensions; role. Whether there is a case for special prudent man rules on asset allocation, with a treatment of pensions relative to other forms of ban on self investment (except for p0' 'folio saving may depend on the view taken that indexation purposes); minimum and contractual annuities have unique features in maximum funding rules tailored to the nature retirement income provision, absent from other of the obligations (given indexation, the IBOj, forms of saving. We consider these arguments but which do not discourage equity holding by persuasive, and hence suggest that pension penalizing temporary shortfalls; independence funds should be tax advantaged even if other of the fund from the employer; insurance forms of saving are not. against fraud; disclosure to members; indexation of accrued benefits for early These link in turn to the arguments for leavers; and vesting periods of around five compulsion. Compulsion is needed if the view years. A Dutch-styie supervisory structure 45 (one regulator, annua' checks on funding, whether the difficulties of an aging population are oversight of rules, occasional on-site really avoided by funding, if funds are invested in inspections) appears a good model to follow. domestic assets. The last point can be answered in two ways; first, property rights may be a more Finally, the advantages of pension funds relative secure basis for retirement than taxation, and, to other forms of saving include the superior second, the difficulties can in principle be avoided retirement insurance they offer, as outlined in by investing in countries with younger Section 1, as well as reduction of the populations.' On balance, it is suggested that demographic difficulties associated with pay-as- pension funds are a suitable means of old age you-go social security, as discussed in Section 2. support for countries at an appropriate state of Of course, the choice of funding itself raises development - where traditional means of numerous policy issues, referred to in Section family support for the old are breaking down, 2(a), such as inter- and intra- generational equity, and there is a reasonable degree of capitalist pressure on domestic rates of return, the costs of industrial development in which to invest - to tax exemption for funded pensions, and the costs supplement basic social security. A degree of to existhig workers in the transition (when they freedom to invest internationally is an essential have to pay both for their own funded schemes counterpart, to avoid demographic difriculties and the previous generation's pay-as-you-go and pressure on rates of return. pensions). Also there are important questions 46 REFERENCES Aldrich B (1982), "The earnings replacement rate of old-age benefits in 12 countries 1969-80", Social Security Bulletin, November, 3-11. Andrews E S (1990), "Pension portability in five countries", in eds. Turner J and Dailey L, "Pension policy; an international perspective", US Government Printing Office, Washington. Avery R B, Elliehausen G E and Gustafson T A (1985), "Pension and social security in household portfolios: evidence from the 1983 Survey of Consumer Finances", Research Papers in Banking and Financial Economics, Board of Governors of the Federal Reserve System, Washington. Barro R J (1974), "Are Government Bonds Net Wealth?", Journal of Political Economy, 82, 1095-1117. Berkowitz, Logue and Associates (1986), "Study of the investment performance of ERISA plans", prepared for the Office of Pension and Welfare Benefits, US Department of Labour, Washington DC. Bernheim B D and Shoven J B (1988), "Pension funding and saving", in eds. Bodie Z, Shoven J B and Wise D A, "Pensions in the US economy", University of Chicago Press. Blake D (1990), "Financial market analysis", McGraw-Hill, London. Blake D (1992), "Issues in pension funding", Routledge, London. Bodie Z (1989), "Pension funds and financial innovation", Working Paper No. 3101, National Bureau of Economic Research. Bodie Z (1990a), "Pensions as retirement income insurance", Journal of Economic Literature, Vol 28, 28- 49. Bodie Z (1990b), "Pension funding policy in five countries", in eds. Turner J and Dailey L, "Pension policy; an international perspective", US Government Printing Office, Washington. Bodie Z (1991a), "Shortfall risk and pension fund asset management", Financial Analysts Journal, May/June 1991. Bodie Z (1991b), "Inflation insurance", The Journal of Risk and Insurance, 57, 634-45. Bodie Z (1992), "Federal pension insurance: is it the S and L crisis of the 1990s?", paper presented at the Industrial Relations Research Meeting, New Orleans, January 1992. Bodie Z, Marcus A J and Merton R C (1988), "Defined benefit vs. defined contribution plans: What are the real tradeoffs?", in eds. Bodie Z, Shoven J B and Wise D A, "Pensions in the US economy", University of Chicago Press. Bodie Z and Merton R C (1992), "Pension benefit guarantees in the United States; a functional analysis", Working Paper 92/094, Harvard Business School. Charkhamn J P (1990), "Corporate governance and the market for control of companies", Bank of England Panel Paper No.25. 47 Charkham J P (1990), "Corporate governance and the market for companies: aspects of the shareholders' role", Bank of England Discussion Paper No.44. Cheetham C (1988), "A framework for pension fund management", Benefits and Compensation International, 18/6, 3-7. Clark R L (1990), "Inflation protection of retiree benefits", in eds. Turner J and Dailey L, "Pension policy; an international perspective", US Government Printing Office, Washington. Clark R L (1991), "Retirement systems in Japan", Pension Research Council, Irwin, Homewood, Illinois. Commissariat Generale du Plan (1991), "Epargnez, investir et croitre", CGP, Paris. Commission of the European Communities (1991), Proposal for a council directive relating to the freedom of management and investment of funds held by institutions for retirement provision", Brussels, 12/11/91. Cornell B and Roll R (1981), "Strategies for pairwise competitions in markets and organizations", Bell Journal of Economics, 12, 201-3. Davis E P (1986), "Portfolio behavior of the non-financial private sectors in the major economies", Bank for International Settlements, Economic Paper No. 17, BIS Basle. Davis E P (1988), "Financial market activity of life insurance companies and pension funds", Bank for International Settlements, Economic Paper No 21, BIS, Basle. Davis E P (1991), "International diversification of institutional investors", Journal of International Securities Markets, Summer 1991. Davis E P (1992), "Debt, financial fragility and systemic risk", Oxford University Press. Dean A, Durand M, Fall . J and Holler P (1989), "Savings trends and behavior in OECD economies", OECD Working Paper No. 67. Deutsche Bundesbank (1984), "Company pension schemes in the Federal Republic of Germany", Deutsche Bundesbank Monthly Report, August, pp 30-37. Diamond P A (1977), "A framework for social security analysis", Journal of Public Economics, 8, 275-98. Economist (1992), "Tomorrow's pensions", The Economist, 20/6/92. Feldstein M (1977), "Social security and private saving; international evidence in an extended life cycle model", in ed M Feldstein and R Inman "The economics of public services", International Economic Association. Feldstein M (1978), "Do private pensions increase national savings?", Journal of Public Economics, 10, 277-93. Friedman B M and Warschawsky M (1988), "Annuity prices and saving behavior in the US", in eds. Bodie Z, Shoven J B and Wise D A, "Pensions in the US economy", University of Chicago Press. Goode R (1992), "Consultation document on the law and regulation of occupational pension schemes", UK 48 Pension Law Review Committee, London. Gravelle J (1991), "Do individual retirement accounts increase saving?", Journal of Economic Perspectives, 5, 133-48. Grossman S and Stiglitz J E (1980), "On the impossibility of informationally efficient markets", American Economic Review, 70, 393408. Hagemann R P and Nicoletti G (1989), "Aging populations; economic effects and implications for public finance", Department of Economics and Statistics Working paper No 61, OECD, Paris. Hepp S (1990), "The Swiss pension funds", Paul-Haupt. Hepp S (1992), "Comparison of investment behavior of pension plans in Europe - implications for Europe's capital markets.", in ed. Mortensen J, "The Future of Pensions in the European Community", published by Brassey's, London, for the Centre for European Policy Studies, Brussels. Howell M and Cozzini A (1991), "Games without frontiers; global equity markets in the 1990s", Salomon Bros., London. Huiser A P (1990), "Capital market effects of the aging population", European Economic Review, 34, 987- 1009. Ibbotson R G and Sinquefield R A (1990), 'Stocks. bonds bills and inflation: historical returns', Dow Jones Irwin. Johnson P (1992), "The taxation of occupational and private pensions in Europe; the theory and the practice", forthcoming in ed. Mortensen J, "The Future of Pensions in the European Community", published by Brassey's, London, for the Centre for European Policy Studies, Brussels. King M A and Dicks-Mireaux L (1988), "Portfolio composition and pension wealth: an econometric study", in eds. Bodie Z, Shoven J B and Wise D A, "Pensions in the US economy", University of Chicago Press. Kollias S (1992), "The liberalization of capital movements and financial services in the Internal Market; implications and challenges for pension funds", forthcoming in ed. Mortensen J, "The Future of Pensions in the European Community", published by Brassey's, London, for the Centre for European Policy Studies, Brussels. Kruse D L (1991), "Pension substitution in the 1980s; why the shift toward defined contribution pension plans?", Working Paper No. 3882, National Bureau of Economic Research. Lakonishok J, Schleifer A and Vishny R W (1992), "The structure and performance of the money management industry", Brookings Papers: Microeconomics 1992, 339-91. Lazear E P and Moore R L (1988) "Pensions and turnover", in eds. Bodie Z, Shoven J B and Wise D A, "Pensions in the US economy", University of Chicago Press. Lusser M (1989), "A critical look at Swiss pension funds' investment policy", address to the Investment Foundation for Trade, Industry and Crafts, Zurich, 22/2/89. Lutjens E (1990), "The Legal Aspects of Dutch Supplementary Pension Plans", Benefits and Compensation 49 International, March, 2-10. Meade J E (1978), "The structure and reform of direct taxation", report of a committee set up by the Institute of Fiscal Studies, London: Allen and Unwin. Merton R C (1992), "Financial innovation and economic performance", Jornal of Applied Corporate Finance, 12, 12-22. Merton R C and Bodie Z (1992), 'A framework for analysis of deposit insurance and other guarantees", Working Paper No. 92/063, Harvard Business School. Metais J (1991), "Funds and portfolio management institutions in France, some recent trends", Working Paper No. RP 91/9, Institute of European Finance, Bangor. Mitra R (1991), "European pensions", Salomon Bros., London. Munnell A H (1984), "ERISA - the first decade: was the legislation consistent with other national goals?", New England Economic Review, November/December 1984, 44-63. Munnell A H (1986), "Private pensions and saving; new evidence", Journal of Political Economy, 84, 1013- 1. Munnell A H (1992), "Current taxation of qualified pension plans; has the time come?", New England Economic Review, March/April 1992, 12-25. Murakami K (1990), "Severance and retirement benefits in Japan", in eds. Turner J and Dailey L, "Pension policy; an international perspective", US Government Printing Office, Washington. Noble R (1992), "Reform of occupational pension funds", lecture to the LSE Financial Markets Group Regulation Seminar. OECD (1988), "Economic Survey of Denmark 1987/88", Organisation for Economic Cooperation and Development, Paris. Pestieau P et al (1991), "Pay as you go social security in a changing environment", Working Paper, CORE, University of Louvain. Pfaff M, Huler R and Dennerlein R (1979), "Old age security and saving in the Federal Repub.' - of Germany", in ed. G M von Furstemburg, "Social security and private saving", Ballinger, Cambridge, Mass. Riley B (1992), "Who gets what in pensions", Financial Times, 5/92. Schleifer A and Summers L (1988), "Breach of trust in nostile takeovers": In ed A Auerbach "Corporate takeovers; causes and consequences", University of Chicago Press. ,mahl W (1992a), "Changing the retirement age in Germany", The Geneva Papers on Risk and urance, 62, 81-103. mAhl W (1992b), "The future development of old-age security; challenges and response.", forthcoming ed. Mortensen J, "The Future of Pensions in the European Community", published by Brassey's, ,don, for the Centre for European Policy Studies, Brussels. 50 Smalhout J (1992), "The coming US pension bailout", Wail Street Journal, 12/6/92 Smith R S (1990), "Factors affecting saving, policy tools and tax reform", IMF Staff Papers, 37, 1-70. Stigum M (1990), "The Money Market", Dow-Jones Irwin, Homewood, Illinois. Tamura M (1992), "Improving Japan's employee pension fund system", Noruma Research Institut Quarterly, Summer 1992, 66-83. Turner J and Beller D J (1989), "Trends in Pensions", US Department of Labour, Washington DC Turner J and Daily L M (1990), "Pension policy: an international perspective", US Department of Labour, Washington DC Van Loo P D (1988), "Portfolio management of Dutch Pension Funds", De Nederlandsche Bank, Reprint 197. Venti S F and Wise D (1987), "IRAs and saving", in ed. Feldstein M S "The effects of taxation on capital accumulation", University of Chicago Press, Chicago. Vittas D and Skully M (1991), "Overview of contractual savings institutions", Paper WPS 605, PRE Working Paper Series, The World Bank, Washington DC. Vittas D (1992), "'The simple(r) algebra of pension plans", Mimeo, 'The World Bank. Zweekhorst K (1990), "Developments in private pensions in the Netherlands", in eds. Turner J and Dailey L, "Pension policy; an international perspective", US Government Printing Office, Washington. 51 NOTES l.The estimates are based on macro data for portfolios of the entire pension fund sector and the returns on the capital market as a whole, and hence do not provide a precise indication of performance of individual funds. 2.In principle, such risks can be reduced by gradually switching to less risky assets such as bonds or deposits prior to retirement. 3.It is suggested by Vittas and Skully (1991) that the "premia" for this insurance include restrictive vesting and transfer conditions. 4.These may be particularly severe for personal pensions, where asymmetric information is likely to be severe, and hence inappropriate products may be mold and/or investment performance may be poor. The main countervailing force is likely to be the institutions's desire to maintain reputation. 5.Indeed, Friedman and Warschawsky (1987) show that annuity markets in the US are far from actuarially fair, due to adverse selection, although they suggest a bequest motive is also needed to explain observed patterns of (low) purchase of annuities. 6.Defined benefit funds nonetheless impose some types of risk on workers, notably sensitivity of pensions to earnings late in the career, in the case of final-salary plans (Section 3(g). In some countries, they may also be vulnerable to inflation (see Section 4(a)). 7.'n practice there are some insurer-provided defined benefit plans in councries such as Switzerland, but they benefit from being compulsory, and are also provided through non-financial companies only, which reduces adverse selection. 8.Justifications for provision of social security benefits include paternalism, overcoming the problem that individuals may not cater for their own retirement due to myopia; information inefficiencies, whereby government helps each individual to save for retirement by providing a base level of benefits, without needing to gather information on the precise nature of each individual's preferences; overcoming adverse selection problems, th3t may plague private annuity markets - the main private sector protection against the risk that individuals will run out of retirement savings before they die; and overcoming free rider problems, that individuals will not save for retirement if they know there is a safety net, and thus they should be forced to contribute to it. (See Bodie and Merton (1992).) Several of these parallel the case for compulsory or tax-advantaged private pension funds (Section 3). Also governments, via the power to tax, have ability to pay pensions that are inflation indexed, and may find social security a convenient means to redistribute income 9.Such redistribution may be justified, for example, if the growth rate is rapid and the young are n,uch more productive than the elderly. lO.Because of their redistributive element, pay-as-you-go schemes generally have to be compulsory. 11.Swedish pension fund assets shown in the table, which are also very large, relate to the national supplementary (ATP) pension scheme, and are included for comparative purposes. 12.The paper does not discuss long-term saving in detail, although it clearly determines the size of total wealth of which pension wealth forms a part. It is likely to depend on factors such as income growth (old age security appears to have a large income elasticity of demand), demographic factors (the proportion of the population in the high-saving groups aged 35-65) inflation 52 and social uecurity provision. 13.Note that in practice the effective retirement age has tended to fall in most industrial countries, as employers seek cheaper and supposedly more flexible younger workers. Final salary schemes may have accelerated this tendency, as assuming wages of the employee would otherwise have risen, early retirement saves on pension payment rates. 14.Many workers also take further employment at a lower wage and status, after "retirement". 15.There is no obligation for such provisions to cover all pension promises. 16.Sometimes called "support funds". 17.It is, of course, compulsory once it forms part of a contractual collective agreement. 18.However, in the United Kingdom the 'personal equity plan' scheme makes a move towards reducing the tax disadvantages of direct equity holdings. Growth of funds has not accompanied a reduction of equity holdings in Japan or Germany; there is no capital gains tax in Japan, while in Germany it only applies to short-term gains. 19.Private pensions in France are also compulsory but operate on a pay-as-you- go basis. 20.Blake (1992) quotes an equivalent figure of £15 billion for the UK. 21.Similar patterns are apparent in Germany, where in a 1982 survey of pensioners, three quarters of former senior managers but only half of former wage earners received private pensions. 22.Note, however, that some of the assets backing unfunded commitments in Germany and Japan are in the form of cross-shareholdings with other firms, i.e., there may be some degree of diversification. Conversely, any rundown of such holdings when pensions are paid may contribute to the unwinding of many of the associated relationships. 23.For example, in France, pension funds must invest at least 34% of their assets in state bonds. 24.Estimates of returns are presented in Section 4(c). 25.In practice, life insurers are more strictly regulated, see Davis (1991). 26.Paradoxically, and despite their indexed obligations (Section 2), they tend to invest conservatively in fixed interest assets (Section 4). 27.The Germans have sought to exempt Pensionskassen from the Pension Funds Directive on the grounds they are insurance companies, but this may be inconsistent with their existing derogation from the Insurance Directives. 28.As discussed below, German schemes are forced to index but unable to fund the IBO tax free. 29.Such rules makes it optimal to hold "real assets" to avoid underfunding. 30.The part of pensions which replaces the state pension is fully indexed in line with average earnings up to retirement. 31.Another company scheme, a personal pension, an annuity or rights in the state earnings related pension scheme. 53 32.Given the speed with which pension funds can change their risk exposure, it could be argued that anything short of continuous monitoring could lead to this hazard. 33.Clearly, if underfunding exceeds 30% of net worth, the operation is profitable for the firm. 34.To the extenit that PBGC insurance leads to imposition of high charges on well funded plans to subsidize underfunded ones, it could be directly counter productive to this goal, as financially sound sponsors wind up their defined befit plans. 35.Bodie and Merton (1992) suggest this is an underestimate, as it is based only on plans that have defaulted to date, rather than estimates of losses on currently insured plans that will default in the future. 36.Custody services include safekeeping, settlement, tax, dividend receipts, dealing with rights issues and stock lending. 37.Note, however, that published accounts did not hide the high level of self investment by the Maxwell funds. The problem for members was of understanding the associated risk - and the difficulty of redress. 38.In addition, the investment management regulatory bodies regulate insurance companies and other financial institutions offering pensions, asset managers and those offering advice to individuals regarding pensions. 39.It also assesses whether funds meet the standards to contract out of the state earnings related pension, checks revaluation and preservation of rights for early leavers, etc. 40.Their fiduciary duties are to hold the assets in trust for members, act impartially, keep accounts, check funding is in place, and seek expert advice when necessary. 41. Since they get the surplus assets in the case of winding up. 42.For example, when a mature scheme is underfunded, and pensioners may prefer it to be wound up (since the ABO gives them all they wish), while the employees may prefer to take the risk that the employer will fund the shortfall later on. 43.In practice, it has been announced that except for plan terminations, the requirement for indexation will not be brought into effect until the implications of the European Court judgment on equal pension ages (the "Barber Judgment") have been clarified. 44.In practice, Danish contributions tend to be around 10-15%. 45.For example, Vittas (1992) shows that in a defined contribution plan with a contribution rate of 10%, 40 years' contributions, 20 years' retirement and real wage growth of 2%, real returns of 3% will only obtain a replacement rate for an indexed pension of 33%, while a 5% real return obtains an indexed pension of 60% of final salary. 46.Foreign yields were constructed using the country's effective exchange rate and the average of yields in the UK, US, Germany, Japan, and Canada (excluding the country in question for its own foreign return). 47.The data in Tables 9 and 10 add up to those in Table 8. 54 48.See Section 5(a). Bodie suggests thiat given such rules, it is a paradox that US defined benefit funds invest in equities. He suggests it is perhaps because management sees a plan as a trust for employees, and manages assets as if it were a defined contribution plan (i.e., for employee welfare), with a guaranteed floor given by the benefit formula. 49.However, Bodie (1990) disputes the utility of equity as an inflation hedge and suggests investment in equities can be seen merely as boosting expected returns for the benefit of members. 50.It reached 84% at end-Ql 1992. 51.Merton (1992) has suggested that using stock index swaps may be a way for LDCs to achieve the benefits of inward international diversification by pension funds from major countries without transfer of capital resources. By separating capital flows from risk sharing, it avoids capital imbalances or foreign intervention in domestic capital markets. 52.For a discussion of life insurance companies' and pension funds' foreign investment see Davis (1991). 53.Hence high volatility for Danish (defined contribution) funds is more serious than in the UK, where funds are largely defined benefit. 54.If reform of the portfolio rules is not possible, and subject to administrative costs, one might question whether pay-as-you-go might not be a better solution in those countries. 55.In Germany, pension funds may not delegate fund management. 56.Fund managers in Japan are themselves subject to restrictions on diversification, and may not have more than 50% of their managed assets overseas. 57.See Bodie (1991a). 58.However, King and Dicks-Mireaux (1987) found little effect in Canada. 59.See Feldstein (1977). However, analysts in countries such as Germany dispute this effect (Pfaff et al (1979)) and suggest social security had no effect on saving. 60.Charkham (1990). 61.Others may be lower population growth, increased life expectancy and social change which reduces the role of the extended family. 62.Large firms already able to access the international capital markets would be less affected. 63.However, Huiser (1990) asserts that there will be further growth in the Netherlands, despite the large size of existing Dutch funds. 64.Note many ldcs also have pay-as-you-go social security. 65.In practice, pension funds are unwilling to invest in LDCs given illiquidity, transfer risk, settlement risk, etc (see Davis (1991). APPENDIX I RETURNS ON LOCAL GOVERNMENT AND PRIVATE FUNDS The data for the UK and US allows a futher comparison of effects of ownership and management methods to be made, this time in the sarne markets, in that local governmnent funds data can be identified separately from private-sector funds. Mean (standard deviaton) of annual real total returns (domestic currency) UK: Local authority funds 4.9 (13.4) Private funds 5.6 (13.0) US: State and local funds 1.2 (12.6) Private funds 2.7 (11.7) * 1967-1988 for the UK In each case, local goverm-nent funds obtain lower returns than private funds. This can be related to more conservative portfolio distributions. UK local authority funds held an average of 52% equity over the sample, while private funds held 56%. For US funds the difference is more dramatic; 25% and 53%. Interestingly, the risks in real tenn3 were higher for the local government funds, given the volatility of real returns on bonds (see Table 7). 55 APPENDIX II This appendix provides average real returns calculated over the period 1971 90 as well as the changes that would be implied by standardized portfolio structures. 56 TAk 1: C0uacsrle of red bla riinml 1971-W Mm (_adard deviaon) of real WuVwhl pnod mum (domesi crency) hr am UIdW UJbd G_may J_p" Cama Nedmrda swum Ihcnk Swbmd Mie: S_w de Fern LoA 3.9 (2.6) 1.2 (5.4) 5.2 (2.3) 0.9 (4.7) 4.2 (3.3) 3.9 (3.9) 3.0 (3.3) 6.5 (3.5) 2.4 (2.1) 2A (33) Monw es 3.1 (14.2) 1.3 (5.7) 4.5 (1.5) 2.7 (5.2) 3.4 (13.2) 4.3 (28) 2.4 (3.1) 6.3 (3.7) 12 (2.5) 3.3 (2.7) 80ties 5.9 (14.9) 10.8 (31.3) 9.3 (20.4) 11.2 (21.0) 5.0 (17.8) 8.6 (30.1) 9.3 (23.7) 9.4 (29.4) 4.7 (22.2) 9.6 (23.3) Bons 1.2 (15.0) 1.6 (11.5) 2.6 (15.1) 0.0 (20.3) 1.1 (12.8) 1.3 (11.5) 0.6 p.6) 4.5 (17.0) -1.7 (13.7) 1.3 (13.9) Showm asset 2.1 (2.7) 1.5 (5.3) 2.9 (2.3) 40.7 (5.0) 2.6 (3.6) 1.7 (4.3) 0.9 (3.7) 1.7 (1.6) 1.1 (2.3) 1.9 (3.4) propel 3.9 (5.5) 5.7 (13.0) 4.5 (2.8) 6.6 (7.2) 5.2 (5.8) 4.6 (15.0) - Foreign bonds 2.2 (16.0) .0.4 (16.8) 3.8 (12.2) 2.9 (15.8) -1.2 (13.9) *0.1 (12.2) 0.6 (13.9) -1.0 (12.9) -1.6 (14.9j 0.4 (14.3) Foreign equhies 9.6 (18.2) 6.2 (17.1) 10.6 (16.0) 8.6 (20.5) 6.4 (15.9) 6.7 (14.9) 7.4 (14.5) 5.9 (14.7) 5.2 (17.1) 7.2 (14.2) Maewomma. inem: laltma (CPI) 6.3 (3.2) 9.8 (5.4) 3.3 (2.2) 5.5 (5.8) 6.9 (3.0) 4.9 (3.3) 3.4 (2.6) 7.9 (3.4) 4.2 (2.6) 8.0 (4.0) Redemption yield o govermesv bonds 2.8 (3.4) 1.9 (4.3) 4.0 (1.4) 1.5 (4.7) 3.3 (3.1) 3.3 (2.9) 2.2 (3.0) 5.7 (2.2) 0.7 (1.9) 3.0 (3.0) Real eaings growth 0.5 (2.2) 2.4 (2.5) 3.6 (2.5) 3.0 (3.5) 1.1 (2.7) 1.4 (2.6) 1.1 (3.4) 25 (3.7) 1.6 (2.0) 56a Table 2: Pension fund returns (1971-90) Mean (Standard deviation) of annual real total returns (domestic currrency) Percent US UK Gennany Japan Canada Netherlands Sweden Denmark Switzerland Using holding period returns on bonds (all countries) and fixed rate mortgages (US & Canada) 4.0 (12.2) 7.4 (15.9) 5.2 (4.7) 4.9 (10.0) 2.7 (10.5) 4.3 (6.3) 0.0 (7.7) 4.9 (13.4) 1.5 (7.7) Average earnings 0.5 (2.2) 2.4 (2.5) 3.6 (2.5) 3.0 (3.5) 1.1 (2.7) 1.4 (2.6) 1.1 (3.4) 2.5 (3.7) 1.6 (2.0) Portfolio return less average eamings 3.5 5.0 1.6 1.9 1.6 2.9 -1.1 2.4 -0.1 Governmentbonds 1.2 (15.0) 1.6 (11.5) 2.6 (IS.1) 0.0 (20.3) I (12.8) 1.8 (l1.5) -0.6 (8.6) 4.5 (17.0) -1.7 (18.7) Market paper 2.6 (2.7) 1.5 (5.3) 2.9 (2.3) -0.7 (5.0) 2.6 (3.6) 1.7 (4.3) 0.9 (3.7) 1.7 (1.6) 1.1 (2.3) Equities 5.9 (14.9) 10.8 (31.8) 9.3 (20.4) 11.2 (21.0) 5.0 (17.8) 8.6 (30.1) 9.3 (23.7) 9.4 (29.4) 4.7 (222) Memo: using rcdemption yioIds on fixed rate instentents 4.8 (7.9) 7.5 (13.8) 5.4 (3.3) 3.5 (2.7) 4.1 (5.5) 4.5 (5.8) 2.6 (3.1) 6.5 (2.5) 2.1 (4.1) 56b Table 3: Ma (stadard devidon) of rea! total returns on divenrifd porfolos 1971-1990 Dometic' Domestic & Memo: Per cant InternatonaP Column 2 less Average Earnings United States 3.5 (13.4) 4.0 (13.2) 3.5 United Kingdom 6.2 (19.6) 5.5 (17.1) 3.1 Germany 5.9 (14.9) 6.2 (13.3) 2.6 Japan 5.6 (16.9) 5.6 (15.6) 2.6 Cauda 3.1 (12.1) 3.0 (11.7) 1.9 France 5.4 (19.7) 5.1 (17.3) n/a Netherlands 5.2 (18.4) 4.8 (16.4) 3.4 Sweden 4.3 (13.5) 4.2 (12.1) 3.1 Switzerlad 1.5 (16.9) 1.5 (14.9) 0. 1 Denmark 6.9 (20.0) 6.0 (17.4) 3.5 1 50% domestic equity, 50% domestic bonds. 2 40% domestic equity, 40% domestic bonds, 10% foreign equity, 10% foreign bons. Note: International diversificadon should cause a convergence in return. by increasing below average returns and lowermg those sbove average. This would occur if purchasing power parity held all the time. In fact, large deviations between exchange rates an purchasing power parities may prevent this convergence from materializing, at least in the short to medium run until corrections in exchange rates take place. 56c Policy Research Working Paper Series Contact Title Author Date for paper WPS1208 Primary School Achievement in Levi M. Nyagura October 1993 I. Conachy English and Mathematics in Abby Riddell 33669 Zimbabwe: A Multi-Level Analysis WPS1209 Should East Asia Go Regional? Arvind Panagariya October 1993 D. Ballantyne No, No, and Maybe 37947 WPS1210 The Taxation of Natural Resources: Robin Boadway October 1993 C. Jones Principles and Policy Issues Frank Flatters 37699 WPS1211 Savings-investment Correlations Nlandu Mamingi October 1993 R. V%' and Capital Mobility in Developing Countries 31047 WPS1212 The Links between Economic Policy Ravi Kanbur October 1993 P. Attipoe and Research: Three Examples from 526-3003 Ghana and Some General Thoughts WPS1213 Japanese Foreign Direct Investment: Kwang W. Jun November 1993 S. King-Watson Recent Trends, Determinants, and Frank Sader 33730 Prospects Haruo Horaguchi Hyuntai Kwak WPS1214 rrade, Aid, and Investment in Sub- Ishrat Husain November 1993 M. Youssef Saharan Africa 34637 WPS1215 How Much Do Distortions Affect William Easterly November 1993 R. Martin Growth? 39065 'iPS1216 Resulation, Institutions, and Alice Hill November 1993 D, Evans Commitment: Privatization and Manuel Angel Abdala 38526 Regulation in the Argentine Telecommunications Sector WPS1217 Unitary versus Collective Models Pierre-Andre Chiappori November 1993 P. Athpoe of the Household: Time to Shift the Lawrence Haddad 526-3002 Burden of Procr" John Hoddinott Ravi Kanbur WPS1218 Implementation of Trade Reform in John Nash November 1993 D. Ballantyne ' ub-Saharan Africa: How Much Heal 37947 and How Much Light? WPS1219 Decentralizing Water Resource K. William Easter November 1993 M. Wu Management: Economic Incentives, Robert R. Hearne 30480 Accountability, and Assurance WPS1220 Developing Countries and the Bernard Hoekman November 1993 L. O'Connor Uruguay Round: Negotiations on 37009 Services Policy Research Working Paper Series Contact Title Author Date for paper WPS1221 Does Research and Development Nancy Birdsall November 1993 S. Rajan Contribute to Economic Growth Changyong Rhee 33747 in Developing Countries9 WPS1222 Trade Reform in Ten Sub-Saharan Faezeh Foroutan November 1993 S. Fallon Countries: Achievements and Failures 38009 WPS1223 How Robust Is a Poverty Profiie? Martin Ravallion November 1993 P. Cook Benu Bidani 33902 WPS1224 Devaluation in Low-Inflation Miguel A. Kiguel November 1993 R. Luz Economies Nita Ghei 39059 WP31225 Irtra-Sub-Saharan Afri-)n Trade: Faezeh Foroutan Ncve-mber 1993 S, Failon Is It Too Little9 Lart Prtchett 38009 WPS1226 Forecasting Volatility in Conrmod!ly Kerineth F. Krcner November 1993 F. Hatab Markets Devir P Kneatfey 35835 St; n C:aessens WPS1227 Designing Water Institutiors: Marie Le gh Livingston December 1993 C. Spooner Market Fail.res and Instviut?nal 30464 Response WPS1228 Compelis.;on. Ccmpetlon Pol cy, B:cad M Hoekrran December 1993 L. O'Connor and the CAl I P&-Os C iIavrclcI.s 37009 WPS1229 'The Struct.~e, Reguia!:un, a~ a E f [),j December 1993 P Intanze Per4ormance of Pension FLnar ,.d 37642 Nlre Irneus!n:i COrJ!r es