CHAPTER SIX Some Considerations Relevant to Prefunded Pensions in France

Paul Masson
Published Date:
October 1995
  • ShareShare
Show Summary Details
Joaquim Levy

Pension schemes in France have provided increasing coverage and benefits during the last 30 years, but their liabilities are essentially unfunded and the aging of the population has cast doubts on the long-run viability of such a pay-as-you-go system. This uncertainty has been acknowledged, but proposals made in the study published by the French authorities in 1991 (Livre Blanc sur les Retraites) were limited to changes in some parameters ruling the current system and did not include the implementation of prefunded pension schemes. This choice was justified by concerns about the guarantees pension funds could ultimately provide to pensioners, income distribution, and the cost of financing a transition from the current system, but apparently did not consider the often salutary effects on capital markets and savings brought about by institutional investors such as pension funds. Therefore, in contrast to the situation in several other developed countries where private pension funds play an important role in providing income for retired workers, large-scale adoption of prefunded pensions in France is not yet certain.

Nevertheless, a significant number of French workers and households have shown growing interest in financial instruments that share several features with pension funds: they have increasingly purchased life insurance and annuities and participated in company-sponsored savings funds. Such an interest is an indication that markets may be ready for a gradual implementation of pension funds, and since recent research has shown that a transition from a pay-as-you-go system toward prefunded schemes can be achieved without hurting any generation while benefiting some, a choice in this direction may be warranted.

This chapter discusses current and forecast conditions of the pension system in France (Section I); the expansion of life insurance and other contractual savings in recent years (Section II); and aspects of a transition toward pension funds (drawing on the experience with funds in other OECD countries), including effects of pension funds on capital markets, and suggestions on how to shift some of the liabilities of the currently unfunded schemes to funded pension funds (Section III).

I. The Pension System in France

Features of the Pension System

In France, the average standard of living of retirees improved significantly during the 1970s and 1980s, and it is equivalent to that of most working persons. Generous revaluations of benefits (30 percent in real terms since 1970), decreases in the minimum retirement age (to 60 years in most cases), as well as the possibility of cumulating pensions from different sources, removed retirees from the poorest segments of the population. In fact, statistics indicate that, in France, adjusted individual income peaks at age 66-70, when it is 25 percent higher than at age 31-40, and that around 80 percent of retirees’ income is provided by pensions, including old age benefits (minimum vieillesse). In addition to having a relatively high income, the majority (70 percent) of those aged in the 60-70-year bracket own at least one house, in contrast to those below 40 years, who in most cases (51 percent) do not own the place where they live. This affluence does not fully extend to those older than 70, who often have lower pensions, but it suggests that during coming years the number of relatively well-off retirees will increase. This bright scenario, however, may be difficult to achieve, given the current method of operation of the pension system in France.

The French pension system comprises a large number of unfunded schemes, which grew up along occupational lines. These schemes can be broadly divided among those for private sector employees and the selfemployed, and those for public sector employees and workers in “special categories,” such as railways and the merchant navy. Schemes in the first group usually comprise two levels: the basic pension (régime général) and the supplementary pensions (régimes complémentaires) grouped around the Association Générale des Institutions de Retraites des Cadres (AGIRC) for employed professionals and the Association des Régimes de Retraites Complémentaires (ARRCO) for other categories.1 Contributions to both levels are compulsory and subject to ceilings (except for the additional 1.6 percent contribution paid by employers (Table 1)).2 Enterprises are required to participate in an industry or assimilated scheme.

Table 1.France: Contribution Rates to Pension Schemes(In percent of wages, 1991)
Contribution RateOverallEmployerEmployee
General pension scheme16.358.206.55
Source: Ministère des Affaires Sociales, de la Santé, et de la Ville.

Additional paid by employer, not subject to a ceiling.

Supplementary pension (schedule C).

Source: Ministère des Affaires Sociales, de la Santé, et de la Ville.

Additional paid by employer, not subject to a ceiling.

Supplementary pension (schedule C).

Benefits from the basic pension are determined by the number of years the beneficiary contributed, as well as by the average wage earned over a certain number of years, indexed by either changes in price levels or average wage increases.3 Benefits from supplementary pensions are proportional to the number of points the individual accumulated before retiring. These points are usually purchased; the ratio between the annual pension each point secures and its cost is known as the return ratio (rendement) of contributions. Cumulation of pensions from different sources is allowed and most persons receive more than one pension: the Livre Blanc indicates that retirees receive pensions from an average of 1.5 basic schemes and 1.3 complementary schemes.

Contributions have risen during the last decade, and currently they average about 2.0 percent of total gross labor income. The supplementary pensions were able to combine the increase in contributions with a fall in the effective return ratio.4 Despite these increases in contributions by workers (and, in the case of the supplementary pensions, the increasing adjustment effort by both workers and retirees), the cost of financing an increasing number of retirees for a longer period of time (due to an increase in life expectancy and decreases in retirement age) has strained the pension system in the last few years. For society as a whole, the increasing cost of pensions was reflected in the growth of pension payments, which, during the last ten years, swelled in real terms at an annual average rate of 4.5 percent, while the economy grew at only a 2.2 percent rate. These effects have been aggravated since 1991 by the slowdown of the economy.

Medium- and Long-Term Perspectives

Simulations published in the Livre Blanc show that by 2040 there would be between 1.3 and 1.7 workers for each retiree, compared with a ratio of 2.15 existing in the early 1990s, and that contributions would have to increase accordingly if pension benefits were to keep growing in line with wages.5 A high fertility rate and a high activity rate would each reduce the dependency ratio (the ratio of retirees per worker) by 10 points, while lower unemployment would have only a marginal impact on it. Changes in dependency ratios were expected to require increases in contributions in a range of between 66 and 127 percent, implying a contribution-to-wage ratio of up to 41 percent (Table 2).6

Table 2.France: Forecasts of Dependency Ratios and Contribution Rates in the Pension System in 2040
Dependency Ratio
Low Unemployment
Low participationHigh participation
Low fertility0.750.66
High fertility0.660.58
High Unemployment
Low participationHigh participation
Low fertility0.780.68
High fertility0.680.60
Contribution Rates
(In percent)

Low Unemployment
Low participationHigh participation
Low fertility40.5 (25.1)135.1 (24.4)
High fertility35.4 (25.0)30.9 (24.3)
High Unemployment
Low participationHigh participation
Low fertility41.9 (26.2)36.3 (25.5)
High fertility36.7 (26.0)32.0 (25.3)
Source: Livre Blanc sur les Retraites.

2010 figures in parentheses.

Source: Livre Blanc sur les Retraites.

2010 figures in parentheses.

In the absence of either increases in contributions or decreases in benefits, large financial shortfalls would develop before 2010, reaching some 370 billion of 1990 francs (about 4 percent of GDP) by that date, and worsening thereafter. These shortfalls are to be compared with a shortfall equivalent to 1.1 percent of GDP in 1990, and are concentrated mostly in the basic system.

The Livre Blanc suggested three main changes in social security in order to balance the system in 15 years: lengthening the contribution period required to obtain a full pension, lengthening the period taken into consideration when computing benefits, and indexing benefits to the consumer price level (CPI), instead of to the average wage level.7

The main suggestions in the Livre Blanc were adopted in 1993.8 However, the problems after “baby boomers” start to retire have not been solved, and maintaining a pay-as-you-go system in the long run would still require contributions above 30 percent of gross wages, which may be infeasible. In fact, it is possible that workers and households have anticipated that, and, taking advantage of capital market liberalization and low inflation (which increased the supply of financial instruments and reduced the risk of inflationary depreciation of returns), looked for options in the private sector even in the absence of private pension funds as they exist in other countries.9 The next section provides an overview of the savings institutions recently favored by workers and households in general. This analysis is followed by a discussion of aspects of a transition to prefunded pensions.

II. Company-Sponsored Funds and Life Insurance in France

In discussing the possible effects of pension funds it is important to emphasize that there have been profound changes in French capital markets in the last few years.10 Among the changes discussed in this section are the growth of life insurance and company-sponsored savings plans (FCPEs, or fonds communs de placement d’entreprise—a sort of closed-end mutual fund—and PEEs, or plans d’épargne d’entreprise). Because some of their features are similar to those of pension funds, and because they are provided by the private sector, both instruments are often considered attractive starting points for the establishment of a widespread system of pension funds in France.11 Their growth in recent years has been strong, and their assets correspond to about 20 percent of GDP (Table 3).

Table 3.France: Stock of Financial Assets(End of period, in billions of francs)
Life insurance7789421,151
Memorandum items
Listed shares1,6351,6192,120
Source: Banque de France.
Source: Banque de France.

The Life Insurance Sector

Life insurance policies and annuities absorbed about 60 percent of household financial saving in 1993, reflecting the growth of the sector in recent years, which can be attributed to three main factors: financial liberalization in the 1980s, tax advantages, and a new perception of the longerterm needs of workers.12

The first factor leading to growth in the life insurance industry was the permission granted in 1985 to banks and other financial institutions to sell insurance policies through their insurance subsidiaries (and the liberalization of the type of product that could be offered).13 Currently, these subsidiaries represent about half of the market. Competition has not only increased the volume of retail outlets, but also the yield provided by the average policy, which passed along the high interest rates available in recent years. The second factor comprises an income tax deduction (prime) proportional to the amount invested and the income tax exemption of capital gains on savings held for at least eight years, as well as favorable treatment of bequests.14 The fiscal expenditure implied by tax-exempting returns on long-term savings in France can be roughly estimated to be around F 20 billion a year; the budgetary cost of tax reductions is of the same order of magnitude.15 Together, they correspond to about 5 percent of the revenues generated by the income tax. The third factor reflects the concern of economic agents about their future income, and the desire of those already owning a house to diversify their assets.

Recent changes in life insurance regulations are expected to increase the length of contracts and the demand for life insurance in coming years, without adding fiscal incentives. A fidelity clause allowing insurers to offer special rates for customers who keep their policies active for many years is expected to extend the length of contracts beyond the eight-year tax exemption period and reduce the risk insurers face of large withdrawals. However, at this stage, the absence of formal estimates hampers a precise evaluation of the effectiveness of that clause. In addition, flexible distribution of surpluses, new disclosure rules, and the recent update of the technical bases for computing annuities and premiums are expected to increase the demand for insurance. Such a flexible distribution of surpluses makes life insurance similar to defined-contribution pensions in the sense that it permits insurers to offer part of the upside of total returns, that is, to shift part of the risk (and expected returns) to investors.16

Features of Company-Sponsored Savings

In contrast to life insurance policies purchased by households on an individual basis, company-sponsored savings programs operate under the framework provided by regulations requiring firms with more than 100 employees to share part of their profits with the workforce.17 This legislation mandates that the share belonging to employees (participation) be frozen for a period of five years. During this period, resources can be held in bank accounts or in FCPEs.18 In early 1993, these resources amounted to about 80 billion francs, half invested in funds.

Workers’ share of profit is not the only source of funds to FCPEs. Firms—on their own initiative, or as the result of an agreement between labor and management—can create special funds to collect the participation and other participatory benefits. These funds—PEEs—enjoy several tax advantages, again reflecting the policy of granting fiscal incentives for long-term savings.19 Since 1959, workers have been able to receive incentive payments tied to productivity measures (intéressement), which can also be deposited in tax-favored investment instruments.20 By 1990 about 2 million employees (about 10 percent of the labor force) had signed agreements to receive incentive payments. Employers’ voluntary contributions (abondements) and voluntary worker deposits are also an important source of financing of PEEs, as reported in Table 4.21

Table 4.France: Sources of Funds for PEEs in 1989(In percent)
Profit-SharingIncentive Payments and Workers’ Voluntary DepositsEmployers’ Deposits
Source: Commission des Opérations de Bourse (COB).
Source: Commission des Opérations de Bourse (COB).

PEEs enjoy great freedom in choosing their investment policies, while—because they are institutions handling contractual savings (as are pension funds)—they are overseen by representatives of employees.22 Funds deposited in a PEE can be invested in individual accounts or in many types of closed-end mutual funds. The vast majority of FCPEs (including those linked to PEEs) are managed by financial institutions outside the sponsoring company, mainly banks and insurance companies, under the supervision of a board that has a majority of representatives of employees.23 Competition among financial institutions for managing these increasingly large funds (Table 5) has led them to supply a variety of savings instruments, allowing companies to offer funds tailored to workers’ demands. It has also led financial institutions to charge relatively low management fees.24

Table 5.France: Total Assets in Mutual Funds of Enterprises(In billions of francs)
Source: COB.
Source: COB.

Portfolios of Life Insurance Companies and FCPEs

The aggregate portfolios of life insurance companies and PEEs indicate how contractual savings are invested in France. Their distributions of assets differ, in part because the former offer a more standard product while the latter offer tailored products that are sometimes a source of financing for the sponsoring firm. A brief analysis of the portfolio of life insurance suggests that the current distribution of assets is close to the efficient portfolio frontier, defined as the minimum variance portfolio for a given rate of return.

The portfolio of the insurance industry is mostly composed of fixed income assets, including both government and private paper (Table 6). This is explained by the liquidity and relatively high returns of those assets in recent years, and not by binding regulations. In fact, since 1990, the obligation to invest at least 34 percent of the portfolio in bonds and notes has been eliminated, and only maximum limits on the other types of assets currently exist. The proportions of real estate and loans have markedly decreased over the years as financial markets developed and low inflation endured, and are well below their respective maxima of 40 and 10 percent. The proportion of stocks is also below the allowed ceiling of 65 percent. Chart 1 suggests that, although conservative, the portfolio of the life insurance sector is efficient, that is, for the level of expected return chosen, the variance is virtually minimized. Chart 1 shows the efficient portfolio frontier generated by holdings of bonds, real estate, and a diversified portfolio of stocks, considering annualized returns for the period 1979–93. It also shows the position of the aggregate portfolio held by insurance companies.25

Chart 1.France: The Portfolio Frontier

(At constant prices, 1979-93)

Sources: Wharton Econometrics Forecasting Associates; Commission des Opérations de Bourse; and IMF staff calculations.

Table 6.France: Distribution of Life Insurance Portfolios(In percent)
Fixed income52.761.3
Real estate19.510.7
Negotiable debt10.3
Source: Ministry of Finance.
Source: Ministry of Finance.

An increase in the proportion of stocks would be desirable as a way to improve the protection of the portfolio against inflation, especially if the average holding period of policies is to be stretched beyond ten years. Recent changes in regulations concerning the distribution of surpluses (see above) are expected to work in this direction, by increasing the attractiveness of real assets at the expense of guaranteed returns. An increase of 5 percentage points in the proportion of stocks—along the portfolio frontier—would permit an increase in expected returns of 0.6 percentage points, while increasing the volatility of the portfolio by about 40 percent. Such a portfolio adjustment would create a demand for about F 70 billion in equity—the equivalent of 150 percent of privatizations in the 1993–94 period.

The aggregate portfolio of FCPEs already includes a significant proportion of stocks (Table 7). As noted before, there are almost no restrictions on the type of financial assets held by PEEs, allowing them to take advantage of the relatively long-term nature of deposits benefiting from tax incentives and invest a significant proportion of their assets in stocks. The average period of a deposit in a PEE oscillates between 7 and 7½ years, thus actually exceeding the five-year blocking period determined by the legislation. Although the legislation does not impose a cap in the share of PEEs’ portfolio invested in stock issued by the sponsoring company, on average this proportion is below 20 percent.

Table 7.France: Asset Composition of FCPEs(In billions of francs)
1986199019931993 Share (in percent)
Mutual funds8.818.727.423.2
Of which issued by the company5.911.823.620.0
Fixed income10.419.634.229.0
Of which issued by the company1.
Source: COB.
Source: COB.

The analysis of the portfolio of PEEs and life insurance suggests that the expansion of contractual savings (e.g., pension funds) could lead to more financing to firms (taking into account the importance of equity in the portfolio distribution of FCPEs, as well as the size of the portfolio of life insurance companies) and that management fees and transactions costs could further decrease as the volume of savings increases and competition to manage these portfolios develops. The effects of pension funds on the capital markets of other developed countries is one of the topics discussed in the following section, together with the macroimplications of a greater reliance on savings instruments as sources of income to retirees, and how this income might be protected against inflation.

III. Aspects of a Transition Toward Prefunded Pensions

This section attempts to address the main concerns usually expressed about the adoption of prefunded pension schemes (i.e., the intergenerational cost of a transition, the degree of protection against inflation such funds can provide, and the relative advantages of defined-benefit and defined-contribution schemes), discusses expected effects on capital markets, and examines some implications of supporting prefunded pensions other than by simply creating yet another tax-favored savings instrument.

Intergenerational Cost

It is well known that when total labor compensation is growing fast—because of either a high rate of population growth or technical progress reflected in sustained increases in real wages—a pay-as-you-go system may be more efficient than a prefunded system.26 However, given the population growth projected for Europe in the coming decades, the above conditions do not seem to apply. It has also been highlighted that contributions to social security most often act as a tax on labor, distorting and reducing labor supply.27 On the other hand, because savings are the result of an intertemporal decision about consumption, and not about labor supply, contributions to prefunded pension plans are thought to be neutral in relation to the latter. Taking these elements into account, and in particular the effects of taxes on labor supply, overlapping generation models with endogenous labor supply indicate that prefunded pensions can be more efficient than a pay-as-you-go system (when population growth is modest), in some cases permitting a switch from the latter to a funded system in a way that does not hurt any generation, that is, permitting a Pareto-improving switch.28

Implementing such a transition (which would not hurt any generation and therefore addresses the concerns raised in the Livre Blanc) would probably require a transitory increase in public debt.29Hornburg (1990) suggested a capital reserve system with government debt (to effect intergenerational transfers), similar to Raffelhüschen (1993) who also suggests the introduction of credit instruments. The increase in public debt (which simply reflects the implied liabilities of the pension system) would not offset the increase in savings, permitting an increase in capital accumulation. Because in France the implied liability of social security exceeds annual GDP (Kuné, Petit, and Pinxt (1993)), a full transition to prefunded pensions without hurting any generation may be infeasible. However, as the Raffelhiischen simulation shows, the effects of a transition are not linear, that is, much can be achieved with a partial reduction of contributions to unfunded schemes and the establishment of pension funds. This cautious approach can also minimize distributional effects due to macroeconomic and financial uncertainty not captured in the above models.

Protection Against Inflation

The opposition to funded pension schemes in France results to some extent from the failure, during the 1930s and early 1940s, of several prefunded pension systems to provide adequate income to their members.30 This failure contributed to the choice, made in 1945, of financing the national social security system through the pay-as-you-go method. Nowadays, the increasing sophistication of financial markets and a lasting decrease in inflation may have reduced the magnitude of the inflationary threat, but it is useful to review how prefunded pensions in other countries have performed in this area.

The protection against inflation afforded by defined-benefit schemes varies according to the way sponsoring firms compute pensions. In most countries, defined-benefit schemes offer discretionary increases in benefits after retirement and some indexation before that.31 Notably, in the Netherlands negotiated increases are the norm, and in Germany indexation is mandatory. However, the ability to deliver inflation protection ultimately depends on the return of portfolios and on the existence of assets whose real returns are not systematically eroded by inflation (e.g., stocks and indexed bonds).32 Indications are that protection against inflation has been effective during the last 25 years, as in most developed countries the average return of pension funds’ portfolios seems to have exceeded inflation.33 In addition, it has also exceeded increases in real average earnings (wages) in several countries (Table 8).

Table 8.Real Returns of Pension Funds and Other Financial Assets and Rate of Growth of Average Earnings (1967–90)
Real Returns
Pension fundsGovernment bondsMarket paperEquitiesGrowth of Average Earnings
United Kingdom5.
United States2.
Source: Davis (1993).
Source: Davis (1993).

Returns depend on the distribution of assets held by pension funds, and this distribution varies from country to country depending on three factors: the extent of indexation of pensions (especially to wages), minimum funding requirements, and the supply of government bonds (Table 9). The first factor favors stocks, whose returns reflect the growth of the real economy. In countries where indexation is not common (as in Canada), the share of bonds tends to be higher.

Table 9.Asset Composition of Pension Funds’ Portfolios(As a percentage of assets, 1990)
Short-Term AssetsEquityGovernmentPrivateLoansProperty
United kingdom7631139
United States9462016
Source: National flow-of-funds data, in Davis (1993).
Source: National flow-of-funds data, in Davis (1993).

The second factor—high funding requirements—usually induces funds to invest in low-volatility assets, hence potentially decreasing the share of stocks and long-maturity bonds.34 In countries such as the United Kingdom, where required funding is limited, stocks are an attractive investment and do not require sophisticated investment strategies; in the United States, higher funding requirements may have discouraged a larger holding of stocks despite the availability of financial derivatives, which can hedge the portfolio against stock price falls.35 The final factor, supply of public debt, works in two ways to make bonds more attractive to pension funds: a higher supply will tend to increase the liquidity and the coupon rates paid by government paper. However, the effect on protection against inflation is not clear.

Several factors suggest that pension funds will continue to provide effective protection against inflation, even in the absence of indexed bonds. They include the proportion of stock holdings evidenced in Table 9; the current yield curve in Europe, which does not predict a surge in inflation; and financial innovations, such as portfolio “immunization” (i.e. the reduction of portfolios’ sensitivity to interest rate changes).36

Effects on Capital Markets

International experience shows that pension funds have fostered the development of capital markets mainly because of the scale of their operations, the nature of their liabilities, and their need and ability to use sophisticated financial instruments. It also suggests that these effects are strengthened if funds are under outside management; effects on aggregate saving and on firms’ governance are considered positive, but difficult to measure.

Because pension funds have long-term liabilities, they can provide longterm company financing; and because funds seek to be able to trade a large volume of assets without affecting prices, they create incentives for the expansion of active markets for bonds and stocks. Indeed, countries whose pension funds are large in aggregate tend to have stock exchanges with large capitalizations (Table 10) and generally low transactions costs (Hepp (1992)). Funds also create a demand for derivatives—used as a way to avoid having to trade illiquid assets (e.g., futures on an asset can be more liquid than the asset itself) or, as noted above, for hedging their portfolios against increases in interest rates.

Table 10.Pension Fund and Life Insurance Assets(As a percentage of GDP)
Pension Fund AssetsLife Insurance AssetsStock Exchange Capitalization1
United Kingdom17235526234297
United States17243524182463
Sources: National flow-of-funds data in Davis, (1993); COB; OECD Financial Statistics; and IMF staff estimates.

Domestic shares.

Includes foreign shares.

Sources: National flow-of-funds data in Davis, (1993); COB; OECD Financial Statistics; and IMF staff estimates.

Domestic shares.

Includes foreign shares.

The influence of pension funds on household saving is likely to be positive, but not one-for-one in relation to the increase of assets held by funds. The main studies in this area are Feldstein (1978), which shows that in the United States saving decreased when unfunded pensions were adopted, and Munnell (1976), which suggests that although holding private (funded) pensions decreases other personal saving, the introduction of funded pensions may increase the pool of capital.37 An illustration of the importance of pension funds as savings institutions in the United States, the United Kingdom, Canada, and the Netherlands is the size of their assets as a ratio to GDP (Table 10). In the United Kingdom and in the Netherlands, claims on pension funds also represent an important share of personal sector assets: about 40 and 30 percent respectively.

Pension funds can also play a role in privatization by providing a group of large and potentially stable investors.38 In France, the current level of privatizations (around F 50 billion a year) could easily be absorbed by pension funds, even if they were to invest the equivalent of only a fraction of contributions currently paid to mandatory supplementary schemes.

Pension funds, like other institutional investors, create incentives for bank disintermediation, that is, because funds create a large demand for securities, they may encourage firms to issue stock or bonds, instead of seeking loans from banks.39 This may increase the transparency of financial markets, but have only indirect effects on the financing of small enterprises.40 However, even if the cost of funds is higher for small firms than for large firms, a larger pool of savings and competition in financial markets should lower the absolute cost of funds so that the overall effect of pension funds could also be favorable to small firms.

The above discussion suggests that the effects of pension funds on French capital markets would be positive. These effects would of course depend on the regulations adopted and on the supervisory authority established. If a system based on defined-contribution schemes, where workers have the right to choose and change the manager of their savings, were adopted (see below), an industry along the lines of that currently managing mutual funds could be expected to develop (this industry includes “boutique” funds, but has a large participation of banks and insurance companies that benefit from their strong reputation).41 In this case, portfolio regulations paralleling those applied to insurance companies could be appropriate. In any case, a cap on self-investment and on investment in nonlisted securities would be desirable.42 The supervision of funds should be under one authority, to enhance protection against fraud and the development of financial disclosure guidelines. Therefore it would be useful to have clarified whether pension funds are to be considered part of the insurance industry, or under the control of the COB, which already regulates mutual funds.

The Choice Between Defined-Benefit and Defined-Contribution Schemes

Private pension funds in developed countries comprise mainly defined-benefit schemes, although defined-contribution pensions (e.g., personal pensions) are increasingly popular.43 The main advantages of defined-benefit pensions are that firms bear most of the risk of providing pensions (including, sometimes, inflationary risk), and that they offer the possibility of achieving some income redistribution; the main advantages of defined-contribution schemes are their portability and transparency. These advantages, as well as issues raised by defined-benefit schemes, are discussed in the Appendix. In the United States, most of the increase in coverage in the last 20 years has been achieved through the setting up of defined-contribution schemes, which nowadays cover as many workers as defined-benefit schemes. Personal pensions were introduced in the United States during the 1970s in the form of IRAs, and in the United Kingdom during the 1980s.

The analysis in the Appendix suggests that a system based on defined-contribution schemes may be easier to implement because such pensions are similar to savings instruments, which most workers are acquainted with, and because they are fully portable. In addition, it suggests that defined-contribution schemes (1) may provide more security to investors than defined-benefit schemes, whose pensions depend on the fate of the sponsoring company; (2) can be hedged against inflation, as long as there are no undue restrictions on pension funds’ investment in stocks; and (3) should be subjected to restrictions on early withdrawals in order to ensure a minimum amount of savings upon retirement (e.g., when changing jobs, pensioners should not be able to cash in their savings).

The Possibility of Contracting Out to Prefunded Schemes

Fostering private pension funds is one among different ways to accumulate enough savings to finance pensions after 2010.44 Another way would be to increase the current reserves of the supplementary pensions by setting aside contributions to accumulate assets.45 Irrespective of the way this accumulation is achieved, the extent of the adjustment required to balance the finances of the supplementary pensions will be large. Table 11 presents two extreme scenarios aimed at illustrating this fact.

Table 11.France: Scenarios for Balancing Supplementary Pension Schemes(Flows and stocks in constant prices)
Demographic indicators1
Birth year for generation retiring at 6019401950196019701980199020002010
Total population (in millions of persons)59.561.763.364.765.665.964.664.3
Population +60/population 20-600.230.260.300.340.360.360.370.37
Expected life after retirement at 60 (in years)2325262728293030
Scenario A: Pensions indexed to wages
Effective contribution rate (as a percentage of wages)9.1211.1013.5916.7518.5218.8819.2719.44
Average yearly change of value of point (in percent)
Ratios(in percent)
New pension/1990 pension113127143161182205231260
Contribution/1990 contribution137188259360449516594675
Wage/wage in 1990113127143161182205231260
Pv of benefits-to-contributions ratio214614112410587766966
Lifetime excess benefit of scheme as percentage of lifetime wage earnings34321-2-4-6-6
Scenario B: Stable contribution rates and funded component
Indicators for unfunded component Effective contribution rate (as a percentage of wages)7.507.507.507.507.507.507.507.50
Average yearly change of value of point (in percent)-1.6-1.0-1.5-
Ratios(In percent)
New pension/1990 pension85878176778594105
Average contribution/1990 average contribution113127143161182205231260
Wage/wage in 1990113127143161182205231260
Lifetime excess benefit of scheme as percentage of lifetime wage earnings3-1-2-2-3-3-3-3
Indicators for funded component
Rate of contribution (as a percentage of wages)
Interest rate (in percent a year)2.752.752.752.752.752.752.752.75
Reserves/pension payments42.67.311.013.013.313.213.013.0
Reserves/PAYG5 contributions2.67.612.516.719.620.621.020.8
Indicators for the aggregate scheme
Ratios(In percent)
New pension/1990 pension8597104115138151167186
Contribution/1990 contribution173195220247279314354399
Pv of benefits-to-contributions ratio2102100939087878585

Based on projections in Dinh (1994).

Pv (present value) of pensions/Pv of contributions, for the cohort retiring in the following five years.

(Pv of pensions-Pv of contributions)/Pv of wages, for the cohort retiring in the following five years.

Including the pay-as-you-go component.


Based on projections in Dinh (1994).

Pv (present value) of pensions/Pv of contributions, for the cohort retiring in the following five years.

(Pv of pensions-Pv of contributions)/Pv of wages, for the cohort retiring in the following five years.

Including the pay-as-you-go component.


The first scenario shows the effects of a policy of indexing pensions to actual wage increases. The second scenario illustrates the sharp decrease in benefits necessary to balance the schemes without increasing the contribution rate, as well as the scope of prefunding in cushioning part of the loss in benefits implied by such a policy. Both scenarios are simplified and do not do full justice to the complexity of the ARRCO system. But they can be seen as boundaries inside which the actual configuration of the schemes in France will probably evolve. Therefore they can be used as benchmarks to evaluate the magnitude of required changes. For instance, in the first scenario, contributions are almost tripled by 2060. Considering that the supplementary pensions are meant to provide only a fraction of total retirement income, it is obvious that a contribution rate around 20 percent of wages is unsustainable. The second scenario indicates the reduction in benefits necessary to keep contribution rates at the current level: pension benefits (at constant prices) would have to be reduced gradually by more than 20 percent in the next 40 years. Such a reduction would imply a halving of the pension/wage ratio. In this case, a funded component to the scheme, based on additional contributions amounting to 4 percent of wages, would permit an increase in the real value of pensions, owing to the accumulation of savings. These financial reserves would stabilize at 20 times the level of annual pay-as-you-go contributions, or about 150 percent of the value of annual wage income.46

The intergenerational transfers implied by the two alternatives outlined above are reflected in the evolution of the net value (to workers) of participating in a pension scheme. This value is measured by computing the net present value of contributions and benefits for each cohort of workers.47 In the first scenario, the net value of the scheme changes from being significantly positive in early years (4 percent of the present value of lifetime wage earnings) to being significantly negative in later decades (up to -6 percent of lifetime wage earnings), so that those persons already retired or retiring soon would benefit from a generous scheme while those retiring later in the future would not recoup the amount of contributions made. The net loss is so large for future generations, even after the population stabilizes, because those generations will be burdened with very high contributions to a system that will yield less than the discount rate.

The net balance of future generations improves when contribution rates are kept constant (they are on average equal to -3 percent of lifetime wage earnings), but that of current retirees and of those now in the middle of their careers worsens by the equivalent of about 4 percent of their lifetime wage earnings. Funded components do not change the absolute net loss of schemes, because these components are assumed to be actuarially fair. But they increase the relative overall return of those schemes, especially in the long run. For instance, even with relatively low prefunding, the average ratio of discounted benefits to contributions stabilizes at a higher rate in the second scenario than in the first. An additional welfare-improving implication of prefunded pensions and low contribution rates is that, compared with a high contribution rate scenario, they free more resources to households, which can use them in optimal ways. However, neither extreme scenario Pareto-dominates the other, and they do not give indications on how the accumulation of assets will be achieved, nor on how financial reserves should be managed.

Private pensions would decentralize the process of accumulation, and could be introduced by permitting a partial contracting out of the system in exchange for the freezing of the contribution rate. In this case, a decrease in future liabilities of the system—instead of a proportional increase of assets and liabilities—would be achieved, because workers would have part of their contributions deposited in defined-contribution funds of their choice, outside current schemes. This switch might increase (contractual) savings, which, as noted above, are usually not perceived as distortional taxes on labor and therefore might contribute to higher economic growth. However, as illustrated in the second scenario, by reducing the inflows into the pay-as-you-go system, it may require decreases in benefits for those already drawing pensions relative to a system without contracting out.

The hoarding of a fraction of labor income in savings instruments should be particularly attractive to younger workers, who would benefit most from the compounding of returns on financial assets. Chart 2 shows how pension benefits (in constant prices) would increase with the number of contribution years to defined-contribution funds, for a given real return, and to pay-as-you-go systems, for a given return ratio (the values are standardized relative to the pension earned after contributing ten years to a defined-contribution plan yielding 2.75 percent a year in real terms).48 It illustrates that even for the generous return ratio of 11.5 percent, longterm users of funded pensions may be better off.

Chart 2.France: Standardized Value of Pensions

(For similar patterns of contributions)

Source: IMF staff calculations.

The contracting-out scenario is one among several that could be adopted (including making prefunded pensions mandatory, as in Switzerland) and complements the gradual increase in the retirement age that will take place as a consequence of extending to 40 the number of years needed to qualify for a full social security pension. It has the advantage of not requiring increasing fiscal exemptions that would probably be necessary to make contributions to funded pensions attractive at a time of rising contributions to mandatory schemes. The approach would also keep the basic pay-as-you-go principle (répartition), focusing it on the basic social security pension, while spreading the cost of pensions across generations and favoring the development of capital markets.

IV. Conclusions

The growth of the life insurance sector and of company-sponsored savings plans suggests that individuals in France have both increasing confidence in financial markets and the desire to use them to guarantee future income, indicating a favorable environment for the expansion of prefunded pension schemes. The fact that French workers have increasingly included PEEs (which share features with defined-contribution schemes) in their collective agreements indicates a perception that such sources of deferred income can be part of labor compensation, which is a common feature in the United States and other countries that have funded pensions.

Their growth also suggests that financial markets are responsive to changes in demand and that the positive effects on capital markets associated with pension funds in other countries could take place in France. In the case of both PEEs and life insurance, despite the still relative dominance of a few large insurance and banking firms, a dynamic market has developed, suggesting that insurance companies are probably ready to manage large pension funds. In fact, since the deregulation in the 1980s, insurance companies have managed large financial portfolios not only for themselves but also for third parties. The size of the insurance market can be gauged by the amount paid in premiums, which corresponded in 1993 roughly to half of social security contributions.

According to the elements presented in this study, pension funds in other developed countries have had a positive effect on capital markets mainly due to the size of their operations, the information they require from firms, and the financial innovations they foster (to some extent, pension funds also tend to help to increase the household savings rate). The effect of institutional investors on French capital markets in recent years suggests that similar developments could be expected in France following an expansion of prefunded pension schemes for French workers.

The fact that most resources in FCPEs come from participatory benefits does not constitute a significant difference in relation to the financing of pension funds in other countries. Mandatory contributions to funds are present in some countries where the basic social security pension is small, and, in France, mandatory contributions could be tied to a sort of rebate (or freeze of contribution rates) linked to the relinquishing of future rights on supplementary pensions. Tax benefits applied to PEEs are not essentially different from those granted to most pension funds. But if these incentives were to be extended to pensions, it might be advisable to do so through a coherent system of taxation of capital gains to avoid unnecessary tax expenditure.

If the above indications are taken as pointing to the establishment in France of a pension system based also on private prefunded schemes, questions arise about the social cost of a transition and the guarantees offered by such systems. The preliminary answer to the first question is that a Pareto-improving transition, associated with increased labor supply and employment, can in principle be achieved. A possible way to initiate a transition would be to permit workers to contract out from the mandatory supplementary pension system, in order to participate in a defined-contribution plan.

The second question can also be partially answered by looking at the regulations adopted in other countries, as well as by drawing on current regulations covering the French life insurance industry and on the practice among PEEs. The experience in foreign countries, and the fact that regulations on French insurance companies have recently been liberalized, suggest that few restrictions on portfolio distribution would be desirable. Among them, however, would be limits on investment in stocks or debt issued by the company sponsoring the pension fund, and in nonquoted securities. In addition, the participation of workers in supervisory boards, their right to change fund managers, and broad disclosure rules would tend to protect the assets accumulated and to increase returns, hence increasing the trust of the public. International experience also suggests that defined-contribution schemes may provide more transparency and be easier to introduce. Because portability of benefits, as well as vesting (the right to receive some benefit after a minimum number of years of work), add some complexity to the management of defined-benefit schemes, these may be more difficult to implement nationwide. It also suggests that having a unified statutory authority responsible for supervising all pension schemes and regulations, possibly along the lines of that in place in the Netherlands, would increase confidence in the system and facilitate its introduction.

Appendix: Features of Defined-Benefit and Defined-Contribution Schemes

Defined-benefit schemes raise three issues: minimum funding levels, the ownership of surpluses, and portability.

Minimum funding levels are actuarial concepts that are applied diversely in different countries, all of them having the objective of guaranteeing that the firm has actually set aside enough resources to fund pension liabilities. Minimum funding levels can work toward protecting the value of pension funds in conjunction with, or as a substitute for, restrictions on portfolio composition: any time funding goes below the minimum, the fund has to be topped up; certain assets, especially those that are more volatile, cannot be held by funds. In the United States, the United Kingdom, and the Netherlands, there are mandatory minimum funding levels but no specific regulations on portfolio distributions; in Switzerland there are both (see Table 12 for features of funds in selected countries).49 The ownership of surpluses is generally given to the sponsoring firm, because it is understood that since the firm is committed to complement any shortfall of pensions, it should also be entitled to surpluses. In the United Kingdom and the United States the firm can either withdraw the surpluses (paying taxes) or adopt a contribution holiday. It can also assume the surpluses when a fund is liquidated.50 In some countries (e.g., the Netherlands), the policy is closer to that followed by insurance companies in France: surpluses are distributed to workers.

Table 12.Features of Pension Systems and Pension Funds in Selected Countries
United KingdomUnited StatesCanada
Social security
Retirement age (men/women)65/6065/6565/65
CoverageAll residentsAll workersAll residents
Replacement ratioFlat pension0.400.35
Supplementary pensions
FinancingFunded (private) PAYG (SERPS)1FundedFunded
(in percent)705545
Indexation after retirementDiscretionary, but common.Discretionary, but common.Not common.
Predominant typeDB2 (SERPS and occupational), DC (Private Pensions).DB (occupational), DC (IRA).DB
Fiscal incentives
Employer contributionsDeductibleDeductible4Deductible4
Worker contributionsDeductibleDeductible4Deductible4
Regulation of portfoliosPrudent man concept; 5% self-investment limit; concentration limit for DC schemes.Prudent man concept; 10% self-investment limit for DB plans.Prudent man concept; 7% limit on real estate, tax on foreign investment above 10%.
FundingObligatory only for contracted-out part.Higher insurance premia if underfunded (ABO).4Obligatory
VestingTwo years. Indexation of accumulated benefits.Five years. No indexation of accumulated benefits.Two years. Little indexation of accumulated benefits.
Social security
Retirement age (men/women)65/6565/6565/62
CoverageAll residentsAll workersAll workers
Replacement ratio0.350.600.60
Supplementary pensions
(in percent)505545
Indexation after retirementAlmost universal.MandatoryAlmost universal.
Predominant typeDBDBDC (60%), DB (40%)
Fiscal incentives
Employer contributionsDeductibleDeductible6Deductible
Worker contributionDeductibleDeductible6Deductible
Regulation of portfoliosPrudent man concept; 5% self-investment limit; 5% foreign investment for public pensions.Majority to be in listed bonds and loans to contributors.30% limit on domestic shares, 50% domestic real estate, 20% foreign currency assets, 10% foreign shares.
FundingObligatory for IBO7 or PBO.8IBO, with contributions adjusted every 5 years.Obligatory for IBO or PBO.
VestingOne year. Indexation of accrued benefits.Immediate, national scheme.Graded vesting between five and 30 years.
Sources: Davis (1993); and Pestieau (1991).

State earnings-related pension scheme.



Accumulated benefit obligations.

Taxed at lower rate.

Deductible up to a ceiling.

Indexed benefit obligations.

Projected benefit obligations.

Sources: Davis (1993); and Pestieau (1991).

State earnings-related pension scheme.



Accumulated benefit obligations.

Taxed at lower rate.

Deductible up to a ceiling.

Indexed benefit obligations.

Projected benefit obligations.

Portability of benefits (the transfer of benefits when leaving a firm) and vesting not only remove disincentives to labor mobility, but also protect workers against abuses. Portability is usually easier to achieve when benefits are based on average wages and not final wages, but in any case an element of arbitrariness is unavoidable (depending on the formulas used) and full neutrality cannot be guaranteed.51 These difficulties have often prevented the extension of coverage of defined-benefit schemes to workers with less stable careers. Regulations on vesting aim at preventing firms from evading the payment of pensions to workers who leave before retirement age, by, for instance, returning contributions without paying interest. Both in the United States and in most European countries, maximum periods before vesting have been imposed, varying from ten years in Germany to one year in the Netherlands (where a more homogeneous system exists).

The main concerns about defined-contribution schemes, and especially personal pensions, are the cost of managing these accounts, and doubts about the ability of workers to save enough and to choose appropriate savings instruments. The experience with personal pensions in the United States and United Kingdom is viewed among some institutional investors as evidence that workers tend to choose the wrong savings instrument—responding more to advertisements than to considerations about longterm returns (Makin (1993))—but evidence supporting such a judgment is scant and Diamond (1993), among others, has noted that freedom of choice is welfare improving. In addition, Lakonishok, Shleifer, and Vishny (1992) noted that returns from stock portfolios held by American definedbenefit pension funds are lower than those of the SP500 index, contrasting to returns of either mutual funds or accounts managed by insurance companies. The authors suggest that even if management costs are lower for defined-benefit schemes, agency problems may imply that total returns to employees are not higher than for defined-contribution schemes. Concerns about the amount saved by workers, which is often seen as too low, and the “excessive” risk borne by employees can be addressed by imposing minimum contribution levels, limiting early withdrawals, providing mandatory education about the savings instruments available to workers, and requiring full disclosure of characteristics of these instruments. Of these, limiting early withdrawals, as is done in the United Kingdom and was recently initiated in the United States (by imposing a 20 percent tax on withdrawals), is probably the most important measure.52


    ArtusPatrick“Bien-être, croissance et système de retraite,” in Epargneed. by PatrickArtusC.Bismut and D.Plihon (Paris: Presses Universitaires de France1993).

    AveryRobertGregoryElliehausen and GustafsonThomas“Pensions and Social Security in Household Portfolios: Evidence from the 1983 Survey of Consumer Finances,” in Savings and Capital Formation: The Policy Optionsed. by FrancisAdams and SusanWachter (Lexington, Massachusetts: Lexington Books1986) pp. 12760.

    BlanchetD.“Retraites, épargne et croissance,” in Epargneed. by PatrickArtusC.Bismut and D.Plihon (Paris: Presses Universitaires de France1993).

    BodieZvi and AliciaMunnellPensions and the Economy: Sources Uses and Limitations of Data (Philadelphia: University of Pennsylvania Press1992).

    BreyerFriedrich and MartinStraub“Welfare effects of unfunded pension systems when labor supply is endogenous,”Discussion Paper Series I No. 252 (Konstanz: Universitat Konstanz, Fakultat für Wirtschaftswissenschaften und StatistikJanuary1991).

    CazesSandrineThierryChauveauJacques LeCacheuxRahimLoufir“L’avenir des retraites dans un modèle d’équilibre général calculable,”Revue d’Economie FinancièreVol. 23 (Winter1992) pp. 10924.

    ChauveauThierry and RahimLoufir“L’avenir du régime de retraite français,”Revue EconomiqueVol. 45 (May1994) pp. 789804.

    Commission BancaireRapport pour l’année 1993 (Paris1994).

    DavisEric“The Structure, Regulation, and Performance of Pension Funds in Nine Industrial Countries,”Policy Research Working Paper No. 1229 (Washington: World BankDecember1993).

    DermineJean and Lars-HendrikRoller“Economies of Scale and Scope in French Mutual Funds,”Journal of Financial IntermediationVol. 2 (March1992) pp. 8393.

    DiamondPeter“Privatization of Social Security: Lessons from Chile,”NBER Working Paper No. 4510 (Cambridge, Massachusetts: National Bureau of Economic ResearchOctober1993).

    DiamondPeter and J.Hausman“Individual Savings Behavior,”paper prepared for the National Commission on Social Security1980.

    DinhQuang Chi“La population de la France à l’horizon 2050,”Economie et Statistique No. 274 (1994) pp. 732.

    FeldsteinMartin“Do Private Pensions Increase National Savings?”Journal of Public EconomicsVol. 10 (December1978) pp. 27793.

    France Ministry of the Budget“Intéressement, participation, épargne dans les organismes soumis au contrôle d’état,”Les notes bleues. No. 611 (Paris1992).

    France Ministry of the Budget“L’évolution de l’assurance-vie en France,”Les notes bleues de Bercy No. 24 (Paris1993).

    GaleWilliam and JohnScholz“IRAs and Household Saving,”American Economic ReviewVol. 84 (December1994) pp. 123360.

    HeppStefan“The investment behaviour of European pension funds: implications for Europe’s capital markets,” in The Future of Pensions in the European Communityed. by JorgenMortensen (London: Brassey’s1992) pp. 15170.

    HomburgStefan“The Efficiency of Unfunded Pension Schemes,”Journal of Institutional and Theoretical EconomicsVol. 146 (December1990) pp. 64047.

    HuangChi-fu and RobertLitzenbergerFoundations for Financial Economics (New York: Elsevier Science Publishers1988).

    KunéJ.W.Petit and A.Pinxt“The Hidden Liabilities of the Basic Pensions System in the Member States” (Louvain-la-Neuve, Belgium: Centre for European Policy Studies1993).

    LahidjiReza“Le financement de l’ economie francaise depuis 1980,”Problèmes Economiques No. 2371 (April131994) pp. 2227.

    LakonishokJosefAndreiShleifer and RobertVishny“The Structure and Performance of the Money Management Industry,”Brookings Papers on Economic Activity: Microeconomics (1992) pp. 33979.

    LefebvreFrancisMémento pratique-social (Paris: Editions Francis Lefebvre1993).

    LegendreR.“Evaluation empirique de quelques réformes de l’impôt sur le revenu,”Economie et PrévisionVol. 110 (1993).

    Livre Blanc sur les Retraites (Paris: La Documentation française1991).

    LynesTony“Paying for Pensions: The French Experience,”Suntory-Toyota International Centre for Economics (London: London School of Economics and Political Science1985).

    MakinClaire“When I’m 64,”Institutional InvestorVol. 18 (October1993) pp. 5259.

    MunnellAlicia“Private Pensions and Saving: New Evidence,”Journal of Political EconomyVol. 84 (October1976) pp. 101332.

    PastreOlivier and P.Moscovici“Epargne salariale et fonds propres,”Ministry of Industry and Trade (Paris1991).

    PestieauPierre“Les prestations de pension privées: leur répartition estelle juste?”presented at the Organization for Economic Cooperation and Development Conference des Experts Nationaux sur les Pensions Privées et la Politique Publique ParisJuly1991.

    RaffelhüschenBernd“Funding social security through Pareto-optimal conversion policies,”Journal of Economics (Zeitschrift fur National-ekonomie) Supplement No. 7 (1993) pp. 10531

    SzpiroDaniel“Points de repère sur le marché des actions des entreprises françaises,”Revued’Economie FinancièreVol. 20 (Spring1992).

    WillardJean-Charles“L’avenir de la retraite complémentaire des cadres: un point de vue technique,”Revue d’Economie FinancièreVol. 23 (Winter1992) pp. 195211.

    YoungH.“Niveaux et formes des allocations de pensions privées: dans quelle mesure sont-ils adéquats?”presented at the Organization for Economic Cooperation and Development Conference des Experts Nationaux sur les Pensions Privées et la Politique Publique Paris July1991.

    ZerahDov and Olivier-AlbanAucoinLe système financier français: dix ans de mutations (Paris: La Documentation française1993).

In some cases, voluntary supplementary schemes are also available. In some large companies (especially in the oil sector), such schemes are similar to funded, defined-benefit schemes available in the United States and the United Kingdom. Schemes for the public sector usually have only one level.

The ceiling wage for contributions to social security corresponded in June 1994 to an annual salary of F 152,160; ARRCO operates under the same ceiling. Contributions to AGIRC are computed based on wages between the social security ceiling and 4 times that ceiling for the so-called schedule B, and between 4 and 8 times that ceiling for (optional) schedule C.

Pensions for public servants depend on the length of the contribution period and on wages in the last few months before retirement. Contributions to public servant pensions correspond to about 40 percent of net salaries, with three quarters of them being financed by the state and the rest by employees.

The effective return ratio is defined as V/(p a), where V is the value of the point, p its price, and a is the call-up rate (taux d’appel), which is a surcharge on the price of the point. Not only has the price of the point increased more than its value—which has been loosely indexed to the CPI instead of to average real wages—but the call-up rate has also increased. A fall (in real terms) in the numerator is borne by retirees, and an increase in the denominator by the active population.

Simulations comprised eight scenarios, reflecting two hypotheses about fertility, labor force participation, and unemployment rates. They indicate that the dependency ratio (beneficiaries/ contributors) will increase sharply after the “baby boom” generation starts to retire in 2005.

These scenarios retained the demographic hypotheses described above, and a 2 percent rate of growth of real wages.

The Livre Blanc focused its attention on the basic system of social security (basic pension) leaving the regimes speciaux (including that for public servants) and the supplementary pensions in the background, in part because some in the first group will be absorbed by the basic system in the long run, and those in the second group are managed by autonomous bodies (ARRCO and AGIRC).

They comprise the indexation of pensions to the CPI instead of wage growth, the gradual increase in the contribution period for a full pension from 37.5 to 40 years (over the next 10 years) and in the period used for the computation of the average wage from the best 10 years to the best 25 years (over the next 15 years).

In 1992, for instance, households directed almost half of their net saving toward the purchase of life insurance and annuities.

See Zerah and Aucoin (1993) for an account of the liberalization process.

The popularity of both PEEs and life insurance reflects not only the fiscal advantages granted to these instruments, but also public trust in financial markets, and, in particular, in basic mechanisms of defined-contribution pension funds, including protection from inflation. As in the United States, life insurance policies in France can be liquidated during the lifetime of the policy holder, and hence are mainly viewed as a savings instrument, not necessarily related to a bequest motive.

Life insurance accounts for about 15 percent of household financial savings, permitting more than 40 percent of households to have some coverage and putting France among those EU countries, including the United Kingdom and the Netherlands, with the highest per capita premiums (France (1993)). Growth has been steady but strong, with annual contributions increasing 10-fold between 1980 and 1992, reaching almost 300 million francs at the later date.

The growth in life insurance and annuities has been shared by another long-term saving instrument, the PEP, which has accumulated deposits of about F 350 billion. This instrument replaced the retirement savings instrument PER (épargne de retraite), introduced in 1986, and benefits from tax advantages similar to those granted to life insurance.

The annual report of the French Bank Commission (Commission Bancaire (1994)) presents a detailed study of the activity of banks in the insurance market.

The tax deduction, corresponding to 25 percent of investments, is limited to F4,000 a year per household, with an additional F1,000 per child. In the case of death of the policy holder before 70 years of age, there are no transmission taxes. In the case of death after 70, a reduction of the taxable amount, up to F 200,000, is granted.

Of which F6 billion is related to life insurance.

Defined-benefit schemes are those where retirement income depends on years of work and/or salary, while defined-contribution schemes are those where pensions vary with the return on assets invested.

In 1986, the requirement was extended to firms with more than 50 employees, effective in 1991.

For details on holding restrictions and tax regimes, see Lefebvre (1993). Blocked accounts are usually used as collateral for investment loans to the firm and remunerated at market rates.

The main tax advantages of PEEs are the following: Employers’ voluntary contributions are not included in employees’ taxable income, and are not added to the wage bill for the purpose of computing social security contributions, if blocked for a period of time. In addition, they are deducted from the firm’s taxable income. Workers’ profit shares deposited in PEEs are not taxable for a period of five years, and investments made by PEEs maintain tax exemptions intrinsic to the financial instruments in which the investments are made. The tax deduction of contributions to PEEs and the tax exemption of reinvested returns are common to pension funds in many countries and reflect the benefits granted to mandatory contributions to pension schemes in France.

These payments can reach up to 10 percent of the wage bill of each firm (except in special cases, when they can reach up to 15 percent), and, at the individual level, up to half of the reference maximum wage for social security contributions. Because PEEs are not mandatory, only 33 percent of companies paying intéressement have one (52 percent in the case of state-controlled institutions and firms), most firms simply sponsoring an FCPE.

Employers’ contributions are limited to F10,000 per employee annually, except if they comprise stock issued by the firm itself, in which case they can reach up to F15,000. Voluntary worker deposits can amount to as much as one fourth of a worker’s gross wages. Workers can also participate in stock option programs and may sometimes be subject to mandatory contributions to PEEs, but stock options have experienced limited success and mandatory contributions are small, being limited to a maximum of F1,000 per year.

The type of savings instrument, management criteria, and restrictions on withdrawals are set out in the internal statutes of the PEEs, which have to be registered with the Stock Exchange Commission.

More than 50 financial institutions manage the 4,000-plus existing FCPEs, with the ten biggest institutions managing about 70 percent of the total funds. In the case of state-owned companies, around 80 percent of the resources in FCPEs are managed by professionals (Pastre and Moscovici (1991)).

Annual fees have been estimated to be around 0.5 percent of the value of portfolios managed by institutions, thus comparing favorably with the management cost of pension funds in regulated countries such as Germany and Japan, but less favorably with those in countries such as the United States (0.4 percent) and the United Kingdom (0.2 percent) (Davis (1993)).

See Huang and Litzenberger (1988) for the properties of the efficient frontier.

This will be the case when the combined rate of growth of the population and of the real wage exceeds the interest rate in steady state; that is, the capital/labor ratio is higher than the so-called golden rule level. See, for instance, Artus (1993).

For instance, the rationale for introducing the CSG—a tax used to finance the French social security system (and, in particular, pensions) and levied on all sorts of incomes—was to shift part of the burden from labor.

Homburg (1990) and Breyer and Straub (1991) prove the existence of transition paths from a steady state that are Pareto improving. This contrasts with results using models that do not consider the utility of leisure, such as Diamond (1993) and Blanchet (1993), where such a switch will always hurt at least one generation, because labor supply does not adjust to changes in incentives. Simulations in Raffelhüschen (1993) confirm the results in Breyer and Straub for the case of Germany. In this exercise, reductions in restrictions on remunerated work after retirement age help the labor supply to adjust to reductions in payroll taxes, leading to a Pareto improvement. Results in Cazes and others (1992) and Chauveau and Loufir (1994) also illustrate this point. The latter paper presents a simulation in which prefunded pensions lead to a decrease in output in the medium term and to unattractive results in the long term, while the former paper presents a much more positive outcome. The result in the latter paper is due in part to the way prefunded pensions are treated, that is, the assumption that contributions to prefunded pensions were invested in a “superfund” whose returns were not necessarily associated with individual contributions. Hence contributions are treated as an additional tax on labor, in contrast to the approach taken in the former paper, where contributions are treated as (contractual) savings.

The choice between financing through debt or additional taxes should be determined by a balance between the distortion induced by new taxes and the (interest) burden on future generations dictated by debt.

In particular, due to the depreciation of government debt caused by the Second World War.

See Table 12 for a comparison of indexation rules in selected countries.

Low returns could limit the effectiveness of the protection against inflation provided by defined-benefit schemes, because they might require those companies offering fully indexed pensions to inject additional resources (from, for instance, operational reserves), possibly impacting on the firm’s financial health and on its future existence. Funding requirements are discussed in more detail in the Appendix.

Of course, the protection afforded by funded pensions against inflation depends to a certain extent on macroeconomic conditions. A war or a continued deterioration of the economy would reflect on the assets held by pension funds and impact on future pensions. Diversification into international markets can, however, hedge pension funds’ portfolios against real shocks to the domestic economy.

Funding requirements may require a company sponsoring a defined-benefit scheme immediately to top up the fund with new resources whenever assets fall below projected liabilities. This automatic reaction is what discourages holding a high-volatility portfolio.

The proportion of pension funds’ assets invested in derivatives is still difficult to measure, because in most cases such investments are off-book items.

Immunization can be used as a hedge against inflation, because it protects the value of the portfolio against a fall in the price of long-term debt when nominal interest rates increase. This is achieved by reducing the maturity of the portfolio of bonds through the use of derivatives or short sales of longer-term bonds.

Results in Diamond and Hausman (1980) and Avery, Elliehausen, and Gustafson (1986) suggest that each dollar invested in pensions is associated with an increase in total saving of up to 40 cents (see Bodie and Munnell (1992)). Recent research on individual retirement accounts (IRAs) (Gale and Scholz (1994)), however, underscores the shifting of taxable forms of savings into sheltered instruments.

Pension funds were among the major players in the privatization process in the United Kingdom, Chile, and Brazil.

Securitization in France expanded throughout the 1980s. Lahidji (1994) reports that the reliance on banks for credit (taux d’intermédiation) decreased from 70 percent to 40 percent in 1991 and to less than 20 percent in 1992.

Institutional investors tend to require firms to publish more information, because they are usually less likely than banks to become directly involved in the management of firms. In general, more information should lead to better investment allocation. However, the dislike of institutional investors for less liquid assets puts small companies at a disadvantage in relation to large companies and may force them to continue to rely mainly on banking credit.

The concentration of the financial sector is stronger in France than in the United States. The top 15 money managers in the United States manage about 30 percent of the assets, with the top firm managing 4 percent of total assets (Lakonishok, Shleifer, and Vishny (1992)). In France, the top 15 financial institutions have a much more dominant position. For instance, the five largest institutions control just over half of the market of mutual funds (Dermine and Roller (1992)) and the five largest banking institutions have 68 percent of total deposits (Commission Bancaire (1994)). In addition, in France, several banks own part of the capital of insurance companies (in addition to their own insurance subsidiaries), and vice versa. The extent to which this concentration affects competition has not, however, been documented. For the case of mutual funds, Dermine and Roller (1992) do not find evidence that the market power of the five top banks affects the fees charged to manage mutual funds. They suggest, however, that because of the well-known fact that mutual funds are one among the many services (often) offered as a bundle, to analyze the return of mutual funds in isolation may not reveal the local monopoly power that could exist at the level of bank branches.

In France, cross ownership of (nonlisted) stock among firms is quite usual (Szpiro (1992)) and if pension funds (in particular, defined-benefit schemes) are not required to hold listed stock, firms could simply expand this practice, increasing the risk borne by eventual pensioners.

See the Appendix (Table 12) for a comparison of main features of pension funds in selected countries.

Another approach would be to rely on foreign savings between 2015 and 2055, which has the drawback that other European countries face the same demographic trend.

The current reserve, which corresponds to about one year of benefits, generates 5 percent of the income of the system, illustrating the scope for such an approach. Technically, the setting aside could be achieved without an increase in contributions, by adjusting the value of the point. Adjustments of contributions and benefits are not new to the system. During the 1950s, AGIRC lowered call-up rates when it wanted to decumulate an incipient reserve caused by the rapid increase in contributing participants (Lynes (1985)).

In a rough calculation, assets would be worth around the value of annual GDP. The second scenario illustrates a modest use of prefunded pensions, adopting a contribution rate of only 4 percent, which is relatively large in the context of supplementary schemes but would be too small if prefunding the general scheme were chosen.

The discount rate adopted is 2.75 percent a year, equal to the (conservative) real interest rate chosen.

The exercise in Chart 2 assumes that wages increase by 1.5 percent a year over the working life and that pensions are on average paid for 20 years—20 percent longer than they are currently paid. The yields adopted (2.75 and 3.75 percent a year in real terms) are conservative when compared with the returns of recent years. The return ratios reflect the evolution of ARRCO ratios, which have been reduced over time. Return ratios are not directly comparable with the yield of a bond (see Section I for a definition).

In the United Kingdom, mandatory funding covers only the relatively small guaranteed minimum pensions (GMPs).

In the United States, this led, during the 1980s, to several company takeovers aimed at liquidating funds with surpluses, until legislation curbed such a behavior by imposing penalties on companies liquidating their funds.

See Young (1991) for simulations for different working profiles.

In the United States, employers are not mandated to contribute, and, moreover, employees can cash in their personal pensions when they change jobs: before the adoption of the 20 percent tax, only 30 percent of the IRA accounts were rolled over after a job change.

    Other Resources Citing This Publication