Pension Reform in Sweden: Lessons for American Policymakers

Sweden has a long tradition of social insurance.¹ Indeed, it was the first nation to implement a mandatory government retirement system for all citizens.² Nevertheless, Sweden is now one of the world's leaders in the global shift to private pension systems.

The Swedish pension program faced financial problems similar to the troubles plaguing the U.S. Social Security system, and the Swedes decided that partial privatization was the best solution. For Americans, the Swedish example shows that pension reform and privatization are sound solutions to a pressing problem.

In Sweden, pension reform occurred because policymakers from both sides of the political spectrum realized that reform was the only way to ensure a safe and comfortable retirement income for today's workers. During the 1980s, Swedish lawmakers became increasingly aware that the country's National Basic Pension and National Supplementary Pension were not well adapted to meet future challenges. After reviewing the possible options for reform, including the recommendations of a committee appointed in 1992 to study the issue, Swedish policymakers decided that both individual workers and the overall economy would benefit if the old-age system were partially privatized.

The Swedish parliament, the Riksdag, reached an agreement to reform the system in 1994 and approved implementing legislation in 1998. The new system, which is fully effective for all workers born in 1954 or later, has four key features:

• Partial privatization. Workers must put 18.5 percent of their income aside for retirement, but they are now able to invest 2.5 percentage points of that amount in an individual account. Beginning later this year, workers will be able to choose the pension fund that best suits their investment preferences.

• Individual accounts. The remaining amount, 16 percentage points, is a tax that funds the pay-as-you-go government program. The Riksdag also altered this portion of the system dramatically. Instead of providing a pre-determined retirement benefit based on number of years in the workforce and earnings history, the new system provides benefits based on the amount of taxes paid during the worker's career. By creating a system of notional accounts (no money is actually deposited) and linking retirement benefits to lifetime income, the parliament has set the stage for further privatization.

• Safety net to protect the poor. As in other nations that have shifted to personal retirement accounts, a safety net will exist under the new Swedish pension system. The government will continue to guarantee a minimum pension funded by general tax revenues.

• Transition to protect retirees and older workers. Although the old pension system no longer exists, current retirees and older workers will continue to receive all or some of their retirement income based on the rules that existed before reform. Individuals born before 1938, for instance, will receive all the benefits promised by the old system. Retirement income for those born between 1938 and 1953 will come from a combination of the new and old systems.

Sweden's reformed pension system will yield substantial benefits for workers, retirees, and the Swedish economy. The key benefits include:

• Budgetary savings. Partial privatization, combined with reform of the government-run, pay-as-you-go portion of the retirement system, is expected to result in a fiscally sustainable system. Future expenditures will be significantly lower, protecting Swedes from higher taxes, higher spending, and large deficits.

• Higher retirement income. The ability to invest privately over a working lifetime will allow Swedish workers to benefit from compounding returns. The average blue-collar worker, for instance, should enjoy 40 percent more old-age income. Swedish retirees will have a safer and more comfortable retirement.

• Economic growth. By reducing the payroll tax rate and creating a direct link between lifetime income and pension benefits, Swedish pension reform will increase incentives to work. Moreover, the shift to a funded system will boost national savings, thus providing capital for future growth.

THE NEED FOR REFORM

All industrialized nations are confronted by the same challenge: Their tax-and-transfer pension programs face serious demographic and financial pressures. In effect, policymakers have only two choices. On the one hand, they can raise taxes and cut benefits in an effort to prolong the solvency of government-run old-age systems. Alternatively, they can give workers private retirement accounts. Faced with this choice, Sweden decided that reform was the best option.

Why did Sweden's lawmakers decide that they had to reform the Swedish old-age pension system? Among the key factors were:

• Long-term deficit. The unfunded pension system liability in 1996 was nearly $500 billion, roughly equal to 200 percent of Sweden's economic output.

• Demographics. In 1950, there were more than five working-age people for each person over age 65. Now there are fewer than four, and the ratio eventually will drop to three working-age people per retiree.³

• Future tax increases. Payroll tax rates already were high-nearly 20 percent-but they would have had to rise even more to keep the existing system in balance. According to Swedish government estimates, the tax rate would have needed to be as high as 36 percent by 2025.⁴

• Economic impact. The system's high tax rates discouraged output and employment. The adverse effect on work incentives was particularly severe because of a weak relationship between the taxes workers paid into the system and the benefits they expected to receive. In other words, workers had an incentive to work less and to underreport their income since increased pay simply meant higher taxes without an accompanying increase in future pension benefits.

• Fairness. Pension benefits for retirees often varied greatly. Because old-age benefits were based on a worker's 15 highest-earning years, two Swedes with equal lifetime incomes who had paid equal amounts of taxes could receive very different pension benefits. This formula discriminated in favor of higher-paid white-collar workers with peak-earning years at the expense of blue-collar workers with relatively stable earnings.

THE NEW PENSION SYSTEM⁵

Under the new Swedish old-age pension system, each worker's future pension will be based on the amount of money accumulated in two separate individual accounts. The bulk of retirement income-although this may change if further privatization occurs-will come from a notional account maintained by the government on behalf of the individual. A significant portion of retirement income, however, will come from a completely private individual account.

The overall contribution rate is 18.5 percent, which is slightly below the tax burden for the old system. The 18.5 percent is split equally between employer and employee, with each paying 9.25 percent on "pension basing income" into the system.⁶ This money is then divided, albeit unevenly, between the government-maintained account and the private account.

The Notional Individual Account

Of the individual's 18.5 percent payroll taxes, the larger part-16 percent of payroll-goes to the government portion of the program. As is the case with pay-as-you-go systems, the money is used to fund benefits for existing retirees.

What makes the government pay-as-you-go portion of the pension program unique, however, is the formula used for calculating an individual's future retirement benefits. Each worker's 16 percent payroll tax is credited to an individual account, although the accounts are notional. These imaginary accounts are credited each year with "earnings" based on Sweden's per capita wage growth.⁷ Beginning at age 61, a worker can retire, and the government uses the money in these notional accounts to calculate an annuity (annual retirement benefit) for the worker.

It is important to note that, because the calculations are based on life expectancy, the longer a worker stays in the workforce, the larger the annuity received. This reform is expected to discourage workers from retiring early, a common occurrence in Sweden that burdened the retirement system. Table 1 shows the pension penalty that workers face for retiring early, and the bonus they receive for retiring later in life.

The new pension system will adjust retirement benefits for inflation, although the adjustment may vary according to real wage growth.⁸ This provision creates a direct link between the growth of labor force income and the taxes paid on that income, and the benefits paid to retirees. This ensures that the system is financially sustainable.

Other interesting features of the notional portion of the Swedish system include the following:

• The notional accounts will be charged a modest administrative fee to cover the government's cost of running the program.

• Account balances are not transferable to heirs.

• Workers who leave the workforce to care for children are credited with contributions to their account.

• The full tax is imposed on all income up to about $30,000 (SEK 273,000). Above that amount, the government collects only the employer share of the tax. The revenue generated by this extra employer tax, however, is not credited to the worker's account.

• The notional accounts include more than just the taxes paid since pension reform was implemented in 1999. Instead, the government used historical data dating back to 1960 to construct the accounts.⁹

The Private Individual Account

Of the individual's 18.5 percent payroll taxes, the smaller share-2.5 percent of payroll-goes into an individual account.¹⁰ This is the "funded" part of the system, meaning that the money is invested in real assets that can then be sold to generate income during retirement. Upon retirement, the nest egg that accumulates in this account will be converted into an annuity.

The government has set up a centralized agency, the Premium Pension Agency (PPA), to oversee the record-keeping for this new system. Beginning later this year or early next year, as soon as the start-up phase is complete, workers will be able to invest their retirement monies in as many as five funds.¹¹ In the interim, the PPA is investing the money that has been collected in low-risk assets.

Interesting features of the funded portion of the Swedish system include the following:

• The system has been structured to discourage funds with high administrative fees.

• Workers can choose survivor's insurance that will provide income to heirs in the case of early death.

• The private pension funds must follow prudent principles and are largely free of intrusive regulations.

• Workers will be able to switch funds but will be charged an annual administrative fee to do so.

Simultaneous Reform of Private-Sector Pensions

Nearly all Swedish workers are also covered by an occupationally based private pension plan. Prior to the pension reform, these were defined benefit plans promising workers a modest pension based on earnings histories. With the exception of the pension plan for government workers, however, almost all of these plans are switching to a defined contribution format that is similar to the privatized portion of the government program. Under the new arrangement, workers can invest 2.0 percent-4.5 percent of their earnings in an individual account.¹² The combination of these occupational plans and the privatized portion of the government system means workers will be able to save 4.5 percent-7.0 percent of their income in personal accounts for retirement. This level of mandatory individual retirement savings is significant by world standards.

The Safety Net

Because of low wages and interrupted work history, some workers earn modest incomes during their careers. Many of these workers will reach retirement with small private retirement accounts and notional accounts. To ensure that pension reform does not adversely affect these people, the new pension system guarantees an adequate old-age income for all Swedes.

For a single retiree, the guaranteed pension is about $9,000 per year. A married couple, meanwhile, receives about $16,000. This means-tested benefit is phased out according to the income available from the notional account and the private account. The means-testing rules that govern the guaranteed pension are relatively generous, to the extent that about 40 percent of workers are projected to receive at least some income from the safety net program.¹³

Gradual Transition to the New System

The new system does not affect individuals born before 1938, and the old system does not affect individuals born after 1953. People born between 1938 and 1953 (inclusive), however, receive a mix of benefits, as shown in Table 2. A person born in 1948, for example, will receive 70 percent of his pension from the new system.

While the new system is financially self-sufficient in the long term, a transition period is necessary for smooth implementation. During this time, the government will finance benefits for current retirees and older workers about to retire using two sources of revenue. The major revenue source will be the 16 percent payroll tax. This tax will fund most of the pay-as-you-go obligations of the old system and all of the pay-as-you-go obligations of the new system.

The second source of revenue will be the assets in the trust fund. The trust fund currently comprises about 750 billion Kronor, a nest egg equal to nearly 40 percent of Sweden's gross domestic product (GDP). In years when benefit payments exceed tax revenues, a portion of these trust fund assets will be sold and the proceeds used to make up the difference. As shown in Chart 1, most of the assets in the trust fund represent real investments, although the trust fund also contains government bonds. These bonds mean that general tax revenues will pay for a portion of future benefits at the time when the bonds are redeemed.

BENEFITS OF THE NEW SYSTEM

Swedish pension reform is not as sweeping as the reforms that have taken place in such countries as Australia and Chile, but the changes are noteworthy and beneficial. The new system provides:

• Greater incentive to work. In the new system, pensions are determined by lifetime income, which means that each year of gainful employment will have a positive impact on future pension benefits. Because pension rights will be recorded in real and notional individual accounts, workers will have much less reason to hide, shelter, and underreport their income. The new system will also discourage workers from dropping out of the labor force.

• Increased national savings. Replacing a tax-and-transfer entitlement system with a partially funded pension system will increase national savings, particularly as the new system matures. Some recent empirical evidence from Swedish household sector data, for instance, indicates that reform will result in a net increase in savings.¹⁴

• Flexible retirement age. The new system neither penalizes nor rewards early retirement. Workers can retire as early as age 61 or stay in the workforce as long as they choose. Early retirement no longer burdens taxpayers since workers who choose to retire early do so in exchange for a smaller pension. Moreover, the new system does not penalize workers who remain in the workforce since they receive a larger pension or, if they so choose, earn income and collect a smaller pension at the same time. These benefits are possible because a worker's notional account becomes an annuity based on life expectancy at the time of retirement. Since the annuities do not reflect differences in life expectancies for men and women, however, women receive more from the new system than men.

• Lower taxes and less government spending. The new pension system will yield large fiscal benefits over time. The Swedish government calculated that the payroll tax rate necessary in order to perpetuate the old system would have reached 36 percent by 2025. This tax rate- and the level of government spending implied by such a tax burden-would have been an enormous weight on the Swedish economy. Even the 16 percent tax in the new system is too high, though the creation of notional accounts minimizes the adverse impact on labor supply. Chart 2 shows how reform will result in less government spending.

• Opportunity for more reform. In addition to partial privatization, the Swedish reform replaced a pay-as-you-go defined benefit (an annual old-age payment loosely determined by work history) with a defined contribution notional account. The system thus no longer implicitly redistributes income from one type of worker to another.¹⁵ Consequently, no demographic group will have a reason to oppose further privatization. Indeed, all demographic groups will have an incentive to support additional reform since private investments earn a higher return than the nominal interest that the government credits to the notional accounts. The only remaining obstacle to further privatization is the difficulty of handling the transition to private accounts, or how to match the short-term loss of tax revenue with the long-term budgetary savings.

• Fairness. The old system allowed some employees, particularly white-collar workers with limited periods of high earnings, to obtain disproportionately large pension benefits. No concept of fairness justifies this reverse income redistribution from the poor to the rich. The new system, in contrast, closely links pension benefits with lifetime earnings. The hidden redistribution of the old system has been replaced by a much more straightforward safety net program that will ensure that all retirees have adequate income.

• More retirement income. Last but not least, the new system allows workers to retire with more income. Table 3 shows the benefits that a Swede who began working in 1960 and retires in 2009 will receive.¹⁶ As the table shows, retirees will have substantially more income thanks to the new system. It is interesting to note that retirement income for blue-collar workers increases more rapidly than retirement income for white-collar workers. This change is due primarily to the fact that blue-collar workers are in the workforce longer and can take advantage of longer periods of compound returns-a benefit that did not exist in the old system.

THE NEXT STEP

Swedish pension reform has had broad support throughout the country. Indeed, it is so popular that most observers believe that it is only a matter of time before lawmakers increase the amount of money that Swedes can put in private accounts.

Such a change would further improve the Swedish pension system. Table 4 shows the pension benefits a worker who began working in 1960 and retired in 2009 will receive under the new system?and under a hypothetical completely privatized system. Income would be much greater in the fully funded system because of higher returns on investments. More specifically, the return on money in the notional account is linked to per capita wage growth, while the return on the money in the private retirement accounts should reflect the long-run return to capital investment.

Other reasons for further privatization also exist; in particular, additional reform would increase national savings and decrease future budgetary costs. Furthermore, pension reform may serve as a guide for reforming other government programs in Sweden. A system of education accounts exists within a few private companies in Sweden, and the legislature is currently considering a national program based on this model. Savings accounts for unemployment insurance are also under discussion, as are comprehensive individual savings accounts to replace the current system of social insurance.

In the spirit of privatization, the Confederation of Swedish Employers has presented a plan for discussion that combines a system of basic security accounts with a national flat tax.¹⁷ Such a system would sharply reduce tax rates, strengthen the economy, provide freedom of choice with respect to social insurance, and improve the welfare of most Swedes.¹⁸

LESSONS FOR THE UNITED STATES

Many nations, including Sweden, Australia, Chile, Great Britain, Mexico, Poland, Hong Kong, Argentina, Hungary, and Kazakhstan, have privatized their pension systems. From their experiences, Americans can learn many lessons.

First, Americans can learn that reform works. As the Swedish experience demonstrates, it is possible to change from a pay-as-you-go entitlement scheme to a system of personal accounts. In the case of Sweden, policymakers not only privatized a portion of the system, but dramatically reshaped the part of the pension program that remains under government control. As discussed above, this will facilitate further reform.

Second, reform is popular. Notwithstanding Sweden?s reputation as a cradle-to-grave welfare state, legislators from across the political spectrum were able to unite in support of the new system. These politicians wanted to create a more comfortable and more secure retirement system?and for them, a better pension system was not a partisan or ideological issue.

Swedish policymakers also deserve credit for the way in which they dealt with the question of the retirement age. Under the new program, workers can choose to retire at age 61 or can stay in the workforce until age 70 or beyond. Because retirees receive less retirement income when they retire early and more retirement income when they retire later in life, there are no adverse consequences for taxpayers.

Moreover, the new system does not force workers into an all-or-nothing retirement decision. Workers can continue working while receiving a pension. They can choose to receive a partial pension (25 percent, 50 percent, or 75 percent of the amount they would receive, depending on their age), a feature that is particularly attractive to workers who wish to withdraw gradually from the labor force.¹⁹ Once they choose full retirement, of course, workers receive a recalculated full pension according to their age and the size of their notional account.

Limitations of the Swedish Reform

While the Swedish experience confirms that personal retirement accounts are good public policy, Sweden has not made a perfect transition to the new system. Poor planning, for instance, has stalled Sweden?s shift to personal accounts. As a result, the individual accounts are not expected to be fully operational until later this year. The government holds the money that has been collected to date and is investing these funds in low-risk assets until the individual accounts are ready. At that time, the government will distribute all of the funds and accompanying earnings to the appropriate personal account.

Two reasons exist for the delay. First, the government agency in charge of implementing the new system, the Premium Pension Agency, fell behind schedule in designing the contract that governs the plan administrator. Second, the company that received the contract underestimated the complexity of the task and did not prepare the software on time.

While these glitches have delayed the full implementation of the Swedish reform, however, they certainly do not suggest that reform is either impractical or undesirable. Instead, Americans need to be aware that those charged with implementing a new system must be realistic about logistical and practical issues.

Furthermore, American legislators also should not permit the Social Security Trust Fund to purchase private-sector assets as the Swedish trust fund does. To be sure, the accumulation of real assets by the Swedish fund makes it easier for the government to finance promised benefits to older workers and current retirees. In the future, instead of collecting more taxes from workers to make up any difference between payroll tax revenues and old-age benefit payments, the government will be able, at least in part, to sell its stocks and private-sector bonds. In contrast, the U.S. Social Security Trust Fund holds nothing but government bonds?IOUs that the government can redeem only by collecting more money from tomorrow?s taxpayers.

Although the real assets in the Swedish trust fund are a benefit, these real assets are accompanied by risks. Politicians control how the money in the trust fund is invested. Countries that have government-controlled investment almost always succumb to the temptation to invest money for political purposes.²⁰ Sometimes politicians steer money toward industries with good political connections, and sometimes they steer it away from industries that are politically unpopular (tobacco, fatty foods, guns, etc.). Regardless of motivation, politically inspired investments harm workers by putting their retirement funds at risk and harm the economy by misallocating savings.

In any event, investing the Social Security Trust Fund is not relevant to the U.S. privatization debate. Even if the money could be protected from political mischief, accumulating real assets would make sense only if the underlying goal was privatization at some point in the future. Because of the looming retirement of the baby-boom generation, American policymakers need to reform the Social Security system as soon as possible. As a result, the current Social Security surpluses should be used to help ensure that the benefits of current retirees and older workers are fully protected during the transition to a system of personal retirement accounts. Indeed, the only lawmakers proposing to let the government invest the Social Security surplus are those that are opposed to reform.

Finally, U.S. lawmakers may not wish to copy Sweden?s approach to retirement annuities. All workers must annuitize their individual accounts, meaning that their nest eggs must be turned into specific annual payments. This makes sense for workers with modest accounts, particularly since one goal of the reform is to ensure that every retiree has an adequate level of income regardless of life expectancy. This approach, however, may be unnecessary for workers with larger accounts. Instead, it might be more appropriate to require these workers to "purchase" an annuity guaranteeing them a certain level of income but then give them flexibility over the disposition of the remaining funds in their accounts. This is particularly true for the privatized portion of the program.

A related problem is the existence of a government monopoly that calculates the annuities. Legislators opted to create this monopoly because they wanted to ensure that unisex life expectancies would be used when calculating the annuities, which discriminates in favor of longer-lived women.²¹ It appears that the government has achieved its goal, but the creation of a monopoly and the requirement that all retirees use one approach will necessarily inhibit the development of new products that may be attractive to consumers.

CONCLUSION

Sweden?s reputation as a cradle-to-grave welfare state would lead many Americans to assume that it would be one of the last nations to privatize its pension system. To the contrary, however, Sweden has made significant and profound progress toward ensuring the future of its retirement program.

In most nations, pension reform is not an ideological issue. The Labor Party, for instance, privatized the pension system in Australia. British privatization is most often associated with Margaret Thatcher, but she merely built upon the reforms that a Labour government had put in place.

Likewise, Swedish policymakers looked at their pension program?s long-term numbers and realized that a partially funded system, augmented by notional accounts for the government-run portion, was necessary. Sweden?s reform will be good for workers, good for taxpayers, and good for the Swedish economy.

Göran Normann, Ph.D., President of Normann Economics International, based in Stockholm and Paris, is an associate professor of economics at the University of Lund, Sweden. He has worked with the Federation of Swedish Industries and the Organisation for Economic Co-operation and Development (OECD). Daniel J. Mitchell, Ph.D., is McKenna Senior Fellow in Political Economy at The Heritage Foundation.

Endnotes

1. For more information, see Social Insurance in Sweden, 1999, available at www.rfv.se/english/pdf/bokeng.pdf.  Additional information on the Swedish pension system may be found at www.pension.nu , www.ppm.nu , www.saf.se,  or www.rfv.se.

2. Marten Palme and Ingemar Svensson, "Social Security, Occupational Pensions, and Retirement in Sweden," National Bureau of Economic Research Working Paper No. 6137, August 1997.

3. Ibid.

4. Sweden Ministry of Health and Social Affairs, The Pension Reform in Sweden Final Report, June 1998, at www.pension.gov.se/in%20English/final.pdf.   The International Monetary Fund projected that the tax rate would need to climb "only" to 26 percent, largely because it assumed the economy would grow much faster?2 percent annually as compared to the Swedish government?s 1 percent growth estimate. See Sweden: Selected Issues, IMF Staff Country Report No. 98/124, November 1998.

5. In addition to the government programs, most Swedish workers have some type of employer-provided pension. The employer-provided pension funds for private-sector workers are pre-funded, while the supplementary pension schemes for public-sector employees are largely unfunded. Many workers also have some level of personal retirement savings. About 75 percent of old-age income, however, comes from the government. See Edward Whitehouse, "The Tax Treatment of Funded Pensions," World Bank Social Protection Discussion Paper No. 9910, April 1999.

6. Labor economists widely agree that workers bear the total burden of both payroll taxes and mandatory savings. Politicians often require employers to pay some or all of such levies, but the funds inevitably reduce total employee compensation.

7. Annika Sunden, "How Will Sweden?s New Pension System Work?" Boston College Center for Retirement Research Issue in Brief No. 3, March 2000.

8. The adjustments for inflation may be greater or lesser than the inflation rate, depending on whether real wage growth is higher or lower than 1.6 percent.

9. Edward Palmer, "The Swedish Pension Reform Model: Framework and Issues," World Bank Social Protection Discussion Paper, June 2000, available at http://wbln0018.worldbank.org/
HDNet/HDdocs.nsf/2d5135ecbf351de6852566a90069b8b6/
5b192d2792690aa5852568ef00626c7a/$FILE/Swedish.pdf   (accessed June 22, 2000).

10. Between 1995 and 1998, 2 percent of each worker?s pensionable income was set aside and invested by the government. This money will be transferred to the fund(s) selected by the worker.

11. Workers that do not choose a fund will have their money invested, by default, in a government fund that will be weighted toward low-risk, low-return assets.

12. Palmer, "The Swedish Pension Reform Model: Framework and Issues."

13. Sunden, "How Will Sweden?s New Pension System Work?"

14. Lennart Berg, Sparandets guldålder (Stockholm: Merita Nordbanken, 2000).

15. Instead, income transfers will be handled in an above-board fashion through the safety net guarantee pension.

16. It is assumed that workers invested the private pension money equally in bonds and stocks in the premium pension system. The rates of return used in this example are the actual rates of return between 1960 and 1998. After 1998, the rates of return are assumed to be 1 percentage point below the average for the 1960?1998 period.

17. Göran Normann et al., Matching Flat Tax with Basic Security Accounts (Stockholm: Swedish Employers Confederation, 1998).

18. For an analysis of the steps in the transition process to this visionary system, see Göran Normann, "From Vision to Reality: How to Implement a Flat Tax and Basic Security Accounts," Swedish Employers Confederation Discussion Paper, forthcoming.

19. Palmer, "The Swedish Pension Reform Model: Framework and Issues."

20. For a more complete analysis illustrating how government-run pension funds around the world and state employee pension funds in the United States have been used for political purposes, see Daniel J. Mitchell, "Why Government-Controlled Investment Would Undermine Retirement Security," Heritage Foundation Backgrounder No. 1248, February 5, 1999.

21. Palmer, "The Swedish Pension Reform Model: Framework and Issues."