This mechanism automatically reduces benefits in times of deficit, and thus avoids what in Sweden is referred to as “solvency deficit”. The mechanism automatically restores indexation as and when the solvency of the plan allows. After eight years of real increases in pensions, this mechanism was triggered for the first time in 2010, reducing benefits and notional pension capital. Despite some modifications, the core policy response has been to retain the basic rules of the system, in particular the automatic balance mechanism, to ensure the financial stability of the pensions system.

As a result, many pensioners have seen their benefits reduced, but the impact has been mitigated by policy responses – a change in the accounting rules for indexation which meant a smaller reduction in 2009 (though to be higher in subsequent years) and tax cuts for retirees. A minimum guaranteed pension for low-income retirees has provided a buffer for low-income pensioners.

This article investigates how the automatic balance mechanism may both reflect and foster a policy-making environment conducive to prioritizing the longer-term goal of financial stability over more immediate concerns with maintaining benefit levels. Clearly, the choice of an automatic balance mechanism implies trade-offs. The Swedish experience highlights the idea that accepting the “cost” of benefit reductions may offer important advantages, including financial sustainability and a degree of political insulation.

Annual indexation of notional accounts and pensions

The Swedish public pension system consists of two different earnings-related benefits: the Inkomstpension , on which this article focuses; and the fully-funded Premium Pension . A tax-financed Guaranteed Pension also provides supplemental support for retirees with low Inkomstpensions .

The Inkomstpension is a pay-as-you-go pension plan with a substantial buffer fund. Because the size of the Inkomstpension benefit is a function of life-time contributions paid, the notional return (or interest rate) on those contributions and life-expectancy, it is often referred to as a notional defined contribution (NDC) plan. Here, the “interest” payments are referred to as indexation.

The basis for calculating indexation for both accounts and benefits is the increase in average income. When the initial pension is calculated, that is, when the notional account value is converted to an annuity, the pension is increased based on an assumed annual growth rate for the income index of 1.6 per cent. This advance interest rate, which raises the initial pension, is then deducted every year from the growth of the index. Thus, Inkomstpensions will be indexed annually by the change in income index minus 1.6 percentage points.

However, under current law (2), certain situations constitute exceptions to the regular income indexation of accounts and benefits. These exceptions are governed by an automatic balance mechanism, based on a solvency ratio of assets to liabilities (see Table 1) (3). The balance mechanism is activated if and when the accounting statement of the Inkomstpension system shows that liabilities exceed assets, at which point the deficit is automatically eliminated by reducing the indexation of both benefits and notional pension accounts of workers. Any subsequent surplus is used to gradually restore the indexation to its prior unreduced level (see Figure 1). This mechanism is intended to ensure an automatically financially stable pension system, in a sense fulfilling the “defined contribution” character of what is essentially a pay-as-you-go system.

Guaranteed pension

For those with very low Inkomstpensions , Sweden provides a minimum guaranteed pension. The Guaranteed Pension is financed by general taxes as a part of the state budget, not by contributions. The size of the minimum pension depends on marital status and years of residence, with reductions for less than 40 years of residence. The guarantee benefit is formulated as a top-up, which tapers off for increasing levels of the Inkomstpension and is annually indexed to the Consumer Price Index. In 2009, 42 per cent of all retirees, and 26 per cent of “new” retirees, had some amount of guarantee benefit.

As long as a retiree has a low enough Inkomstpension to qualify for the guarantee top-up, an increase or decrease in the Inkomstpension , whether due to normal indexation or balancing, is either completely or partially offset by a concomitant increase or decrease in the guarantee amount.

Policy responses to negative Inkomstpension indexation in 2010

The automatic balance mechanism for Inkomstpension was triggered by the solvency deficit as of 31 December 2008, marking the first time the balance ratio fell below unity since the introduction of the new system. The value of total assets was 3.28 per cent less than the value of liabilities (balance ratio of 0.9672).

This deficit, at least in part, reflected the impact of the global financial crisis. The buffer fund suffered heavy losses in 2008, posting a negative return of 21.6 per cent. While the impact on the buffer fund was significant, it was by no means the only reason assets fell short of liabilities. The shortfall can also be traced to (artificially induced) excessive indexation of accounts and benefits – due in large part to the smoothed and lagged structure of the income index – as well as to a change in the tax code that increased registered , but not real, average pensionable incomes.

According to the rules governing the balance mechanism, the ratio as of 31 December 2008 should have been applied to reduce the indexation of accounts and benefits for 2009/2010. However, in light of the buffer fund losses, policy-makers questioned whether it is wise to allow the value of the buffer fund on a single occasion, the last trading day every year, to affect the indexation of accounts and benefits over the following year. Alternatives calculation methods, including using an average of the fund value during the year or over several years, had indeed been discussed in the original bill in which the automatic balance mechanism was proposed. In 2009, the government, supported by the pensions group (4), proposed that the balance ratio be calculated instead based on the average of the values on 31 December of the last three years. The proposed legislative change would, in effect, reduce the size of the indexation reduction due to the loss in buffer fund assets, with the drawback of making those reductions continue for a longer period.

The change was enacted despite expert criticism (5). Consequently, the balance ratio increased from 0.9672 to 0.9826, reducing the balancing effect in 2010 from 3.28 to 1.74 percentage points. Adjusting for an income index growth rate of only 0.3 per cent in 2010, and the advance interest rate of 1.6 percentage point, the final indexation ofInkomstpension benefits was minus 3.0 per cent in 2010 (6), significantly lower than the anticipated 4.5 per cent had the legislated change not been introduced. Likewise, because the CPI fell by 0.8 per cent, the indexation of the Guaranteed Pension was also nominally negative in 2010.

A second policy response was to reduce taxes for retirees aged 65 and older. The tax cut more than compensated those with the lowest Inkomstpensions , and consequently a large Guaranteed Pension, for the pension reduction. Retirees with higherInkomstpensions and little or no Guaranteed Pension still suffered a loss in pension income net of taxes. Pension income after tax dropped by approximately 1 per cent for an average retiree, while the loss was greater for retirees with higher pensions.

From the standpoint of state revenues, the tax cut has reduced the annual tax revenue by an estimated 3.5 billion SEK (7). However, the negative indexation of theInkomstpension reduced annual benefits by some 6 billion SEK, which in turn triggered increased Guaranteed Pensions and housing allowances at an estimated cost of 0.3 billion SEK.

Policy responses to continued negative Inkomstpension indexation in 2011

The 2008 financial crisis ushered in a recession that slowed earnings and employment growth through 2009. That same year, contributions to the pension system fell 0.2 per cent. This development, coupled with the decision to smooth the buffer fund value, which in turn postponed part of the elimination of the deficit incurred in 2008, put further strain on the system’s solvency. Indeed, the shortfall of assets relative to liabilities at the end of 2009 increased from 7.6 to 10.4 per cent of GDP, where the balance ratio (0.9549) now showed a deficit of 4.5 per cent of liabilities. As mentioned, a large share of this substantial deficit can be attributed the previous excessive indexation of accounts and benefits due to the above mentioned lag structure of the income index (8). Nevertheless, the buffer fund’s high rate of return of 19.3 per cent in 2009, a strong recovery from the previous year’s losses, helped to avoid an even larger deficit.

In late July 2010 the income index for the following year was finalized, showing an increase of 1.9 per cent. With a balance mechanism reduction of 4.5 percentage points, the account balances will be indexed at minus 2.7 per cent, while Inkomstpensions will be indexed at minus 4.3 per cent (2.7 per cent – 1.6 per cent). Guaranteed Pensions limits will increase by 0.9 per cent, reflecting an equivalent increase in the CPI over the period from June 2009 – June 2010.

How the Swedish Parliament will respond remains to be seen, although tax reductions, especially for retirees, are clearly on the agenda. In May the Government submitted a bill proposing a further reduction of taxes for retirees, and both the Government Alliance and the opposition have pledged in the current general election campaign to reduce taxes to compensate retirees for the losses to their Inkomstpension .

Lessons and Conclusions

While any conclusions are necessarily tentative, the Swedish experience offers several preliminary lessons for countries considering embarking on a similar course.

First, choosing an automatic balance mechanism inevitably implies some degree of benefit reductions in adverse times. When those times come, the challenge is to “stick to the rules” – to resist potentially strong public and political pressures to deviate. In the Swedish case, it is clear that the financial and economic crisis has had a significant and negative impact on the Inkomstpension system and especially on the insured in this plan, potentially undermining public confidence in the system. In the face of benefit cuts, the core policy response in the pension legislation has nevertheless been to retain the basic rules of the system, in particular the automatic balance mechanism, to ensure the financial stability of the pension system. Still, the crisis, as the first real test of the balance mechanism, presented an opportunity to evaluate the mechanism’s more technical features, resulting in a minor modification of the buffer fund value.

Second, in Sweden, pressures to veer off course were undoubtedly minimized by the minimum guaranteed pension and tax reductions for pensioners, suggesting that automatic balance mechanisms may be more resilient when additional support is available. A well-designed minimum guaranteed pension protected the poorest retirees from the brunt of the crisis, while tax reductions, some of which have already been enacted and others proposed for next year, offer another source of relief. Taken together, these measures have indeed mitigated the effect of reductions of Inkomstpensions . However, it should be noted that political pressure behind tax reductions stemmed not only from the automatic reductions of the Inkomstpension , but from retirees’ demands for equal taxation of work income and pensions. Since 2006, workers have received larger tax reductions than retirees, and even with the proposed additional tax reductions for retirees, retirees will bear a heavier tax burden than workers. Nonetheless, since benefit reductions have been somewhat offset by tax reductions, whether the Swedish pension plan has truly stood the test of managing benefit reductions in adverse times is still somewhat open to question.

Third, while the benefit reductions incurred in Sweden will likely have a negative impact on the public’s trust of their pension system, so would, arguably, unattended deficits and perpetual pension reform debates. In accepting automatic and rapid adjustments (reductions) of present and future benefits, the pain is less likely to be prolonged beyond next year. Moreover, the automatic financial stability of the Swedish system largely shields politicians from having to introduce what many perceive to be inevitable benefit cuts in hard times. Sweden has been able to skirt some of the political battles facing recession-stricken defined benefit and “notional” schemes, which must attend to the increased need for benefit reductions and or contribution/tax increases. Without the predefined rules, these decisions would not only be significantly more difficult, but as a result would consume more valuable and scarce political time and energy.

Finally, even given the tax reductions and the smoothing of the buffer fund value, it is nevertheless striking that two consecutive years of significant pension reductions were accepted, in light of the context: an election year, basically strong government finances, a large buffer fund and moderately optimistic long-term pension plan projections. In contrast to many other policy-making contexts, the broad political agreement and relatively large societal agreement on pensions in Sweden, coupled with the logic of the Inkomstpension system and particularly its accounting method and presentation, hopefully creates an environment in which benefit reductions for those with pensions above the guarantee level are widely considered an acceptable response to financial and economic constraints. Needless to say, what works in Sweden may not in other contexts.

Text contributed by Ole Settergren, Director of Research and Development, Swedish Pensions Agency.

(1) This article presents only limited contextual information. For such information, please refer to the Orange Report – The Annual Report of the Swedish Pension System 2009.The report can be down loaded from http://www.pensionsmyndigheten.se/Publications_en.htlm, especially pages 4–9 and 38–40 will facilitate understanding of this text.

(2) Parliament introduced the automatic balance mechanism principle in 1998 alongside the income indexation law, although the balance mechanism did not officially become law until 2001.The reason for this time lag was that the research and details of the balance mechanism had not been completed in 1998.

(3) System liabilities are the account values of all participants, workers and pensioners. The system assets consist of the value of the buffer fund and a calculation of the present value of the Inkomstpension system’s claim on 16 per cent of the future contribution base, called the “contribution asset”.

(4) The pensions group is the permanent committee consisting of legislators representing parties supporting the pension reform. The pension group is thus backed by some 85 per cent of the members of Parliament. It is chaired by the Minister of Social Insurance and entrusted with the task of “managing the pension reform agreement”.

(5) The change was criticized by a majority of those institutions given the chance to comment on the proposal. There were more drawbacks to this change than the mentioned – including the existence of alternative, more efficient, smoothing mechanisms.

(6) It is somewhat misleading to say “minus” since indexation is calculated as [Income/balance index (t) / Income/balance index (t-1)] / 1.016.

(7) Promemorian Ytterligare sänkt skatt för pensionärerna, Finansdepartementet (2009-09-15).

(8) The Swedish Pensions Agency proposed that the Government investigate the lag structure of the income index, since it causes financial strain – if necessary, resolved by the balance mechanism – when there is a shift from high earnings growth to low earnings growth.