Working Paper 37: Investment Regulations and Defined Contribution Pensions - Executive Summary

The ongoing financial crisis is having a dramatic impact on individual accounts or defined contribution (DC) pension plans. Pension funds in countries with mandatory DC systems have experienced investment losses in 2008 as high as 20-25% (Antolin and Stewart, 2009; OECD, 2009). Larger losses have been experienced by the more aggressive portfolios with high equity exposures. A collapse in the value of their pension savings is of greatest concern for workers close to retirement as well as those already in the pay-out phase that have not shifted to conservative portfolios or bought life annuities.
The crisis has happened at a time of rapid expansion of DC pensions throughout the world. Such plans are even becoming part of mandatory retirement income systems in some countries. Despite their many advantages, DC systems subject retirement benefits to a great deal of uncertainty. Regulations can be designed so as to limit some of these risks and avoid situations where older workers and retirees are exposed to major losses on their retirement income. However, designing suitable investment regulations for DC plans is a complex task. It requires careful consideration of various factors. Following a quantitative assessment of different regulations using a stochastic model of replacement rates, this paper reaches the following conclusions:
The regulation of investment choice and default options in mandatory DC plans should be carefully designed
The weight of the DC portion in total retirement income should be a key deciding factor in the design of default investment strategies. In Chile and Mexico, where the mandatory DC pension is very large in relation to total income, the default fund for a worker ten years from retirement has a maximum 20% and 0% allocation to equities, for each country respectively. In contrast, some European countries with mandatory DC systems like Estonia, Hungary and the Slovak Republic have set the conservative portfolio (with no equities) as the default for all ages. Such portfolios may be inadequate for younger cohorts as they imply lower expected retirement benefits. In Australia, the default option for the mandatory DC pension system - which provides a large part of retirement income – is not regulated, and is often in practice a balanced fund with a large equity allocation (over 60% in some cases). In Sweden, where the mandatory DC system accounts for a very small part of total retirement income (contributions equal 2.5% of wages) the equity allocation of the default fund is even higher (around 90%). The default equity exposures in these countries should be carefully reviewed, taking into consideration the risk that pension benefits are exposed to.
In a retirement context, the risk-return trade-off of different investment portfolios and strategies should be evaluated by looking at projected retirement benefits
This paper departs from the traditional approach to assessing investment strategies that focus on short-term investment return and risk. In a retirement context, the objective is to maximise retirement benefits subject to a given risk level. Therefore, using a stochastic model to produce retirement income estimations, this paper maps the risk-return trade-off of different investment strategies using the replacement rate at the 5th percentile as the risk measure and the median replacement rate as the return measure.
The analysis corroborates that an investment portfolio may be mean-variance efficient in the short-term but inefficient when looked at through the lens of a pension plan member. Both very low allocations to equities (below 20%) and very high ones (above 80%) look unattractive in terms of the trade-off between replacement rate expectations and risk. In between, however, there is a wide range of options for plan members and regulators to consider.
Alternatives to conventional life-cycle investment strategies should be evaluated, especially when used as default options
More careful analysis is also needed on the design of suitable life-cycle investment strategies, especially when used as default option. Modelling results show that a naïve life-cycle investment strategy - reducing equity investment to zero over the last ten years before retirement - may not be the optimal investment strategy for an individual contributing regularly to a DC plan and intending to purchase an annuity at retirement. While life cycle strategies do indeed reduce retirement income risk they do so at the cost of lower pensions on average.
Various quantitative regulations can be established to limit retirement income risk in DC systems
Quantitative investment regulations can be used to restrict investment policies to those that provide a certain combination of potential retirement income and risks. Risk adverse regulators and supervisors will aim at policies that reduce the downside risk or that minimise the risk of unfavourable outcomes from DC plans. Such regulations come at the cost of renouncing potentially higher replacement rates that are attainable but at a higher risk of unfavourable retirement income outcomes. Less risk adverse regulators and supervisors, on the other hand, would aim at lower probability requirements as regard the downside risk (e.g. a security level of 80% instead of 95%), which will increase the range of investment policies available and thus the share of riskier assets.
Simple quantitative regulations such as a ceiling on risky asset classes have some advantages over risk-based regulations
Policymakers must also consider that regulations could be efficient a priori but inefficient a posteriori depending on whether real events fail to validate the modelling. They must also assess the complexity and cost of implementing and monitoring these different risk measures. Simple regulations (e.g. quantitative limit on equities) could achieve the same results than risk-based regulations (such as minimum returns, a Value at Risk (VaR) ceiling and a maximum replacement rate expected shortfall), but only in the case that the model is validated by real events. Risk-based regulations can also lead to pro-cyclical investment strategies, especially when applied over short periods.
The regulatory approach should vary depending on the length of the contribution and accumulation period and the type of benefit pay-out allowed
The impact of regulations minimising the risk of unfavourable retirement income outcomes through restrictions on investment risk depends on the length of the contribution and accumulation period. Long periods render possible investment policies with a larger share of riskier assets, increasing the potential for high replacement rates, but also risks. Short contribution periods combined with risk adverse regulators would steer people and pension funds towards conservative investment policies. The impact of different regulations is also likely to change once one considers alternatives to life annuities for the pay-out phase.
The design of DC investment regulations should take into account various country-specific factors
It is important to stress that there is not a single correct risk-retirement income trade-off to guide public policy decision. This trade off depends on the country context and on risk aversion levels. In countries where payments from DC pension plans are the main source of retirement income, the cost to the society of unfavourable outcomes are much larger than in countries where they have other sources of retirement income, such as public pension provision. Other factors such as incentives to achieve desired participation levels, cultural attitudes to financial risks and the nature of the pension promise also affect this trade-off between risk and retirement income. When participation in DC pension plan is mandatory concerns about risk may outweigh the desire to reach potentially high replacement rates.