You are here

POLICY BRIEF: Lifecycle Portfolio Choice and Pension Design

October 5, 2016

KEY POLICY INSIGHT: The foundation of better pension systems should be, in part, based on lifecycle portfolio choice models. These models provide clear guidance for policymakers on how to promote higher savings rates, foster appropriate investment risk-taking during working years, and encourage an alignment of institutional and individual discount rates.

Savings for a Secure Retirement

Despite a number of well-intentioned changes to pension policy, beneficiaries are still shortchanged. For example, in U.S. defined contribution (DC) plans, savings rates are not high enough to support a secure retirement. Similarly, inflows to U.S. defined benefit (DB) plans are not enough to provide sufficient retirement funds for beneficiaries.

The lifecycle portfolio choice framework helps identify the impact of alternative policies on savings choices that individuals make for themselves. Consequently, it provides insight on how to set savings rates for DC plans, the role of target date funds and annuities in individually designed retirement programs, and the tradeoffs between DC and DB plans. Finally, this modeling approach is useful for the development and provision of actual investment management solutions.

The simple models in this note suggest that current pension policy (at least in the U.S.) does not encourage enough savings. The consequences of insufficient savings are most obviously felt on individual pension beneficiaries. Moreover, insufficient savings also have public finance consequences for publicly funded DB plans.

Policy Levers and their Effects

Regardless of the type of pension system, retirement income is the result of saving (and therefore forgoing consumption) during working years. There are three key variables that drive savings decisions.

  • Desired level of retirement income
  • Rate of return on invested assets (i.e. savings)
  • Value of consumption in retirement relative to consumption during working years

The last variable is measured with an interest rate called the discount rate. The following simple example illustrates the interactions between these three variables.

Suppose an individual began working at age 25, retires at age 65, and lives to age 85. The 20 years of retirement are financed by capital accumulated and invested during the person's 40 working years. In this case, the annual income is $40,000 when the person begins working and is assumed to grow 2.4% each year during the working years. Retirement income is assumed to be 50% of the person's income in the final working year. To simplify discussion, it is assumed that each year's retirement income is consumed during that year. Finally, we will abstract from the effects of inflation.

Table showing the effects on savings rates of four alternative scenarios: base case, high returns, high discount, and low income.
Exhibit 1 - Effects on savings rates

Exhibit 1 shows the effects on savings rates of four alternative scenarios. In the base case, investment returns are assumed to be 3% per year, and retirement income is discounted at an annual rate of 2.15%. In this scenario, the savings rate is 18% per year. As the exhibit makes clear, the savings rate decreases when investment returns increase, discount rates increase, or the desired level of retirement income decreases.

While the example is simple, it illustrates the limited set of choices facing policymakers. If the metric "secure retirement" is defined in terms of accumulated wealth at retirement, then policymakers can choose which of the following mechanisms to influence.

  • How assets are priced (i.e. investment return)
  • How retirement consumption is priced relative to working life consumption (i.e. the discount rate)
  • How much income is set aside during the working period

Lifestyle Portfolio Choice

Lifecycle portfolio choice studies the decisions that individuals make over their lifetime (i.e. including working and retirement years) (1). This framework assumes that individuals make optimal choices. Thus, policy analysts can evaluate the impact of alternative policies on individual welfare.

In many respects, the example in the previous section is a simple version of a lifecycle model. Proper lifecycle portfolio choice models differ from this example in five important ways.

  1. Rather than assuming that wage growth, investment returns and discount rates are known with certainty, lifecycle portfolio choice models allow uncertainty. The net result is that there is risk associated with portfolio choices.
  2. Retirement income is assumed to be an outcome of a portfolio choice exercise, rather than a predetermined and fixed percentage.
  3. The allocation to risky assets can vary across time, as can the accumulation of assets during the working period.
  4. The savings rate can vary across time, rather than being a fixed percentage.
  5. Inflation, and uncertainty about inflation, can be explicitly introduced into the model.

An example of a policy that is well-suited for analysis with lifecycle portfolio choice models is whether savings rates and retirement income should be fixed percentages. Mandating savings rates (e.g. requiring all consumers to save a fixed percentage of income) carries with it a welfare loss (2). The first application of lifecycle portfolio choice models is to calculate the level of welfare loss.

Pricing Assets and Liabilities

Exhibit 1 raises a second set of issues that can be directly addressed with lifecycle portfolio choice models - the pricing of assets and valuation of retirement consumption. These issues are especially important in analyzing the benefits of defined benefit (DB) and defined contribution (DC) plans.

From the perspective of an individual investor, DB pension plans are simply repackaged DC plans, albeit with a number of choices made for the investor. To see this, take the perspective of a DC investor. Upon retirement, the investor faces two uncertainties - mortality date and annual income (3). Self management of assets exposes the investor to potential swings in annual income. Furthermore, because the investor's mortality date is uncertain, self-management of assets implies accumulating excess assets (as insurance against outliving assets). By contrast, an annuity guarantees a smooth path of income through the mortality date. A preference for one strategy versus another depends on how the annuity is priced (4).

In a DB plan, a number of choices are made for the investor. The DB plan sponsor explicitly chooses a path for savings, an investment portfolio and a retirement date. Upon retirement, the pension beneficiary receives a fixed annual income until the beneficiary's mortality date. Plan sponsors are able to offer this implicit annuity because they are pooling risks across investors.

Suppose that discount rates and savings rates are harmonized to what individuals would do for themselves. All else equal, the potential advantages of DB plans result from a broader investment opportunity set and (possibly) cheaper annuities. The role of public policy is to use frameworks such as lifecycle portfolio choice to design mechanisms that encourage sufficient savings. The effects of those mechanisms will be felt through their impact on investment returns and discount rates (i.e. pricing). An important part of pension policy design is to ensure that these mechanisms (a) rely on market pricing and (b) do not distort choices between DB and DC options.

Implications for Policymakers

A recent study by Wilshire Consulting suggested that the average actuarial return target (used to discount pension obligations) for U.S. public DB plans is 7.4%. By contrast, a recent paper by Vanguard advocated a 3% discount rate for their target date funds. Finally, a study by PIMCO recommended using the interest rates implied by a ladder of zero-coupon bonds for calculating retirement needs. As the earlier case illustrated, such differences in discount rate assumptions have consequences for savings rates that are most obviously felt on individual pension beneficiaries. Moreover, insufficient savings also have public finance consequences for publicly funded DB plans.

The foundation of better pension systems should, in part, be based on lifecycle portfolio choice models. These models provide clear guidance for policymakers on how to promote higher savings rates, foster appropriate investment risk-taking during working years, and encourage an alignment of institutional and individual discount rates.

See Time for a Fresh Look at Pension Design - Issue 1 of 3 and Pension Reform for Aggregate Welfare - Issue 3 of 3 for more information.

Footnotes

(1) More specifically, lifecycle portfolio choice shows the optimal decisions about consumption and savings that an individual should make over a lifetime.

(2) Lifecycle portfolio choice models assume that the investor chooses paths for consumption, savings, and the investment portfolio to maximize expected lifetime utility. The welfare change from any specific policy refers to the change in expected lifetime utility.

(3) Income is derived from the pool of assets accumulated during working life.

(4) The annuity discussed here solves an intragenerational risk sharing problem. As structured here it does not solve an intergenerational risk sharing problem.

References

Bansal, Ravi and Amir Yaron. 2004. Risks for the Long Run: A Potential Resolution of Asset Pricing Puzzle. Journal of Finance.

Cocco, J.F., F.J. Gomes and P.J. Maenhout. 2005. Consumption and Portfolio Choice over the Life Cycle. Review of Financial Studies 18 (2): 491-533.

Cochrane, John. 2001. Asset Pricing. Princeton University Press.

Merton, Robert. 1969. Lifetime Portfolio Selection under Uncertainty: The Continuous Time Case. Review of Economics and Statistics.

Policy briefs from the Heller-Hurwicz Economics Institute are intended to convey policy relevant research done by researchers associated with the Institute and do not represent the position of the Institute.