Close Close
ThinkAdvisor

Retirement Planning > Spending in Retirement > Income Planning

Into the Unknown: Rethinking Return Assumptions for Financial Plans

X
Your article was successfully shared with the contacts you provided.

What You Need to Know

  • Asset classes and investments with more attractive capital market assumptions should receive higher portfolio weights than those with less attractive CMAs.
  • Historical long-term average returns can provide an unreasonable estimate of expected returns for investors, especially in fixed income.
  • Nearer-term estimates provide a useful context for more immediate investment expectations, they are not necessarily appropriate for an investor with a longer-term investment horizon.

Future market returns are uncertain, but estimating expected returns is an essential planning exercise for most investors. Here I will explore some new research that digs relatively deeply into this topic.

Return forecasts, commonly called capital market assumptions, impact investors in a variety of ways. The most obvious implication has to do with optimal portfolio design. Asset classes and investments with more attractive CMAs should receive higher portfolio weights than those with less attractive CMAs.

CMAs can (and should) also affect other financial decisions, such as how much a household has to save for retirement, can spend in retirement, etc. (i.e., a financial plan). Financial plans for households can exceed 50 years in length. For example, a 30-year-old might plan on retiring at age 65 and then having a retirement that would last until age 95, suggesting a 65-year planning period.

The returns used in this financial plan should be the expected returns for investors and can vary as expectations evolve into the future.

The Breakdown

While historical long-term average returns are commonly used by financial advisors, this approach can provide an unreasonable estimate of expected returns for investors today, especially for fixed income investors.

For example, the average yield on 10-year U.S. government bonds from January 1870 to July 2021 has been 4.5%, based on data from Robert Shiller’s website. This is significantly higher than current yields, which are closer to 1.7% as of Oct. 21.

Any type of plan using historical long-term averages is likely to paint an overly optimistic picture for investors, resulting in lower required savings rates and higher available safe withdrawals, especially for investors with shorter time horizons (e.g. retirees).

While it’s possible the gap between current and historical yields will narrow in the future, a financial plan whose return assumptions are based purely on long-term averages would imply an investor owning 10-year U.S. government bonds (before fees, taxes and inflation) today could effectively earn a 4.5% return. In reality the actual expected return, based on today’s yields, is likely to be considerably lower.

More on this topic

A number of asset managers create CMAs today, but most of their forecasts tend to be relatively near-term estimates covering roughly the next five to 20 years. While these nearer-term estimates provide a useful context for more immediate investment expectations, they are not necessarily appropriate for an investor with a longer-term investment horizon, such as a financial plan for a middle-aged family through retirement which, again, can exceed 50 years.

Another Approach

Therefore, one approach to consider would be to use two (or more) sets of CMAs in a financial plan: a near-term forecast (e.g., for the next 10 years) that incorporates the current market environment (i.e., is conditional in nature) and a longer-term forecast that is more unconditional in nature, focused on return expectations in equilibrium (i.e., for the long term).

For those advisors who aren’t interested in estimating their own CMAs, PGIM Quantitative Solutions releases updated estimates quarterly, which are available free online.

Two final points when thinking about return assumptions in a financial plan.

First, many CMAs are geometric returns. Geometric returns are the expected realized returns after considering portfolio volatility and are not a suitable input into any type of Monte Carlo planning tool.

The input should be the expected arithmetic average and the realized return will be the geometric average. Expected geometric averages should be the assumption in any type of time value of money (i.e., deterministic forecast).

Second, all return assumptions should incorporate expected fees and expenses, as well as potential investment alpha, and should be documented accordingly.

It’s impossible to know what the future holds for investors, but it’s your responsibility as a financial advisor to ensure your assumptions are reasonable. Good luck!


David Blanchett is managing director and head of retirement research at PGIM DC Solutions.