Retirement Planning in the Post-4% World

Today’s low bond yields and high equity valuations have led many to jettison the traditional 4% initial safe-withdrawal rate (SWR) assumption. But I will show that the optimal “safe” withdrawal rate depends considerably on the retiree.

Yields on 10-year U.S. government bonds are close to 1%, which is well below their long-term historical average yield of closer to 5%.

Lower assumed returns will reduce initial withdrawal rates (and increase required savings), ceteris paribus, and a growing body of research has suggested that 4% may no longer be a “safe” initial withdrawal rate for retirees.

While it’s definitely important to use expected returns in projections (don’t be lazy and use the historical long-term averages), it’s also critical to understand how adjusting other key assumptions will impact the results. For example: retirement doesn’t always last exactly 30 years, retirees can adjust withdrawals as needed, many retirees don’t increase spending annually with inflation, and common metrics (like the probability of success) don’t often accurately portray the impact of a bad outcome.

If I update returns to reflect today’s investing environment, then 3% (or even 2%) becomes the “new 4%”; however, depending on other assumptions and the preferences of the retiree, a 5% initial withdrawal rate (or higher) could work as well.

I acknowledge that 4% is a good starting point with respect to initial SWRs (i.e., you need 25-times your pre-tax portfolio income goal when you retire) because that’s what it should be … a starting place. But the optimal withdrawal rate will vary based on the unique preferences and situations of each retiree.

4% for all!

Early research by Bill Bengen resulted in what has become known as the “4% rule.”[1] This suggests a retiree can withdrawal 4% of the initial portfolio balance upon retirement, where that amount is increased by inflation (annually) for the length of retirement, which is assumed to last 30 years. This seminal work, along with other subsequent research, relied on historical returns, and the Ibbotson stocks, bonds, bills, and inflation data series (which goes back to 1926) was commonly used.