In the current environment of low bond yields, what is a safe investment withdrawal rate during retirement? In their January 2013 paper entitled “The 4% Rule is Not Safe in a Low-Yield World”, Michael Finke, Wade Pfau and David Blanchett model the risk of exhausting wealth for different retirement durations, withdrawal rates, stocks-bonds portfolio mixes and assumptions about future market conditions (stock and bond returns). They model retirement portfolio dynamics via Monte Carlo simulations applied to log-normal return distributions, based on an inflation-adjusted withdrawal rate set at a percentage of investment portfolio value at retirement and annual rebalancing to a target asset allocation (with failure whenever a withdrawal results in a zero balance). They assume withdrawals cover any taxes due and ignore any portfolio fees/frictions. They focus on a 30-year retirement under current market conditions, with bonds yielding an average annual real return of -1.4% (based on current yields for 5-year Treasury Inflation-Protected Securities) and stocks yielding an average annual real return of 4.6% (6.0% equity risk premium over the bond yield). They also consider a case based on 1926-2011 average annual real returns of 2.6% for bonds and 8.6% for stocks, and cases for which current low returns revert to historical averages five or ten years into retirement. Using the specified modeling assumptions and parameters, they find that:
- Based on the conventional 4% withdrawal rate over a 30-year retirement, with stock allocations of 30%, 50% or 70%, retirement portfolio failure rates are:
- 6%, 6% and 8%, respectively, assuming historical average annual real returns of 8.6% for stocks and 2.6% for bonds.
- 47%, 33% and 28%, respectively, assuming 6% average annual real returns for stocks and 0% for bonds.
- 77%, 57% and 46%, respectively, assuming the current annual real returns of 4.6% for stocks and -1.4% for bonds.
- 22%, 18% and 18%, respectively, assuming the current annual real returns for the first five years of retirement and historical returns for the next 25 years.
- Based on a 4% withdrawal rate over a 30-year retirement under current market conditions (annual real returns of 4.6% for stocks and -1.4% for bonds), even the optimal allocation of 100% to stocks results in a 40% retirement portfolio failure rate (see the chart below).
- For 95% assurance that the retirement portfolio will not fail, the maximum sustainable withdrawal rates and optimal stocks/bonds allocations are:
- 5.0% withdrawal rate and 27%/73% stocks/bonds allocations for a 15-year retirement.
- 3.6% and 32%/68% for a 20-year retirement.
- 2.7% and 33%/67% for a 25-year retirement.
- 2.2% and 44%/56% for a 30-year retirement.
- Assuming current market conditions persist, replacing bonds with single-premium annuities (for longevity protection) may improve retirement outcomes.
The following chart, taken from the paper, plots retirement portfolio failure rates during a 30-year retirement for stock allocations ranging from 0% to 100% and withdrawal rates (WR) ranging from 3% to 6% under current market conditions (annual real returns of 4.6% for stocks and -1.4% for bonds). Squares indicate the stock allocations for different WRs that minimize portfolio failure rates. For example, even a 3% WR with optimal 70% allocation to stocks results in a portfolio failure rate over 20%. The probability of exhausting wealth during retirement for the conventional 4% WR and allocations of 60% stocks/40% bonds is 50% .
In summary, modeling based on current U.S. bond and stock market returns indicates that the safe investment withdrawal rate over a lengthy retirement is much lower than the conventional 4% of wealth at retirement.
Cautions regarding findings include:
- As noted, analyses assume that withdrawals pay for taxes and optimistically ignore portfolio fees/frictions.
- The assumption of a 6.0% U.S. equity risk premium based on historical data may be optimistic.
- Analyses assume tame return distributions. To the extent that actual distributions are wild, the modeling methodology breaks down.
- Spending requirements may vary considerably during retirement, as addressed in “Allocating Assets for Retirement”, such that a constant withdrawal rate does not match needs.
Frequently Asked Questions
What is the 4% withdrawal rule for retirement?
The 4% withdrawal rule is a conventional retirement guideline in which a retiree withdraws 4% of the portfolio's value at retirement in the first year and adjusts that dollar amount for inflation each year thereafter. In the Finke, Pfau and Blanchett analysis, a portfolio fails when a withdrawal results in a zero balance. The rule was tested over retirement horizons of 15 to 30 years.
Is the 4% withdrawal rule still safe when bond yields are low?
The 4% withdrawal rule is not safe under current low-yield conditions, according to Monte Carlo simulations over a 30-year retirement. With real returns of 4.6% for stocks and -1.4% for bonds, failure rates are 77%, 57% and 46% for 30%, 50% and 70% stock allocations. Even a 100% stock allocation fails 40% of the time, versus 6% to 8% failure under historical average returns.
How are safe retirement withdrawal rates calculated using Monte Carlo simulation?
Safe withdrawal rates are calculated by running Monte Carlo simulations on log-normal stock and bond return distributions, with an inflation-adjusted withdrawal set as a percentage of the portfolio value at retirement and annual rebalancing to a target allocation. A portfolio fails whenever a withdrawal results in a zero balance. The model assumes withdrawals cover taxes and ignores portfolio fees, and a safe rate is the highest one with a 95% chance of not failing.
How do safe withdrawal rates differ for a 15-year versus a 30-year retirement?
Safe withdrawal rates fall as retirement length increases, based on 95% assurance of not exhausting the portfolio under current market conditions. The maximum rates are 5.0% for 15 years, 3.6% for 20 years, 2.7% for 25 years and 2.2% for 30 years. The optimal stock allocations rise from 27% to 44% across those horizons.
What are the limitations of using a fixed withdrawal rate model for retirement planning?
Fixed withdrawal rate models rest on assumptions that may not hold, so their results carry several cautions. The analysis ignores portfolio fees and frictions, assumes tame (non-wild) return distributions, and uses a 6.0% equity risk premium based on historical data that may be optimistic. A constant withdrawal rate also may not match spending needs, which can vary considerably during retirement.