Asset allocation

Stocks vs. bonds in retirement: choosing a mix and what it does to your plan

The Stock / bond mix is one of the five Market Assumptions in Runway, and of everything on that step it moves the results the most. It is also the one with the least settled answer. Here is what the research says, what Runway's engine finds for the sample couple at five different mixes, and an uncomfortable feature of the results: the two methods Runway uses disagree about which mix is best.

Quick answer. Stocks have grown faster than bonds over long periods but swing more, and retirement is when swings hurt most, because you're withdrawing. William Bengen's 1994 research, which gave us the 4% rule, concluded that "the historical record supports an allocation of between 50-percent and 75-percent stocks as the best starting allocation." Runway's engine, for a couple spending $95,000, finds the chance their money lasts is about 97.1% to 98.9% from 20% to 60% stocks and falls to 91.3% at 100% stocks in simulated markets, while the median ending balance climbs the whole way, from $1,023,916 at 20% stocks to $3,242,298 at 100%. In a replay of 124 real starting years, the 40%, 60% and 80% mixes lasted in 98.4% of them at $95,000, 100% stocks in 96.8% and 20% stocks in 86.3%. A higher stock share buys a bigger expected ending balance and costs more bad-case risk, and where you land is a judgment, not a calculation.

What does the research say?

Bengen's 1994 article in the Journal of Financial Planning (reprinted in the Journal's "Best of 25 Years" series) tested withdrawal rates against historical U.S. returns at different stock allocations. His advice to planners: "accept a stock allocation as close to 75 percent as possible, and in no cases less than 50 percent. Stock allocations lower than 50 percent are counterproductive." He found that more than 75% stocks was also counterproductive: accumulated wealth kept rising, but the portfolio's longevity in the worst years, like the Depression, deteriorated.

A different question is whether the allocation should stay fixed. Wade Pfau and Michael Kitces ("Reducing Retirement Risk with a Rising Equity Glide Path," Journal of Financial Planning, January 2014) tested 121 glide paths across 10,000 Monte Carlo simulations and found that portfolios that start at 20% to 40% in equities and rise to 60% to 80% generally performed better than static or declining ones. Both studies are historical or simulated, they use their own assumptions, and neither is a guarantee. They are useful because they point at the same trade-off Runway shows below.

What does Runway's Monte Carlo show?

For the sample couple Runway loads (Pat 65, Alex 64, $1.46 million, plan through 95), here are 2,000 simulated markets at five stock shares, using Runway's defaults: stocks 7.5% and bonds 2.93% a year after inflation, with 16% and 6% volatility. The columns show the unlucky (10th percentile), median and lucky (90th percentile) ending balances.

What does Runway's Monte Carlo show?
StocksLasts at $95,00010th pctMedian90th pct
20%98.7%$308,099$1,023,916$1,959,356
40%98.9%$507,337$1,566,878$3,276,085
60%97.1%$509,967$2,194,630$5,384,208
80%94.3%$323,237$2,772,321$8,367,464
100%91.3%$84,726$3,242,298$12,555,308

Two things stand out. The median ending balance rises with every step toward stocks, more than tripling from 20% to 100%, and so does the 90th percentile, from $2.0 million to $12.6 million. But the bad case, the 10th percentile, peaks at 40% to 60% stocks (about $510,000) and then falls: at 100% it is $84,726, well under a fifth of that. And the chance the money lasts is highest at the conservative end (40% stocks, 98.9%). When spending is tighter against the portfolio, the shape is the same but the stakes are larger:

What does Runway's Monte Carlo show?
StocksChance money lasts at $110,000Median ending balance
20%53.9%$52,271
40%72.9%$437,582
60%77.8%$929,223
80%77.3%$1,367,462
100%75.9%$1,739,739

At $110,000 a year, the best odds of lasting are at 60% stocks (77.8%), with 80% (77.3%) close behind, and both very low and very high stock shares do worse. That is the sequence-of-returns problem in numbers: a portfolio with more stocks is more likely to leave you a lot, and also more likely to run out (why the first decade matters most).

What does the historical replay show?

Runway's Historical Backtest starts the same plan in every real year from 1872 and runs it for 31 years on actual U.S. stock and bond returns. The picture is different:

What does the historical replay show?
StocksLasted, $95,000Lasted, $110,000Median end, $110,000
20%86.3%40.3%$0
40%98.4%55.6%$186,780
60%98.4%68.5%$796,168
80%98.4%76.6%$1,668,841
100%96.8%78.2%$2,183,714

In the replay, more stocks is at least as good at every step at $110,000: 40.3% of starting years succeed with 20% stocks, 76.6% with 80% and 78.2% with 100%. This disagrees with the Monte Carlo result, in which the odds peaked at 60% stocks and fell after that. The reason is what each method assumes. Runway's Monte Carlo uses a 7.5% real return for stocks and 2.93% for bonds, a gap of about 4.6 points. In the historical data (1872 to 2025), U.S. stocks averaged 8.6% a year after inflation and bonds 2.8%, a gap of nearly 6 points, and bonds actually lost money in real terms from 1940 through 1981, at an average of -1.7% a year. A 31-year replay therefore rewards stocks more than the simulation does. If you believe the future looks like the past, the replay is the better guide; if you expect returns to be lower than they have been, the simulation is. Neither is the "right" one, which is the point: how much stock you hold depends on a forecast nobody can make.

So what mix should I choose?

Try different mixes on your numbers. Change the stock share on Market Assumptions, or slide it in Explore Your Plan, and compare the simulation and the backtest.

Try the planner

What does this leave out?

How were these numbers computed?

The Monte Carlo uses 2,000 simulated paths and seed 2026, with stocks at 7.5% a year after inflation (about 10.2% before it, with inflation at 2.5%), bonds at 2.93% (5.5% before it), volatility of 16% and 6%, and a correlation of -0.10. The historical backtest uses annual real, dividend-reinvested total returns for U.S. stocks and 10-year bonds from Robert Shiller's public dataset, in 124 starting years from 1872 through 1995, each running 31 years. Both methods use the sample couple, married filing jointly, plan through age 95, with the long-term-care cost left out. Amounts are in today's dollars. The Runway app now shows these same plans in future dollars (each year raised by the inflation assumption), so its figures run larger than the ones here.

Runway content is educational only and is not financial, tax, or legal advice. Consult a qualified professional before making financial decisions.

About the author

Lei Huang is a former professor, turned founder and developer. He builds Runway, a DIY retirement income planner whose planning engine computes the figures in these posts. He is not a financial advisor.