compoundsimStock-Bond Allocation by Age: Drawdowns and Failure Rates
Is 100 minus your age the right share of stocks? Five stock-bond mixes over 1,000 futures show what matters more than age.
Published
Is 100 minus your age the right share of stocks? Five stock-bond mixes over 1,000 futures show what matters more than age.
Published
A common rule of thumb says your percentage in stocks should be 100 minus your age. At 30 that is 70% stocks; at 65, 35%, with bonds growing as you get older. To see how well that holds up, we compared five stock-bond allocations, from 100% stocks down to 20%, across 1,000 simulated futures.
More bonds made the falls along the way shallower and the growth smaller. But what split the results was not age itself: it was how many years were left before the money was needed, and whether it was drawn down along the way. For money needed in five years, more bonds meant fewer paths ending below the starting amount: 8.0% with 100% stocks, 3.8% with 20%. For a retirement drawing 4% of the starting sum a year, raised with inflation, for 30 years, cutting stocks below 60% made the money run out more often: 5.9% of paths with 60% stocks, 20.3% with 20%. A 65-year-old who follows the rule with 35% stocks and draws down over 30 years lands on the side where money runs out more often. Age can stand in for the years until the money is spent, but it says nothing about how long you will draw it down or how deep a fall you can sit through without selling.
The comparison works like this:
Monthly returns come from a model that switches between a calm regime, when prices move quietly, and a wild regime, when stock prices swing more and fall more often. The model is fitted to the full history; the article on Monte Carlo simulation explains how it works.
Over 20 years, we measured two things for each path: its max drawdown, the largest fall from a previous high, and its longest recovery, the most months it took from a high to get back to it. A path still below its high at the end of the 20 years counts the months up to the end.
| Stocks | Max drawdown (median) | Max drawdown (10th percentile) | Longest recovery (median) | Longest recovery (longest 10%) |
|---|---|---|---|---|
| 100% | -29% | -45% | 43 months | 90 months |
| 80% | -24% | -38% | 40 months | 85 months |
| 60% | -19% | -30% | 37 months | 81 months |
| 40% | -14% | -22% | 34 months | 76 months |
| 20% | -10% | -16% | 34 months | 75 months |
The median max drawdown (the middle result of the 1,000) got shallower with every step toward bonds, from -29% with 100% stocks to -10% with 20%. The 10th percentile (the fall one path in ten went beyond) followed the same pattern. Why bonds tend to hold up in a crash, and the exceptions such as 2022, when bonds fell too, are covered in Does diversification work in a crash?
Recovery times differed far less than the depths did. The median longest recovery was 43 months with 100% stocks and still 34 months with 20%. A bond-heavy mix falls less, but it also climbs more slowly afterward, so getting back to the old high still takes time.
In the same experiment, this is the share of paths that ended below the starting sum, by how many years it was held.
| Stocks | 5 years | 10 years | 20 years |
|---|---|---|---|
| 100% | 8.0% | 3.3% | 1.5% |
| 80% | 6.5% | 2.4% | 1.2% |
| 60% | 5.1% | 1.7% | 0.8% |
| 40% | 3.9% | 1.2% | 0.5% |
| 20% | 3.8% | 1.7% | 0.2% |
Over five years, 8.0% of paths ended below the start with 100% stocks, against 3.8% with 20%. Over a short span, a stock-heavy mix can fall hard and run out of time before it climbs back. Over 20 years, few paths in any mix ended below the start, because there was time to recover.
What bonds reduced, though, was the chance of ending below the start, not how low the bad outcomes went. After five years the 10th percentile was about 1.1× the starting sum in every mix. The median, meanwhile, was 1.7× with 100% stocks against 1.3× with 20%, and after 20 years the two had spread to 8.2× and 2.5×.
The gap between 5 and 20 years is one reason behind the advice to own more bonds as you age. For most people, getting older means fewer years until the money is spent, and less time to make up a fall. Money that won't be touched for decades is a different case: in every mix, few paths end below the start, so what more bonds change there is mainly how deep the falls go along the way, and how much less the money grows.
With 100% stocks, the share of paths ending below the start is lower than in history. Looking at the actual S&P 500 (SPY, dividends reinvested) from every starting month since 1993, 14% of five-year holding periods lost money, and 8% of ten-year ones did. Those periods overlap, though: every ten-year period that lost money started between 1998 and 2000. Read the table for the differences between mixes, not for the levels.
Next, drawing the money down in retirement. The setup is the one in the 4% rule article: withdraw 4% of the balance in the first year and raise that amount by 2% a year for inflation. The failure rate is the share of paths whose money ran out before the end, which is 100% minus the success rate the simulator shows. The 60% row is the same set of paths as in the 4% rule article.
| Stocks | Failure rate, 20 years | Failure rate, 30 years | Failure rate, 40 years | After 30 years (10th percentile) | After 30 years (median) |
|---|---|---|---|---|---|
| 100% | 2.3% | 5.3% | 8.3% | 52% | 683% |
| 80% | 1.5% | 5.1% | 9.3% | 39% | 412% |
| 60% | 1.0% | 5.9% | 12.5% | 19% | 227% |
| 40% | 0.5% | 8.7% | 20.5% | 4% | 108% |
| 20% | 0.4% | 20.3% | 48.7% | 0% | 37% |
The balance after 30 years is in today's money, as a share of the balance at retirement.
Over 20 years, more bonds lowered the failure rate a little: 2.3% with 100% stocks, 0.4% with 20%. Over 30 years, everything from 100% down to 60% stocks failed at about the same rate, and the rate rose to 8.7% at 40% stocks and 20.3% at 20%. Over 40 years the gap grew wider still.
The cause is that bonds grew too slowly to pay for the withdrawals. US aggregate bonds (AGG) returned 2.9% a year with dividends reinvested since 2003, which leaves little after 2% a year of inflation. The withdrawal is 4% of the starting balance in today's money every year, so the more bonds in the mix, the less it earns against what it pays out, and the longer the span, the more paths run out. Even at the 10th percentile, the stock-heavy mixes had more left after 30 years.
The more stocks in the mix, the deeper the falls along the way (the first table). The failure rates above already include paths with a crash right after retirement; the article on sequence of returns risk explains why the timing of a crash changes how a retirement's withdrawals turn out. Whether you could hold on through such a fall without selling more than the withdrawals does not show up in this table.
The 100-minus-age rule reads as using age as a stand-in for the years until the money is spent. But two 65-year-olds, one about to spend part of their savings and one drawing down over 30 years, are in the two cases above where more bonds worked in opposite directions. 35% stocks falls between the 40% and 20% rows, and over 30 years of withdrawals those failed in 8.7% and 20.3% of paths. Age also says nothing about whether you have a pension or other income, or whether you would hold on through a fall.
Three things are worth checking when you compare mixes for your own situation: