An asset allocation rule of thumb that’s been used for decades is to set the bond percentage of your portfolio equal to your age, then adjust upward or downward based upon your tolerance for portfolio volatility. So if you’re 40 years old, roughly 40% of your portfolio should be in bonds, and roughly 60% in stocks.
As generally-applicable guidelines go, it’s a pretty decent one. But I think it leaves out something important:
How long do you expect to live?
Let’s imagine two investors: Lucy and Kelsey. Lucy was born in 1950, and Kelsey was born in 1990. As a result of their differing birth dates, Kelsey’s life expectancy is 7 years longer than Lucy’s.
In order to ensure that she has enough savings to pay for those 7 extra years, Kelsey will have to do (at least) one of the following:
- Retire later,
- Have a lower standard of living in retirement, or
- Have more money saved by the time she retires.
Options #1 and #2 aren’t particularly appealing to most people (though personally I’m rather fond of the “don’t retire” idea). That leaves Kelsey with option #3. Before she can retire, she’ll need more money than Lucy needed–even after adjusting for inflation.
In order to accumulate that extra savings, Kelsey has two options:
- Invest more inflation-adjusted dollars each year than Lucy did, or
- Shift her asset allocation more heavily toward equities–especially during her working years.
Of course, “invest more money” is easier said than done. As a result, it would seem reasonable to me for Kelsey to have a different asset allocation than Lucy.
Unmentioned assumptions
Every article, book, or blog post discussing asset allocation includes assumptions about life expectancy.
For example, I recently finished reading a 1990 edition of Burton Malkiel’s A Random Walk Down Wall Street. In the book, Malkiel suggests various asset allocations for investors of different ages. However, what was an appropriate asset allocation for a 50-year-old investor in 1990 might not be appropriate for a 50-year-old investor today, given a difference of 6 years in life expectancy.
The same thing happens even with articles that were written very recently. For example, if an expert suggests a given asset allocation for a 65-year-old, it’s significant to note whether the expert is suggesting that allocation for somebody who is 65 today, or somebody who will be 65 a few decades from now.
The problem: Laziness
Unfortunately, I doubt that there’s much value in analyzing articles as to their assumptions about life expectancy. Why? Because (I’d bet) most authors aren’t even thinking about it when they write. (And I must admit, I’m guilty too.) We often assume that what is appropriate today will be appropriate 30 years from now, even though one of the fundamental factors–how many years’-worth of money a retiree will need–is constantly changing.
A few years is a big difference.
It may be tempting to look at a 6 or 7-year difference in life expectancy as no big deal. After all, it’s less than a 10% increase. But what’s important is that it can mean a very large percentage increase in terms of length of retirement.
Just look at it this way: It takes a lot of money to pay for 6 years of living expenses.
Takeaway
I don’t have any quick-fix here–no formula where you can plug in your numbers and get the perfect asset allocation. I just think we need to remember that asset allocation isn’t a two-factor decision (i.e., age and volatility tolerance). It depends upon life expectancy as well.
Hi. I'm Mike Piper, the author of this blog. I'm 
I actually think its more helpful to think about the longevity of the money vs the longevity of the investor. For example, an 80 year old might have a 40 year horizon – because she is investing knowing that her grandchildren are the beneficiaries. On the other hand, a 30 year old might have a very short time horizon because she needs the money to buy a home.
Thanks. This piece was really well written.
Excellent distinction, Neal.
I wrote the article exclusively with retirement savings in mind. But you’re absolutely right.
I found your blog through monevator.com, and I’d like to begin by saying how interesting and informative your posts are. This post is another example: I’d not thought at all about life expectancy and asset allocation before (though I’m 21 and not investing yet so no harm done!).
What I would like to add is that increasing life expectancy is hitting British pensions especially bad. Here in the UK the standard way of converting a pension pot into income is to buy an annuity (indeed most people are obliged to use at least 75% of their pension pot to buy an annuity). But because of longer lives (and recently falling gilt yields) annuity rates have halved over less than two decades for those retiring at 65 – in 1990 you could get 15%, on a standard level annuity (i.e. not inflation-linked); now 6.5% is the best rate around. (See http://www.guardian.co.uk/money/2009/apr/12/buying-annuities-pension-advice )
Moreover, the longer a pensioner is going to live after retirement, the more they lose to inflation unless they get an index-linked annuity, which pays an even lower rate. It’s a shame that something so fantastic as our ever-improving health is causing such financial problems.
Hi Niklas.
Thanks for providing the UK perspective. In the US, people seem to have an aversion to annuities due to the fact that, for decades, most of the annuities offered here carried exorbitantly high fees. Now, however, a low-cost variable annuity is likely one of the best ways to make it through a very lengthy retirement.
Regardless, we’re faced with the same issues as you: Longer retirements means more years of expenses plus more negative effect from inflation.
“It’s a shame that something so fantastic as our ever-improving health is causing such financial problems.”
Well said.