Investment returns in retirement?

Just wondering what other people are using for investment return rates, in retirement.

Here’s my take: If you believe in Bengen’s 4% rule (that for a 30 year retirement, you can safely spend 4% of a 50% equities/50% bonds portfolio the first year, then the same dollar amount indexed for inflation over the next 29 years), then running a 30-year scenario with an Excel spreadsheet and experimenting with different rates of return suggests that you get equivalent results by plugging in a real return of approximately 1%, i.e. with that rate of return every year, no ups or downs, you will be broke after 30 years (if you withdraw, say, an inflation-adjusted 40K every year from a million dollar account).

Some folks who do this sort of thing for a living feel that nowadays the 4% rule is too optimistic, because the run-up of the equity risk premium in the latter half of the 20th century is unlikely to continue in the 21st, and because bonds produce such meager returns, and suggest changing the 4% rule to the 3% rule. Playing with the Excel spreadsheet, this corresponds approximately to a linear real return of NEGATIVE 1%.

Anybody care to comment on this? Obviously, the choice of a reasonable rate is dependent on how aggressive a portfolio you want to choose, how lucky you feel, how many years you think you might live, etc., but I’d appreciate any feedback on my thoughts.


I'd change the question to try and understand how various combinations of inflation and rate of return impact your situation. Basically a sensitivity analysis. At certain points (e.g. low inflation, high returns) you are probably in good shape, other combinations would be okay, and still others (e.g. high inflation, low returns) will have a big hit on your consumption / standard of living. By doing this, you can estimate where you’ll be okay or better and where you’ll have serious issues.

Unfortunately there's no simple Bell curve for inflation / RoR and poor combinations seem more plausible today although we really don't know if they are or not.

Overall, depending on your vulnerabilities, you can try to adjust to manage those areas that seem most critical.

Specifically regarding Bengen, you may want to consider his book "Conserving Client Portfolios During Retirement". He goes into much more detail that may be of interest.


Well, I think I hear what you’re saying, but let me rephrase my question, see if it changes your response.

I believe in the Bengen rule, except that I would now call it the 3% rule (see above). What I mean by that is that I think it provides a reasonable starting point for a retirement period of 30 years, which I hope to start in a year or two. Like any reasonable person, I would modify my spending in subsequent years based on actual investment returns, and of course there all kinds of scenarios for the next three decades, but you’ve got to start somewhere.

I’m comfortable with at least a 50% equities allocation. I think 30 years is a reasonable upper limit for my remaining life expectancy. I think planning on a level spending pattern over the 30 years is a reasonable guess, anticipating that a decline in spending on travel, etc., would be offset by an increase in medical expenses, and in paying people to do things that I now do for myself. May not work out that way, of course, but, again, you have to start somewhere.

If I had just a pile of stocks and bonds, that would make for an easy calculation: Just take 3% of the value to find your first year’s spending. But I have a mortgage with a number of years left, and three immediate annuities, two already kicked in, one to begin in a few years (which are not inflation-adjusted). That makes it very hard to just apply a 3% rule. And of course the 3% rule skews your discretionary spending lower at first, since taxes come out of the 3%, and taxes generally decline as your portfolio shrinks with time, hence the amount left for spending after taxes increases with age. And with a substantial mortgage, thinking about taxes becomes even more complicated.

Hence my purchase of ESPlanner. And my question is: What choice of annual investment return will give you a spending recommendation that is about as conservative as the 3% rule. This is just to provide a ballpark number for first-year spending. Based on fiddling with some sample spreadsheets, I conclude that negative 1% is about right. But I’ve been known to screw up spreadsheets, and maybe I’m not thinking about this clearly in the first place. So I was interested to know if anyone else has thought along these lines, and has reached different conclusions. Your comments are certainly interesting, but I’m not sure they help me with my need to have ESPlanner make some sort of recommendation for spending in my first year in retirement.

These are thoughtful questions. As I mentioned above, you may be interested in Bengen's book "Conserving Client Portfolios during Retirement". This came many years after his initial 4% rule and talks about assets invested in tax-sheltered vs. taxable accounts, different time horizons, various withdrawal options (e.g. higher / lower with annual ceilings based on market performance), rebalancing, and other related topics.

To your question (and based on your description with a mortgage, annuities, etc.), I wouldn't focus overly on the 3% figure. Here's why:

First, ESPlanner is already trying to smooth your living standard per adult for the rest of your life (including spouse). It is almost a certainty that you could raise your living standard by making various choices (e.g. delaying Social Security, deferred annuity, reverse mortgage, choices around retirement contributions/withdrawals and many other variables). I call this profile optimization.

Once you have optimized your profile, the "Total Income" report will show specific year-by-year recommendations for withdrawals. By comparing the withdrawal amounts to your assets you can see how close they are to your 3% figure. As an example, if you retire at 66 and delay Social Security until ages 68-70 plus continue paying off your mortgage, you may need/want "special withdrawals" to bridge your income gap for a few years. This would lead to higher than 3% withdrawals for a few years, then (ideally) 3% or lower once you maximize your Social Security and eliminate your mortgage. Of course, RMDs will occur, but you just have to withdraw those amounts and not necessarily spend it each year.

If you can safely raise your standard of living sufficiently through "optimization", then you may determine that you can accept a lower rate of return and still be "safe". In a previous post (, I wrote out some ideas on how to manage risk in ESPlanner that might be of interest including a safety factor approach of spending less than ESPlanner's recommendations (including withdrawals) and how to model this.

What I'm getting at here is an approach to raise your standard of living while managing your risk. I compare my results against Bengen's values (from his book) as an additional sanity check.

Since we are all making guesses about the future, I use the inflation/rate of return option that I mention above to see how I would do under different scenarios. For example, 6% rate of return/3% inflation (2.913% real return) might be considered somewhat conservative. However, if inflation averages 5% and you have a fixed 7% rate of return (1.905% real return), your results will change. In my testing, this exercise gave me another sanity check.

I know this is a round about way of trying to answer your question. Years ago, I asked essentially the exact same question having studied Bengen before using ESPlanner. Now I use profile "optimization" to raise the living standard per adult, then back it off somewhat using a "safety factor" approach. By comparing ESPlanner's recommendations with benchmarks (such as Bengen's), it may provide an additional level of confidence.


dan royer's picture

Nice Brian!

Brian –

Thanks for the thoughtful reply. And thanks for elaborating on your recommendation of the Bengen book; I’ve just ordered a copy.

I’ve already looked into most of your suggestions. I did delay social security until I turned 70, so that’s maximized. And 70 is even further behind for my wife, so likely the 30-year horizon is overly optimistic – or pessimistic, if you look at it from a spending standpoint! I’ve maximized my SEP-IRA contribution at work, but it just comes right back out in my RMD (which, of course, reduces the amount of equities I need to sell to meet the RMD). I just got a QLAC quote from a strong insurance company last week, for the maximum 125K premium, and plugged the numbers into ESPlanner. It bumped my spending by 1 to 3% for a variety of scenarios, and I’m seriously considering it. Something doesn’t seem quite right, though; my taxes before it kicks in at 85 should be lower, because I’m removing 125K from my IRA’s, hence should have lower RMD’s, but ESPlanner shows that the taxes are actually a little higher. Probably my mistake, I’ll go over the details when I have time and try to figure out what happened.

I’m a little puzzled by your description of a 6% return/3% inflation scenario as “conservative.” As an average, yes, it’s conservative. But the reason the 4% in the “4% rule” is so low, while returns were on the average much higher in the 20th century, is of course because of the sequence of returns risk. So a nominal 6% return is conservative as an average, but not (I think) as a straight-line, year after year assumption, the way ESPlanner (in the non-Monte Carlo planning) does it.

Parenthetically, I might add that I’ve discovered that thinking about retirement spending is one thing when it’s years away, but quite another when it gets to be close at hand, and you’re thinking of pulling the trigger on being unemployed…

I think you'll find the Bengen book insightful. He really digs into the content. Years ago I created a "Bengen calculator" spreadsheet with all the factors in the book and still use this when assessing ESPlanner's withdrawal values.

Regarding the "somewhat conservative" 6%/3% figures, it really depends on your point of view. Historically, this would have been quite conservative in the US, but who knows going forward. Just to double check, I looked at the ESPlanner MC return values for large cap stocks and intermediate-term government bonds from 1929-1945. Bengen shows a 60/40 portfolio with rebalancing would have been okay with a 4% withdrawal rate. showed that a 4% withdrawal rate worked in 109 out of 115 time periods (30 years in length) from 1926 to today. Of course, the initial conditions aren't the same and a 3% withdrawal rate may well be more realistic as a safe amount. William Bernstein seems to agree among others. To cope with the uncertainty, I use the risk management approach mentioned above.

One idea that might be interesting for some is to have another set of eyes analyze their profile. After understanding their profile and goals, it may be possible for others to think of areas for improvement...especially in safely raising living standards and dealing with risk. This seems somewhat impractical today, but could happen at some point in the future.

Good luck!


Food for thought. Thanks for all your comments. Leaving tomorrow to go hiking in the Gobi Desert in Mongolia, perhaps that will be an environment to do some thinking...

And of course there's always the "work 'til you drop" option.

dan royer's picture

Just curious, I looked at the first year smooth withdrawal calculated by ESPlanner in my own model. If I assume an annual nominal return of 6% (3% inflation) It has begin with a first-year draw of 4.45%. If instead I assume 5% nominal return, it has me draw 3.4% the first year. I guess I wouldn't approach it this way because I better trust my intuition about what 5% nominal or 6% nominal means going forward than I do backing into an assumed rate with a 4% rule--which seems like what you would use if you didn't have a calculator like ESPlanner.

We use cookies to deliver the best user experience and improve our site.