The steady-savings retirement plan

By Felix Salmon
February 22, 2011
Wade Pfau has a fascinating paper out called "Safe Savings Rates: A New Approach to Retirement Planning over the Lifecycle".

" data-share-img="" data-share="twitter,facebook,linkedin,reddit,google" data-share-count="true">

Wade Pfau has a fascinating paper out called “Safe Savings Rates: A New Approach to Retirement Planning over the Lifecycle”. It’s really just the bones of such an approach: the details need to be fleshed out a lot. But I love the idea that we should get away from thinking about “the number” we need to be able to live comfortably in retirement. The effect of the number is to break life into two — pre-retirement and post-retirement: your goal pre-retirement is to reach the number, while your goal post-retirement is to spend it down slowly enough that it doesn’t run out before you die.

Pfau’s insight is that thanks to mean reversion, the number you need at the end of a bear market is actually lower than the number you need at the end of a bull market — if the market’s about to head up, your retirement savings can grow even post-retirement, while if the market is about to fall, you’re liable to lose much more than just your annual expenditures. Instead, says Pfau, stop thinking about stock, and just think about flows. Save a set percentage of your salary every year, stick to it, and, it turns out, you’ll be fine:

Starting to save early and consistently for retirement at a reasonable savings rate will provide the best chance to meet retirement expenditure goals. You don’t have to worry so much about actual wealth accumulation and actual withdrawal rates, as they vary so much over time anyway. But the savings plan should be adhered to regardless of whether it seems one is accumulating either more or less wealth than is needed based on traditional criteria.

What’s the percentage? That’s the crucial question. Pfau makes a very basic calculation that for someone on a constant real wage, saving for 30 years and then living for another 30 years on 50% of their final salary, saving about 16% of your salary each year into a portfolio of 60% stocks and 40% bonds will put you into safe territory.

Of course, real wages aren’t constant over time, and all the other figures are highly variable too. But the bigger message certainly resonates with me: spend less effort on trying to boost your annual returns, when you have very little reason to believe in your alpha-generation abilities, and spend more effort on maximizing your savings every year.

Investing can be exciting, especially when it’s done wrong. You follow the markets rising and falling, you obsess about your retirement-fund balance, you rotate out of this and into that, you read books and magazines and blogs to try to learn more about what to do. You might even, in a moment of weakness, find yourself watching CNBC. Budgeting, by contrast, is like going on a diet: it’s a drag, and it’s hard to get any pleasure or excitement out of it. But the latter is much more likely to get you well-set in retirement than the former.

Update: Matt Yglesias has a this-thing-looks-like-that-thing moment.

28 comments so far

Felix, thank you for writing about my article. I think sometimes people feel like they get misrepresented by journalists, but you’ve nailed it here. I’m in full agreement with everything you wrote and with the way you portrayed my article. Best wishes! Wade

Posted by wpfau | Report as abusive

I applaud Wade for this advice – but (and this is not a criticism in any way!) it appears remarkably similar to Scott Adams’ 9-point financial plan, which the Dilbert author announced back in 2006. Scott listed 9 points, the 8th of which is almost exactly the portfolio breakdown suggested.
More importantly, Scott made a list that is really easy to follow, and reads more like a diet than an investment strategy!
Wade’s insight is to explain why a simple approach should work. Three cheers for that!

Link to an article on Scott Adams’ financial plan: s-9-point-financial-plan-worthy-of-econo mics-nobel?siteid=myyahoo&dist=myyahoo

Posted by andy_mn | Report as abusive

Wow! Saving 1/6th of your income (hopefully before taxes) is a huge chunk, yet it only allows living on half your income when you retire. That will definitely mean a cut in standard of living.

Also, while the 60/40 equity-bond split may be a good average over one’s life, it strikes me that one would want to be higher in equities early on and higher in bonds in retirement (when there is no “long-term”).

So, while I’m generally comfortable with the approach, I think there are some kinks that need to be examined.

Posted by Lilguy | Report as abusive

The big problem is that even if you do everything right– e.g., follow Scott Adam’s plan + rebalancing, you’ll fall short by some 10 to 20% no matter whether you look at capital totals or capital flows. The numbers just don’t add up.

Posted by MattF | Report as abusive

Lilguy, those numbers jive with my own financial planning. We target a savings rate of 30% (or more) of our gross income, and aim to replace 70% of our income in retirement. Different methodology, but very similar results.

I’ve struggled with the equity-bond split myself, and suspect that is a highly individual choice. One approach is to calculate the present value of your planned savings stream (I would use at least a 10% discount rate for this calculation) as part of the total. Focus primarily on equities until your actual savings is equal to the present value of your future savings, then adjust to a 60-40 split after that.

But then my mood swings with the market. I’m much more comfortable with 80-20 when the market is low, uncomfortable (as I was in 2007 and am again today) when the market is on a bull run.

Note that even in retirement, you still need to take a long view. Most people can expect to live 25+ years in retirement. Many will live 30-40 years.

Posted by TFF | Report as abusive

I agree, TFF, on some level that it feels more comfortable to be in a higher stock ratio when the market is low, and if one is counting on mean-reversion that’s the way to go — I’m hoping to read the paper this afternoon, but Felix implies that’s an assumption the paper makes — but it feels against the spirit of the result, which is to focus on savings rate and not on optimizing returns. Perhaps a bit less against the spirit would be to adjust the ratio as one moves from ten years before likely retirement to ten years after retirement — not so much trying to time the market or beat a long-return as simply acknowledging a change in risk tolerance. Even this, though, might detract from the message; a simple-minded 60/40 split might well do better than what a lot of people would end up with if they tried to do something just a little bit more sophisticated.

Posted by dWj | Report as abusive

I prefer focusing on a certain lifestyle, or standard of living rather than a number, or even a percentage of income (of course, as long as that lifestyle requires less than what you earn). Over a period of 30 years or more, you learn what lifestyle makes you feel comfortable, and you more or less stick to it.

At my current income, that works out to saving around 25-30 percent of gross, but I don’t care about the number. The savings for retirement come as a side benefit to the lifestyle you choose, and it ensures that you will need less in retirement than what you are earning.

Posted by Curmudgeon | Report as abusive

dWj, I agree. A 60/40 split is a solid, simple approach that offers peace of mind for the typical investor. It might not be quite optimal, but it shouldn’t be too far off. (Note that annual rebalancing helps capture mean reversion.)

The paper doesn’t assume a whole lot — it analyzes the given strategy using data from the last century. Not a whole lot of independent 60-year samples there, so there is a chance that returns in the future will be weaker (and the necessary savings rate will be higher), but at least it suggests what has worked in the past.

Posted by TFF | Report as abusive

Curmudgeon, that is our approach as well.

The problem is that if your lifestyle eats up 80% of your income, you aren’t going to be saving enough to support it in retirement. So it is good to run some numbers as a check on what you are doing — and if you aren’t saving enough, then you need to adjust.

Posted by TFF | Report as abusive

Thanks for the comments everyone. Just to clarify a couple of things:

The 60/40 asset allocation assumption has no special meaning. That is just the classic balanced portfolio for a conservative investor. I agree strongly that reducing stock allocations as one ages is useful. Also, some of you are suggesting to alter your asset allocation based on whether stock market valuations are high or low. I’m doing some work on that issue as well now. But to illustrate the basic point for the article, I thought keeping things simple with 60/40 would be okay.

Also, about whether a 50% replacement rate is too low. Please keep a few things in mind: after retiring, you no longer have to save for retirement or pay the Social Security payroll tax, and I am assuming that you will receive Social Security benefits on top of this 50%. So in the end, it should work out to be a pretty decent replacement rate in terms of what you could actually spend before retiring.

Posted by wpfau | Report as abusive

Great comment thread here… while I agree that returns will be lower for the current working generation than for the boomers hopefully better tools will help make up that difference.

My company recently adopted auto-enrollment in our 401k plan… new hires now have to fill out forms and sit with HR NOT to be signed up rather than the reverse. Better still we now have an auto increase feature in our 401k that will bump your contribution percentage every year without you haveing to do a thing.

We also have auto rebalance so if you want that 60/40 bond split your 401k would have been selling equities into 2007′s overheated market and buying equities in the dark days of 2008 – 2009 when most small fish were jumping out of the frying pan and into the fire.

Posted by y2kurtus | Report as abusive

Finance: rediscovering what actuaries have known for years.

This can be simplified greatly–assume that investment returns match inflation, no more and no less. That’s not as radical a statement as it might seem: real returns for fixed income are at best 1-2% more than inflation, and real returns for equities average to something less than 4 or 5% over inflation. Portfolios should be heavily skewed towards risk when investors are younger and skewed more towards fixed income near retirement, so dollar weighted returns are much closer to the 1-2% mark. If you make the conservative move of assuming that real returns are 0, then (in constant dollars) retirement is just (percentage saved)*(years of saving)/(number of years to live).

The big driver thus is not the return–it’s what percentage you save, how long you save and how long you live. Everyone is focusing on returns, but the big picture is dictated by other factors–length of service, percent of income saved and longevity.

Which is why buying a life annuity, especially one with a COLA benefit that adjusts for inflation, makes sense. The largest single controlable risk is outliving one’s savings. That risk can be controlled by replacing an uncertain death date with a weighted average death date–which is essentially what buying a life policy adjusted for inflation does.

Posted by loopguy | Report as abusive

I think Wade’s study is a huge advance. The Old School safe withdrawal rate studies fail to consider valuations and are thus dangerous as all get-out (the historical data shows that the valuation level that applies on the day the retirement begins is the single biggest factor that determines the true SWR). Wade’s approach makes sense because it indirectly takes valuations into consideration.

I have a New School safe withdrawal rate calculator (“The Retirement Risk Evaluator”) available at my site. The Risk Evaluator takes valuations dirinto account in determining the SWR. Thus, rather than reporting the SWR as always being 4 percent, it shows that it can drop to as low as 2 percent at times of high valuations and can rise to as high as 9 percent at times of low valuations. calculator.html


Posted by RobBennett | Report as abusive

Incidentally, Felix, remember a year or two ago when you were surprised that target date funds don’t move entirely to cash as you approach the target date? This is related to why they don’t do that; it’s not the amount of wealth at retirement that should be optimized, or even a concave function of that, but the amount that that will secure for retirement.

Posted by dWj | Report as abusive

Great article. It seems that the 60/40 split is reasonable and that long-term it can provide the flow of income that I would need to sustain a healthy retirement. My financial advisor has been helping me deal on the psychological side of things when I would worry about the market ups and downs. But he stresses that savings are by and large the major part of getting me to the retirement I want, not so much as cherry picking investments.

Posted by johnjacobson | Report as abusive

There are some pretty good retirement articles at

Posted by johnjacobson | Report as abusive

Investing would be so easy if we could just assume that markets are mean reverting, like Pfau. Unfortunately, Taleb’s insight, that mean reversion is bullshit, is far more persuasive.

Posted by maynardGkeynes | Report as abusive

“If you make the conservative move of assuming that real returns are 0, then (in constant dollars) retirement is just (percentage saved)*(years of saving)/(number of years to live).”

Let’s see… 30% saved * 30 years of saving / 30 years of retirement = 30% replacement. I can’t live on 30% replacement, can you?

Real returns may “only” be 2% to 4% over a lifetime of savings and investing, but that 2% makes all the difference!

Posted by TFF | Report as abusive

“Which is why buying a life annuity, especially one with a COLA benefit that adjusts for inflation, makes sense.”

And what is the payout rate on that? Insurance companies demand a large profit on the annuities that they offer, and an additional profit on top of that if they are taking any risk at all (such as promising inflation-adjusted payouts).

The least risky strategy is to give your life savings to charity and join a convent. You are 100% guaranteed to die poor if you do that. Buying COLA annuities from insurance companies is barely any better than that.

Posted by TFF | Report as abusive

I have been taking a similar approach for the past decade. It is clear that both equity valuations and interest rates need to factor into savings and withdrawal strategies.

Doug Short occasionally posts a “Total Return Roller Coaster” chart on his website that looks at S&P 500 rolling 10, 20, and 30 year total returns: ite-annualized-total-return-roller-coast er.html . Since few of us have a large lump sum that can just sit over 30 years without inputs or outputs, I think the 10 and 20 year charts are the most valuable for evauating investing and withdrawal stragies for the typical person.

A couple of things leap off the page of these charts:

1. Subzero annualized real total returns over 10 year periods can persist for up to a decade – it only recently went below zero in 2008; and

2. Once the 20 year real total return drops to below 5% annualized, it invariably goes to near zero within a few years indicating a long period of poor returns. It has only recently dipped below 5% in 2008-2009.

Both of these charts indicate to me that we are likely to have another significant market decline over the next few of years. After that we are likely to begin a long upward trend that likely could generate real total returns of 8% or more per year for a decade or more.

Similarly, interest rates are very low now (BND yield = 3.5%) so income from bonds over the next decade should be assumed to be very low with the potential for capital losses over that period.

As such, my savings strategy is simple: I need to be saving 20%-30% of my current income for the next decade as I cannot count on the stock or bond markets to deliver decent total returns to build up my nest egg during that period. However, this makes it a good period for long-term investing as I expect that after retiring in 10-15 years, I will likely be able to pull money out at greater than Bengen’s 4% safe withdrawal rate (probably at 5%-6%) if the equity valuations are relatively low (less than Shiller’s PE10 of 16) at that time.

At retiremement, the expected total return from equities based on their PE10 valuation at that time plus the interest rate on a total bond fund is a good starting point for determining whether or not a safe withdrawal rate can be bumped up to 5% or 6%.

The current PE10 of about 23 ( ite-pe-ratios.html) and the BND yield of 3.5% tell me that the 8% returns assumed by many pension fund managers today are ludicrous at best and criminally fraudulent at worst. Just another example of the governing class of people gaming the system to buy themselves another handful of years raking money off the top before their bubble world bursts.

At least I can make a decent stab at understanding how much my 401k/IRA is really worth looking at it in a multi-decadal way. My other source of valuation is to pay attention to what insurance companies are willing to payout in annuitized payments. They seem to be the only part of the financial sector living in the same world that the average person inhabits.

Posted by ErnieD | Report as abusive

If Wall Street had not been polluting the the world with self serving theories about stock returns “for the long run,” 90% of the posts on this thread would not be here. The entire debate is debased and devalued. Hopeless. If there’s an original idea out there, I could sure use it now.

Posted by maynardGkeynes | Report as abusive

Original idea for you, MGK: asset prices are meaningless, only cash flow counts. Cash flow is the life blood of any business, any household. Cash flow from operations, not from financing.

Whether you invest in the stock market, in bonds, in real estate, or in “hard commodities”, be certain that the free cash flow generated is sufficient for your needs. Anything beyond that is speculative.

Posted by TFF | Report as abusive

@TFF, It’s a good thought, certainly more original (and insightful) than the comments here.

Posted by maynardGkeynes | Report as abusive

Assume you begin investing at 25 and at 45 you would hold a balanced portfolio with some risk aversion coefficient, lambda, and the investor is passive. Back out the expected returns of the balanced portfolio with reverse optimization (u=lambda*S*w where S is the covariance matrix and w is your balanced portfolio). However, they dynamically adjust their risk aversion coefficient by lambda*2^((age-45)/20). Then, using the expected returns found above, optimize the portfolio given the lambda here. When they are 25 their risk aversion coefficient is half that at 45 and when they are 65, their risk aversion coefficient is twice when they are 45. If the optimizations are unconstrained, this would mean the 25-year old (65-year old) portfolio is exactly twice (half) as risky as the 45-year old portfolio.

Any views on the investment outlook should be relative to the balanced portfolio and considered a separate problem (though not a separate optimization). Further, the investor could also dynamically adjust the risk aversion coefficient of the balanced portfolio (though not in the reverse optimization process) based on market conditions.

To the extent that he would prefer a constant savings contribution to take advantage of mean-reversion, then he is not necessarily passive. He does have an opinion on the return generating process. I would prefer to keep active and passive decisions separate. For instance, you could assume that the returns over your original optimization time horizon will be distributed as your inputs into the optimization (which could be influenced by views) and then normally distributed according to the original covariance and reverse optimized mean until say 30 years after death. You could then set your savings rate such that the probability that your wealth 30 years after your death is less than 50% of your final salary is less than 5%.

This is obviously more than any individual investor could do, but it wouldn’t be so hard to program into something a financial planner could use.

Posted by jmh530 | Report as abusive

“…on 50% of their final salary”
Q: Felix, if one is a renter, not a homeowner, …

Posted by AmicusAlso | Report as abusive

Good insight, Amicus, and another reason why simplistic income-replacement guidelines don’t hold water. Every individual situation is different.

Note that investment income used to pay rent is taxed (and may trigger taxation of Social Security benefits). Yet if you “invest” in your home, the “income” of low-cost housing is entirely tax-free and does not count towards your AGI.

Yet another example of how income taxes create truly perverse incentives. The homeowner is likely already wealthier than the renter, so the renter is rewarded by paying higher taxes in retirement?!?

Posted by TFF | Report as abusive

In the paper “Safe Savings Rates: A New Approach to Retirement Planning over the Lifecycle” the author, Wade Donald, seeks to critique his interpretation of current financial planning practice with respect to retirement planning only to replace it with, in effect, the same process.

From page 3:

“I am suggesting that the following retirement planning process, which is one which isolates the accumulation and decumulation phases, is not appropriate.”

The author goes on to summarise current financial planning practice to be (the steps listed below have been reworded so as to read from the third person);

Step 1: Estimate the withdrawals needed from financial assets to pay for planned retirement expenses after accounting for Social Security, defined-pension benefits, and other income sources. Planned retirement expenses are defined as a replacement rate (RR) from pre-retirement salary.

Step 2: Decide on a comfortable withdrawal rate (WR) shown to be sufficiently capable in the historical data.

Step 3: Determine the wealth accumulation (W) retirement goal, defined as W = RR / WR.

Step 4: Determine the savings rate (SR) needed during working years to achieve this wealth accumulation goal.

Donald suggests replacing Steps 2 to 4 of the above process with the following:

Step 2: Decide on a comfortable savings rate (SR) based on what has been shown to be sufficiently capable of financing desired retirement expenditures in the historical data.

On page 4, Donald then goes on to define what a sufficiently cable savings rate should be.

“A particular savings rate was successful if it provided enough wealth at retirement to sustain 30 years of withdrawals without having the account balance fall below zero. Actual wealth accumulations and withdrawal rates may vary substantially for different retirees.”

This statement is simply a restatement of Steps 2 to 4. Put another way, this is what Steps 2 to 4 do leading the reader to wonder what, exactly, is new with this approach.

Posted by Galician | Report as abusive


There is a subtle difference, though it has a big implication. The difference is that you don’t independently choose a withdrawal rate and a wealth accumulation goal, but just directly save enough to finance your desired retirement expenses. It leads to the differences between the black curve and the blue curve in Figure 5.

Posted by wpfau | Report as abusive
Post Your Comment

We welcome comments that advance the story through relevant opinion, anecdotes, links and data. If you see a comment that you believe is irrelevant or inappropriate, you can flag it to our editors by using the report abuse links. Views expressed in the comments do not represent those of Reuters. For more information on our comment policy, see