There’s a better way to get wealthy

I’m Todd, and I created Financial Mentor to give you a step-by-step blueprint for building wealth that actually works.

Free Instant Access

More than 10,000 people have already used this blueprint to jumpstart their financial freedom.

Are Safe Withdrawal Rates Really Safe?

189 Flares 189 Flares ×

Reveals the Little-Known Risks Hiding Behind The 4% Rule That Can Jeopardize Your Retirement Security. Discover The Simple Solutions To Correct The Problem…

What do retirees in 1921, 1966, and 2010 have in common?

Very little… and that’s the problem.

Each faced a different life expectancy and invested in a different economic climate with varying inflation expectations, interest rates and market valuations.

The truth is these dates weren’t chosen at random: One had the highest safe withdrawal rate in recorded history, the other the lowest, and the third barely survived the ravages of inflation.

Each of these three retirees lived through dramatically different economic times; yet, according to conventional wisdom they all share the same safe withdrawal rate in retirement – roughly 4%.

It makes no sense.

How can a static, one-size-fits-all solution to a problem as varied and complex as knowing how much money you need to retire be correct? How could retirees in 1921, 1966, and 2010 share the same safe withdrawal rate when market valuations, interest rates, inflation expectations, and expected lifespans were completely different.

It’s impossible. It’s wrong.

Yet, that is the conventional wisdom in the financial planning profession. It is known as the “4% Rule”, and it is widely considered to be “the truth” in safe withdrawal rates for retirement.

The problem is it’s not the truth and every day people risk a lifetime of retirement savings on it. There are better solutions…

In this article I will reveal the problems hiding behind the 4% Rule and provide you with practical solutions you can implement for your retirement security.

Why Safe Withdrawal Rates Are Critically Important

Every new retiree needs an answer to the same critical question – what is the maximum retirement income I can withdraw from savings without running out of money before I run out of life?

It is the single most important question I get from retirees and near retirees.

The reason is because safe withdrawal rates impact every aspect of retirement planning – from the lifestyle you can afford to the amount of savings needed to fund it. Small errors in safe withdrawal rates multiply over many years causing huge financial impacts…

  1. Lifestyle You Can Afford: Few people realize that a mere 1% safe withdrawal rate change makes a big difference in spending during retirement. It seems counter-intuitive because the number is so small. However, a 1% variation from the industry standard 4% assumption will increase (or decrease) your income in retirement by 25%. That can make the difference between a world traveler lifestyle or living at home on hot dogs. In other words, it pays to calculate your safe withdrawal rate as accurately as possible because small changes equal huge changes in the lifestyle you can afford.
  2. Savings Required: The amount you can spend each month from savings and the amount of savings you must build to support your retirement are flip-sides of the same coin. One implicates the other mathematically. For example, the “4% safe withdrawal rate” is mathematically equal to the “Rule of 25” (you need 25 times your first year spending in savings). Similarly, a 3% safe withdrawal rate equals roughly 33 times your first year retirement spending in savings. Using the two examples above, a mere 1% change in safe withdrawal rate when spending $100,000 per year in retirement is the difference between building a nest equal to $2.5 million vs. $3.3 million – obviously a big deal. Conversely, knowing you can safely spend 6% would knock the savings requirement down to $1.7 million. That’s why it is so important to figure an accurate safe withdrawal rate: It can change the amount you need to save for retirement thus shortening the time it takes for you to reach your financial goals.
  3. Risk To Financial Security: If you withdraw just 1% more than your actual safe withdrawal rate you will go broke before you die. It is a mathematical truth. A mere 1% less would have allowed your nest egg to last a lifetime. That’s why accuracy is critical – because the razor thin margin between 1% too much and getting it right is literally the difference between poverty and financial security.

Like Goldilocks, there is a sweet spot in retirement spending somewhere between “too much” and “too little” that is “just right”. In an ideal world you would exhaust your last penny from retirement savings as you exhaled your last breath. That is the theoretical objective of safe withdrawal rates.

It is a high-stakes game where the quality of your life during retirement is dependent on getting the answer right. For that reason, there is probably no question more important in retirement planning.

Unfortunately, the conventional wisdom can be dangerously misleading…

Lesson Learned:Withdrawing the right amount of money from savings is one of the most important retirement planning questions you will confront. Take too much and you blow up: take too little and you leave lifestyle on the table. You need to get as close to the right amount as possible. It’s worth the effort – and it’s worth reading this long article to understand the issues.

A Quick History Of Safe Withdrawal Rate Research

The state-of-the-art knowledge in safe withdrawal rates has progressed dramatically since the early 1990’s. What began as a naïve exercise in simple amortization has progressed to sophisticated regression analysis and Monte-Carlo research.

You may not want to learn the research history behind safe withdrawal rates and skip straight to the results and conclusions – but it’s not that simple. The 4% rule became the sacred cow for a reason, and a new generation of research is teaching important principles that can make big differences to your financial future.

It pays to learn this stuff. It sets the context for the rest of the discussion that follows and is essential to your making smart decisions with your money.

To keep these ideas accessible I’ve organized them into three distinct generations of knowledge…

Generation 1:

The first generation in safe withdrawal rate research was naively based on mortgage amortization – but in reverse. Rather than pay down a mortgage (like your home) you would draw down an asset account (like your retirement savings). The idea seemed intuitively correct even thought it was fatally flawed (due to volatility and returns sequencing risks as shown below).

Even the venerable Peter Lynch (1995) succumbed to the intuitive appeal of the mortgage-style model when he falsely stated a 7% withdrawal rate would be prudent for an all stock portfolio. He was forced to retract this obviously incorrect statement when 2nd Generation research proved conclusively how such a withdrawal rate could land you in the poor house.

Generation 2:

Bengen (1994) ushered in the second generation of safe withdrawal research when he published a groundbreaking Journal of Financial Planning article that still sets the basic framework employed by most research to this day. He used historical simulations of long-term U.S. securities index data to define “Safemax” as the highest withdrawal rate, expressed as a percentage of the account balance on the first day of retirement, and adjusted for inflation annually, that allowed for a minimum of 30 years of withdrawals over all rolling historical periods in the database.

Bengen concluded the maximum safe withdrawal rate was about 4.1% for stock allocations between 37%-67% and later upgraded that amount to 4.5% when small cap stocks were included in subsequent research.

Cooley, Hubbard, and Walz (1998) took the 2nd Generation model to the next level by applying Monte Carlo analysis (ed. note – please see comments below for correction) to overcome the sequencing risk in data (as defined below). It showed a 95-98% chance you won’t run out of money applying a 4% withdrawal rate. This study, dubbed the “Trinity study”, and popularized by Dallas Morning News columnist Scott Burns, was updated in the April 2011 Journal of Financial Planning with similar results.

Other 2nd Generation researchers have also come to similar conclusions depending on assumptions applied. Results vary slightly from study to study based on asset allocation, data sources, whether or not fees are deducted, frequency of portfolio rebalancing, and much more.

The key point defining all 2nd Generation research is that each study applies the same basic premises thus producing extraordinarily consistent results. This consistency caused the 4% Rule to become conventional wisdom and be mistaken as “truth” when it is really just a product of the research premises.

To understand the problems with 2nd Generation research we need look no further than the amazing breadth of dubious assumptions behind the results…

  • Safe withdrawal rate research was based on U.S. securities data history. No foreign market data was included.
  • The research typically assumes a static allocation to stock and bond indexes as the only viable asset classes. Alternative assets are not included.
  • It assumes 30 years of retirement spending regardless of expected longevity.
  • It assumes no investment expenses by using historical index data – obviously not true for any real-world investor.
  • It assumes a fixed spending amount that grows with inflation and does not adjust based on changes in portfolio value or age – not true for most retirees.
  • It seeks to determine the highest beginning spending amount that can be adjusted for inflation without ever running out of money that works across all time periods in the database – a sort of least common denominator concept.
  • All data periods are created equal with no adjustment for valuations or interest rates at the time you retire.
  • You must have 30 years of subsequent data to know if the withdrawal rate was actually safe meaning retirements beginning after 1982 (this was written in 2011) are unknown and theoretical (out of sample).

While these assumptions make for expedient research there is a clear sacrifice of accuracy when compared to the real-world retirement you will face.

Below I will examine each of these assumptions in detail to show you the implications so that you can decide how relevant 2nd Generation research conclusions (The 4% Rule) are to your retirement planning…

Generation 3

3rd Generation safe withdrawal research has attempted to correct the assumption limitations outlined above thus providing a more accurate picture of how safe withdrawal rates might apply to real-world retirees like you and me.

In essence, 3rd Generation models stand on the shoulders of the excellent research that came before it by recognizing the limitations, correcting the flaws, and producing more accurate models. The 4% rule has serious problems and 3rd Generation research seeks to correct those errors.

Rob Bennett was an early pioneer in 3rd Generation modeling by advocating (through various online forums) that withdrawal rates must be adjusted for market valuations consistent with research by Campbell and Shiller (1998).

Also, Wade Pfau (2010 – 2011) broke new ground by applying safe withdrawal rates to international market data with shocking results. He also applied valuation, interest rate, and inflation metrics in regression analysis to form a dynamic and robust safe withdrawal rate model.

The key point illustrated by 3rd Generation research is that a deeper level of complexity underlies the sacred cow “truth” known as the 4% rule. It was the best answer for its day, but those days are gone. It is a second generation model whose shortcomings have been proven well enough that it must be retired.

With that said, the third generation research is in its infancy and has only corrected certain shortcomings from the 2nd Generation model. There are many remaining assumptions you must still individually answer to determine your personal safe withdrawal rate.

For that reason I will examine each of the shortcomings of the 2nd Generation models below and use 3rd Generation research to demonstrate viable solutions where they exist. In the process I will explain all issues remaining unresolved and point you toward viable answers to consider. The result will be a safer withdrawal rate than the simple rule-of-thumb provided by conventional wisdom.

Lesson Learned:Safe withdrawal rate research has been through 3 generations of growth and sophistication in a very short time. However, the 2nd Generation model, known as the 4% Rule, is unfortunately locked in most expert’s minds as “truth” when 3rd Generation research has already discovered more dynamic models. Regardless of the model used some underlying assumption problems remain that must be adjusted for.

In the paragraphs that follow I will examine each of the underlying problems built into 2nd Generation safe withdrawal rate research so that you can adjust your retirement plan for greater financial security and peace of mind…

Backcasting – Data Sampling Problem

Data limitations are one of the most obvious problems with all safe withdrawal research. 2nd Generation models have been tested exclusively on U.S. data – sometimes as far back as 1871. While that is a solid long term study it completely ignores international data.

The implied assumption is the future should be no worse than the past as evidenced by U.S. asset price growth.

But is that really true?

Well, the truth is the U.S. was the prom queen of the economic world for the last 130 years. She led a privileged life that likely introduces an overoptimistic bias to research based exclusively on her data. Consider the following…

  • U.S. stock market capitalization grew from 22% of world total in 1900 to 54% in 2003.
  • The U.S. stock market grew at 6.3% real compounded and never had a losing 20 year period. This compared to an average growth of 5.4% for other developed countries with only 3 providing all positive 20 year holding periods.
  • U.S. stocks compounded at a higher rate than all but 3 other countries.
  • U.S. stocks had lower volatility of returns than all but 4 other countries.
  • Australia was the only country that had both lower volatility and higher returns than the U.S.
  • U.S. bonds had higher real compound returns than all but 3 other countries and lower bond volatility than all but 2 countries.
  • Only Switzerland had both higher bond returns and lower bond volatility than the U.S.
  • Only 2 countries experienced lower compound annual inflation than the U.S.

Because the U.S. enjoyed the highest returns with lowest volatility for stocks, bonds, bills and inflation, it is simple math to conclude any research into safe withdrawal rates based on this data would likely provide some of the most optimistic outcomes of any data set. This isn’t opinion: it is just the way the math works. It is a fact.

The U.S. led a charmed existence with all the right connections and resources to bubble to the top of the economic heap. She was the prom queen of the world economy. Her experience is not representative of the rest of the world, and it is highly questionable the U.S. will repeat her performance during your retirement for many reasons including the following…

  • Interest rates (as of this writing) are below U.S. historical averages, dividend yields are below historical averages, and P/E ratios are above historical averages. All of these facts imply less than average investment returns (see next section on valuation models).
  • The last 100 years were built on cheap oil and plentiful natural resources. The U.S. used a disproportionate share of those resources to fuel its economic growth and consumption demands. Maybe technology will solve these problems, or maybe natural resource limitations will inhibit overall economic growth.
  • The U.S. also enjoyed a stable political environment with no destructive wars on its soil compared to its developed competitors.
  • The world economy has become more homogenized. The degrees of separation have diminished. U.S. companies sell and manufacturer world-wide while their competitors do the same. It is truly a global economy in ways that didn’t exist historically.
  • U.S. debt levels have ballooned by every measure compared to economic history placing a serious question mark around the stable inflation assumption of the past.

Are you willing to bet your financial security on the charmed economic stats of the prom queen during her heyday when the world is a changed place today? Should you bet on the prom queen returning to the podium or is it more realistic to expect some kind of mean reversion that could lower asset returns to average expectations?

This is more than a cute analogy. Your financial security in retirement depends on it.

Wade Pfau (source for much of the data cited above) was the first to put this idea to the test. He applied a research methodology similar to Bengen but with the critical change of using historical international data instead of U.S. data. The results were alarmingly different.

In other words, similar methodology as 2nd Generation models + different data = dramatically different conclusions.

Using 109 years of data for each of 17 different developed countries Pfau determined that a 4% withdrawal rate with a fixed 50/50 asset allocation would have failed in all 17 countries. Yes, a 100% failure rate. You could have run out of money before you ran out of life using the conventional assumptions on foreign country data.


Lesson Learned:The 2nd Generation models used to prove the 4% Rule showed a surprising failure rate when applied to international data. Basing your retirement expectation on results from U.S. data alone is the economic equivalent of basing your expectations for the High School Prom on the prior 10 years prom queen’s experience. It is overoptimistic. This can be proven either by comparing U.S. economic statistics to the rest of the world or by actually running the models on the actual asset price data. They both point to the same conclusion – if you bet your retirement on becoming the prom queen you stand a good chance of being disappointed.

The Dramatic Impact of Sequencing of Returns on Safe Withdrawal Rates

In the last section you learned the critical role that data assumptions play in safe withdrawal rates by seeing how international data indicated a potentially lower return expectation than U.S. data.

In this section you will discover how safe withdrawal rates are actually dynamic – not static as commonly taught. You will learn how the sequencing of investment returns and inflation during your early retirement years will make or break your financial security.

The sequencing of returns problem is best illustrated in this example from William Bernstein. Assume you have a $1,000,000 portfolio with an average return of 10% split evenly between 15 years at +30% and 15 years at -10%. This would give you a compound return of 8.17% (compound is less than average due to volatility effects). More importantly, when you vary the returns sequences you get something truly shocking…

  • If you are unlucky and start your retirement with 15 straight losing years you can only withdraw 1.86%. Same annual returns, same average return, different sequence of returns, different result.
  • Conversely, if you are lucky enough to start your retirement with 15 straight winning years you can safely withdraw 24.86%.

These are astounding results!

Sequencing risk causes your safe withdrawal rates to vary from an average of 8.17% to as low as 1.86% (in this example) or as high as 24.86%. This variation is solely caused by the exact same returns occurring in a different order. Nothing else changed. Amazing!

4% Rule & Safe Withdrawal Rate Sequence of Returns Infographic

You can embed this infographic on your own site – just copy/paste this code into your post or page…


As shocking as these numbers are it is really just common sense when you think about it. Imagine 15 years of no net investment gain (not hard to do with the stock market’s performance since year 2000) while still withdrawing 4% per year for spending. Even without inflation adjustments you would wipe out 60% of your account just in spending alone. When you add inflation and investment losses to the equation the overall destruction to equity would be the retirement equivalent of death by strangulation.

By the way, this is not some strange statistical mumbo-jumbo that has no bearing on your retirement. This is real-world stuff that is critical to your understanding. It can make-or-break your financial security. Real people retired in 2000 applying the conventional 4% wisdom and destroyed their nest eggs in the process because of this exact problem.

Sequencing of returns risk is a huge factor in explaining why actual safe withdrawal rates in U.S. historical data vary from the 3% range at the low end to over 10% at the high end (depending on assumptions and the date chosen to begin retirement).

Sequence of returns is determined by the date you retire, cannot be known in advance, and will be one of the most significant factors affecting your financial security in retirement.

It’s a big deal.

The truth is safe withdrawal rates are all over the map depending on what date you retire and what happens to your investment returns in the early years of your retirement. Pfau (2010) concludes that retirement success is highly dependent upon early investment returns showing that wealth remaining after 10 years of retirement combined with cumulative inflation during those 10 years explains 80% of the variation in safe withdrawal rates. This is very similar to Bernstein above.

The importance of this issue cannot be overstated.

The problem is your next 10 years investment returns are unknowable: you don’t get to know the sequence of returns until after the fact. The future can’t be predicted with any accuracy (and it certainly isn’t dependent on the last 100+ years of U.S. average historical data!)

Lesson Learned:Your real safe withdrawal rate for 30 years is highly dependent on the first 10 year’s sequence of returns and inflation rate. One size does not fit all. The 4% conventional wisdom is a static, least-common-denominator approximation, but actual safe withdrawal rates are highly variable. It is one reason why retirees in 1921, 1966, and 2010 face such dramatically different safe withdrawal rates.

What’s a near-retiree to do? As it turns out, all is not lost. There are answers provided in the next section below (but they are not the same as conventional wisdom would lead you to believe!)…

Market Valuations and Safe Withdrawal Rates

2nd Generation research assumed all time periods were equal regardless of market valuations, interest rates, and inflation. According to conventional wisdom a retiree in 1921, 1966, and 2010 should all withdraw the same percentage of savings even though history proves this assertion is patently false (with perfect hindsight, of course).

3rd Generation research concludes differently by attempting to show that safe withdrawal rates vary widely depending on economic conditions on the date of retirement.

Let’s set the stage for this argument with the obvious logic first. Everyone intuitively understands that investment returns are a primary determinant of safe withdrawal rates. The more your portfolio earns during retirement the more you can afford to spend during retirement. That much is clear.

The argument isn’t whether investment returns affect safe withdrawal rates. Everyone agrees that is true. The problem is guesstimating what investment returns will be since you only know in hindsight and your withdrawal rate must be chosen in advance. That’s the problem.

This problem is why 2nd Generation models chose to define the highest withdrawal rate that could survive all historical data periods. The assumption was the best and worst performing periods could not be determined in advance so the only safe choice was the lowest common denominator that survived all time periods.

Fortunately, that assumption is false. Future investment returns are not “luck” or random as many would guess. As it turns out, market valuations at the time you begin your investment holding period are inversely correlated to the return you can expect over the following 10-15 years.

Notice how this 10-15 year time period is identical to the critical time period identified by Pfau in the research on sequencing of returns risk cited earlier – no coincidence.

Campbell and Shiller (1998), Hussman, Easterling, and many others have documented the effect of valuations on subsequent investment returns. It is a well know phenomenon that is valid across data samples and is something you can apply in your own retirement planning.

Is it statistically perfect? No, nothing is. Is it robust enough for me to bet my own retirement on it? Absolutely yes!

The implications are startling. It means the 2nd Generation safe withdrawal rate models were well intentioned but somewhat misguided. Retirees don’t need to know the historical “least common denominator” withdrawal rate that survived most data samples (as the 2nd Generation models taught). They need to know the forward-looking investment expectation given the actual data that exists on the day their retirement begins (which is what 3rd Generation models teach).

These are completely different questions…

Michael Kitces considered the valuation effect alone by testing both optimal asset allocations and safe withdrawal rates based on Shiller’s P/E 10. He concluded 4.5% for P/E 10 above 20, 5.0% when P/E 10 is between 12 and 20, and 5.5% when P/E 10 is below 12 were all safe.

These are not revolutionary variations from the conventional 4% Rule, but they point a clear direction – valuations matter.

In other words, the higher the market valuation the lower the safe withdrawal rate. Rob Arnott (2004) asserted similarly that sustainable withdrawal rates are not a fixed number but evolve with changing market conditions.

Similarly, Rob Bennett provides a safe withdrawal calculator based on regression analysis by John W. Russell that varies output using valuations at the start of your retirement. The key difference between Bennett and Kitces, however, is Bennet’s conclusions are more dramatic with a safe withdrawal rate at the 2000 peak in market valuations using a large allocation to stocks of only 2%.

This is half of what conventional wisdom would claim. Yikes!

More recently, Wade Pfau created a robust model using regression analysis and 3 valuation metrics – PE 10 (price divided by average real earnings for previous 10 years), dividend yield (dividends divided by stock price), and interest rates (on 10 year government bonds) – to explain variation in safe withdrawal rates across time periods reasonably well. It wasn’t perfect, but most of the results were within 1% of being accurate (which is far more accurate and informative than blindly following historical averages).

For example, our theoretical retiree in 1921 enjoyed an astounding 10.42% safe withdrawal rate largely because of historically low market valuations when he retired. Our 1966 retiree faced a difficult future with high valuations and rising inflation causing a 3.53% safe withdrawal rate. This is a difference of 3 times the spending capacity from the same nest egg simply because of the date you retired!

And if that isn’t shocking enough, the 2010 retiree is looking at a 1.8% safe withdrawal rate according to Pfau’s research. No, that is not a misprint – 1.8% – far below the conventional wisdom of 4% based on historical research. It is caused by persistently overvalued markets and razor thin interest rates that simply don’t exist in the historical data.

What this means is 2010 retirees have serious cause for concern when you consider a healthy couple at 65 has decent odds of one spouse outliving the 30 year lifespan assumption and none of the models thus far include administrative or transaction fees (both of these issues are explained in detail below, and they both lower the safe withdrawal rate even further).

Suffice it to say, 2010 is a tough year to begin retirement. The conventional “wisdom” could dangerously mislead you to overspend which would put you at risk of running out of money long before you run out of life.

Lesson Learned:Safe withdrawal rates vary with market valuations, interest rates, and inflation at the time your begin retirement. This connects to the previous lesson because valuations are a strong indicator of subsequent 10-15 year investment performance. When you put these two facts together you have a dynamic model that is more accurate and can be adjusted based on your actual retirement situation.

The safe withdrawal rate you must use is not a fixed number like conventional wisdom claims that can be neatly packaged into a simple rule-of-thumb. It is not the same across all time periods in all conditions. One size does not fit all.

It is a dependent relationship on economic conditions and asset valuations prevailing at the time you retire. This is completely contrary to the 4% rule taught elsewhere.

The Understated Problem With Inflation

In the previous section we discovered how our 1921 and 2010 retirees faced a more than 5 fold difference in expected safe withdrawal rates because of market valuations and interest rates. In this section we look at our 1966 retiree to learn how inflation affects safe withdrawal rates so we can begin assembling a more complete picture.

The key concept to understand about inflation is how it throws a one-two punch at retirement planning.

  1. First off, withdrawal amounts are adjusted annually to reflect cumulative inflation thus forcing a progressively larger annual withdrawal from savings to maintain real purchasing power.
  2. Secondly, periods of high inflation are correlated with lower asset returns (see Unexpected Returns: Understanding Secular Stock Market Cycles).

The combined effect is to increase your annual withdrawals from savings while simultaneously reducing your investment returns – a very difficult situation for any retiree to face. It is the worst of both worlds all at one time because your retirement savings are getting squeezed from both ends simultaneously.

The best historical example illustrating the ravages of inflation included the period from 1966-1996. Surprisingly, it was harder on new retirees than the Great Depression.

Inflation caused the annual withdrawal amount to rise from 4% to above 10% of savings within 15 years of retiring. This also coincided with a nominal return on the S&P 500 of 6.81% roughly equaling the inflation rate and putting the real return at roughly zero. Numbers like these are unsustainable and spell financial destruction.

The only reason these retirees survived was because Paul Volcker wrenched down inflation thus setting off one of the greatest bull markets in history beginning in 1982. The subsequent outsized investment returns offset the outrageously high withdrawal rate that had been caused by inflation in the preceding 15 years thus bailing out a near-death experience for 1960’s retirees.

Consider the 1960’s a warning shot over the bow because the next round of inflation may not end so gracefully. Men of Paul Volcker’s caliber seem in short supply these days and record breaking bull markets, by definition, are extremely rare occurrences and should not be relied upon to repeat.

The other insidious fact about inflation is that it’s not predictable. PhD economists routinely miss their forecasts just one year into the future. The idea of a 30+ year inflation forecast (the duration of your retirement) is an absurd joke.

In short, inflation is an incredibly dangerous three headed monster: it can’t be forecast accurately and it multiplies your spending while reducing average investment returns.

Lesson Learned:Taking inflation adjusted withdrawals over a 30 year period is only realistic in a stable inflation environment otherwise the withdrawal rate grows to an unsustainably large percentage of savings. The bulk of U.S. economic history has seen stable inflation so there is little historical precedent to judge the seriousness of the problem. Foreign history includes bouts of inflation and the effects indicate safe withdrawal rates below 4%. Given the unprecedented government debt levels you should carefully reconsider any safe withdrawal rate that blindly increases spending during inflation.

How A Long, Healthy Life Is A Financial Problem…

Longevity is a key idea when spending principal from savings (as most safe withdrawal research assumes.)

Research shows (based on U.S. historical data – see above caveats) that safe withdrawal rates for 10 year retirements approach 10%. 20-25 year retirements push 5% and 30 years or more dip under 4%.

The rule is simple: the longer your money has to last, the lower the percentage you can withdraw. It is inherent in the mathematics of amortizing a fixed pool of capital (your savings).

Now that we know the math, let’s look at the problem… people are living longer.

Remember our 1921 retiree? He had a life expectancy of little more than 65 years. When Social Security was created they set the retirement age at average life expectancy. It was never intended to fund 30+ year retirements.

Since that time average life expectancy has increased by roughly 1/3 of a year for every year thus increasing by 30 years in the last 100.

Our 1921 retiree didn’t need a lot of savings because he could spend a large chunk of principal every year. Our 2010 retiree doesn’t have that luxury.

Today, a healthy couple retiring at 65 has a good probability of at least one spouse living into their 90s. That means today’s retirees must budget for 30+ years and be extremely careful about any strategy that amortizes their savings by spending principal.

Additionally, this average life expectancy is a moving target expected to increase by the time your date with destiny arrives. If history provides any guide it is reasonable to expect the average to rise another 10 years over the next 30 moving our 2010 retiree into a 100+ year lifespan. This may sound extreme but with developments in biotechnology and nanotechnology it may actually prove to be a conservative estimate.

Finally, understand that all this discussion is about averages, but half the population outlives the averages. Already the 95% confidence interval life expectancy is over 100 years (and rising).

The risk has never been higher that you could outlive your savings. Extremely long retirements exceeding 30+ years are entirely reasonable to plan for; yet, all safe withdrawal research to date is based on the premise that you spend your assets to zero at 30 years. Yikes!

This could be very dangerous.

For many people this assumption could cause you to run out of money long before you run out of life.

Lesson Learned:Both 2nd and 3rd Generation research into safe withdrawal rates has assumed 30 year retirements as the maximum. Trends in human longevity and developments in medicine make that a dubious assumption at best and dangerous at worst. The longer your life expectancy the lower percentage you can withdraw from savings.

The bottom line is a safe withdrawal rate that spends principal is an oxymoron when longevity expands beyond 30+ years. Any spending of principal is not safe over very long time periods. You should adjust your investment strategy and withdrawal rate accordingly.

The Zero Fees And Expenses Assumption

How many of you invest with zero fees and expenses?

Not too many hands raised…

Amazing that most safe withdrawal rate research supporting the consensus 4% rule assumes zero fees and expenses. This is another example of an assumption made for the purpose of expediency in research but having no real-world application.

The solution to this problem is not complex. Real-world withdrawal rates must be reduced compared to theoretical research to reflect real-world investment management and transaction expenses.

This issue may seem small but it is not. Imagine you’ve invested your portfolio with an adviser who charges 1% management fees while investing in mutual finds with 1-2% total expense ratios. That is 2-3% of annual expenses when compared to a 4% withdrawal rate. The difference is huge.

Fortunately, you don’t have to subtract the expenses directly from the theoretical withdrawal rate because the math doesn’t work that way. (It is a common mistake.) Instead, you subtract expenses from the investment return first and then calculate the sustainable withdrawal rate. The reduction in withdrawal rate is significantly less than the actual expenses.

For example, Pfau adjusted his 3rd Generation research results for administrative fees of 1.6% for stocks and 1.2% for bonds (similar to recent Morningstar averages) and it reduced his safe withdrawal rate by .66 percentage points – far less than the nominal expenses.

Lesson Learned:If you invest in low cost ETF’s without additional advisory fees then you may be able to ignore the investment expense issue since its impact should be limited. However, if you invest with an advisor in expensive mutual funds then this issue is a serious consideration that could reduce the amount you can withdraw each month by 10-20%. It is an important issue to consider that few advisors will explain to you… for obvious reasons.

Humans Are Rational… Sort Of!

The final 2nd Generation research assumption built into the 4% Rule that makes no sense is the idea that you withdraw a fixed percentage of your savings adjusted for inflation (but nothing else) each and every year.

This is nonsense! Nobody lives this way.

In real life we adjust our spending based on the success or failure of our careers and the income they produce. Why should retirement be any different?

If your assets got hammered by inflation and bad investment performance during the first 10 years of retirement causing the percent withdrawal to rise above 10% are you going to march like a lemming to the cliff of financial destruction? Of course not! That would be foolish.

You would reduce your spending based on adverse circumstances in your early years of retirement. It is the prudent, common-sense thing to do (but it is not included in the research because it is difficult to model).

Similarly, do retirees consistently spend more each year as they get older? No, quite the opposite occurs. Retirees reduce spending as they age. Why, then, do safe withdrawal models plan for ever increasing spending? It is not how real retirees manage their money.

The point is that a 4% withdrawal rate on the first year of retirement that is adjusted every year for inflation has no real world applicability. It is a fiction of academic research.

Real world retirees increase spending when their assets have a good run and cut back spending when assets get clobbered. They spend more in the early years of their retirement when their health is strong and world travel beckons, and they reduce spending as their energy and health decline with age.

Flexible spending opens up many possibilities not modeled in the 2nd Generation research.

  • You could spend more in your early years and then reduce spending (or forego inflation increases) in later years when you don’t need as much money.
  • You could spend more in the early years and ratchet your spending down if you are unfortunate enough to endure an adverse returns sequence in the first 10 years.
  • You could start at 4% and increase spending if your first 10 years enjoy high investment returns and/or low inflation.

These are just three of many possible variations on how to approach withdrawing money. The bottom line is you don’t have to be a lemming and mindlessly follow the 4% rule into a financial abyss. Such blind obedience could leave a fortune on the table or risk unnecessary financial ruin. Instead, be smart and adjust your spending based on the actual results you experience.

It is just common sense.

Lesson Learned:There are many possible spending alternatives that offer real-world practical solutions to the fixed academic model of blindly increasing withdrawals based on inflation each and every year. You can’t determine the risk of ruin for rational retirees from a model based on irrational behavior.

The truth is risk of ruin is just as dependent on retiree behavior as it is on market dynamics – something not considered by the research and certainly ignored by the 4% rule. You must remain flexible during retirement and use your brain. Correct and adjust your spending based on the growth or decline of your portfolio – be rational and your risk of running out of money will be reduced.

Putting It All Together…

We’ve covered a lot of ground by examining the 3 generations of safe withdrawal rate research and the various problems associated with the 2nd Generation safe withdrawal models.

The reason I focused on the 2nd Generation model is because it has been elevated to the status of “truth” in the financial planning industry. The 4% Rule is quoted regularly in the financial media and used as a benchmark by which all other retirement planning models are compared.

The problem is it’s not really safe. The 4% rule could cause you to leave a fortune on the table or run out of money long before you die. The 4% rule is a static conclusion in a world that is dynamically evolving.

In summary, there are specific assumptions built into the research supporting the 4% model that must be seriously questioned…

  • 4% safe withdrawal rates are based entirely on U.S. investment data covering a time period when the U.S. was the prom queen of the economic world. Research by Pfau on long-term international data shows U.S. results optimistically on the high side of world estimates and found material risk of failure when the same model was applied to international data.
  • 4% safe withdrawal rates fail to account for rational retiree behavior in response to adverse (or positive) returns sequences.
  • 4% safe withdrawal rates are a least-common-denominator approach to finding a static answer. In other words, the 4% Rule is the highest withdrawal rate that survived most data, but that doesn’t mean it is the highest withdrawal rate that might be safe for your situation. It doesn’t take into account market valuations, interest rates, or inflation at the time you retire. Wade Pfau demonstrated such factors can provide dramatically different safe withdrawal rates (both higher and lower) when compared to the 4% rule.
  • 4% rule cannot account for out of sample data. This was first identified by Wade Pfau using international data and surfaced again in regression analysis on the record breaking market overvaluation and low interest rates of the last 10 years. Both situations produced expected safe withdrawal rates well below 4%. In addition, it can’t possibly account for the potential for out of sample inflation that might occur in the future due to government financial mismanagement and excessive debt levels. Don’t believe that 100+ years of U.S. economic history is as bad as it can get. The past is not the future.
  • 4% safe withdrawal rates usually exclude investment management fees and transactions costs. This is not a big deal if you self-manage your portfolio using low cost ETF’s but is a very big deal if you use professional advisors and invest in mutual funds. Adjust accordingly.
  • 4% safe withdrawal rates assume 30 year life expectancy. The problem is average life expectancies hold no relevance for your particular date with destiny. You could live much longer. In addition, a 95% confidence interval for a healthy couple at 65 pushes the number beyond 30 years, and increasing longevity coupled with medical research breakthroughs can extend this number even further. Budgeting for 30 years may cause you to run out of money before you run out of life.
  • 4% safe withdrawal rates assume conventional asset allocation to U.S. stocks and bonds and cannot be extrapolated to include less conventional investment approaches. Skill based investment strategies (active management), commodities, TIPs, and real estate can imply dramatically different conclusions.
  • 4% safe withdrawal rates assume irrational retiree behavior. A rational retiree who adjusts spending based on actual results can significantly decrease his risk of ruin – even at higher initial withdrawal rates.

The truth is your actual safe withdrawal rate will likely be very different from the 4% rule depending on when you retire, your spending patterns, and your investment strategy.

The 4% rule is promoted as a one-size-fits all answer to the question of “how much can you spend in retirement?” Unfortunately, the exact opposite is true. The answer is dynamic – not fixed. It is dependent on all the factors discussed above.

One size fits all is naïve and dangerous.

Don’t believe it… even if it is conventional wisdom.


So what should a retiree do?

If the 4% rule isn’t the answer then what is?

Unfortunately, no simple “plug-and-play” model has surfaced to replace the 4% rule (which probably explains why it has persisted despite inaccuracy). Below I will provide you with a four step process to serve as a guideline in determining a reasonable approximation for a safe withdrawal rate. It includes various adjustments you can make to determine a reasonable withdrawal rate for your retirement situation…

  1. Longevity: Decide first how long you need your money to last. I’m personally budgeting for a 100+ year lifespan because anything less is financially risky (see longevity assumption above). When your retirement time horizon extends to 30+ years then spending principal isn’t safe. As your time horizon shortens (i.e. you get older) then spending principal becomes viable again. Build these facts into your safe withdrawal rate.
  2. Market Valuations: The next step is to assess market valuations at the time you retire to determine if you are in a high risk or low risk period. You can reference Wade Pfau’s research or Rob Bennett’s calculatorfor benchmarks.

    In periods of record high market valuations the low safe withdrawal rate (under 4%) indicated by 3rd Generation research may make alternative asset mixes to the traditional stock/bond portfolio at least temporarily prudent and allow a higher withdrawal rate. For example, during periods of high asset valuation you might consider…

    • Limiting your spending to the premium over inflation on Treasury Inflation protected securities.
    • Investing in inflation adjusted fixed annuities and leave the worry to the insurance companies.
    • Limiting your spending to the dividend income from an internationally diversified dividend stock portfolio.
    • Or you can choose to invest in alternative assets that aren’t governed by the same investment return limitations as traditional paper assets. For example, some retirees live off the positive cash flow from directly owned real estate.

    Conversely, in periods of low market valuations you may be able to prudently increase your withdrawal rate above the 4% rule and allocate a higher percentage to stocks.

    In summary, the second step is to assess the risk level inherent in market valuations so you can decide an appropriate asset mix and withdrawal rate given your longevity expectation and the economic environment.

  1. Refine: Now that you have a benchmark withdrawal rate consider subtracting for other factors discussed above. For example, if you pay investment adviser fees and invest in high cost mutual funds then you may want to reduce your withdrawal rate accordingly. The same with other assumptions described above that fit your situation.
  1. Correct And Adjust: Once you’ve picked your investment strategy and withdrawal rate don’t make the mistake of setting it and forgetting it. Remember the 3rdGeneration research by Wade Pfau showed how the bulk of your financial “blow-up” risk is determined by the sequence of returns and inflation during the first 10 years. You may need to adjust your strategy based on actual results and you certainly shouldn’t blindly increase the amount you spend every year by the inflation rate as the 4% rule would indicate. Revisit you plan every few years so that you never spend too large a percentage of your savings in any one year regardless of what the research tells you. You may even want to consider employing some alternative spending models…
    • You can replace the “4% rule” with the “3% rule” during times of excessive market valuations thus increasing your safety but lowering your income.
    • You can eliminate or reduce the annual inflation adjustment factor so that your spending is fixed in nominal terms but declines in real terms as you age (something most retirees do naturally anyway). One idea to consider is only increasing your spending by inflation during years when your portfolio grows in value.
    • You can change your withdrawal strategy to a fixed percentage of principal which virtually eliminates risk of failure but causes variability in income based on portfolio fluctuations. As your assets rise you will withdraw more and as your assets fall you will withdraw less. Whether or not your spending keeps up with inflation would be determined by the growth of your assets.

The key point is to use common sense.

That means use the research and calculators as guidelines only. Don’t apply static models based on blind faith just because they have become conventional wisdom and everyone says they are true. Your retirement is too important to incur that much risk.

Your retirement will be different from your grandfather’s. Your expected longevity has made any strategy that spends principal in the early years questionable. Financial markets are more volatile and interest rates are at record lows making fixed stock/bond allocations questionable. In addition, inflation could look completely different from historic norms due to unprecedented government debt levels.

In short, the premise that all 2nd Generation research is based upon – that the future should be no worse than the past 100+ years of U.S. economic history – is extremely dangerous. Already the last 20 years has produced a period of record market overvaluations and low interest rates not seen in the historical data. This produced a theoretical safe withdrawal rate less than half that indicated by the 4% rule. Never mind the possibility of out of sample inflation in the future making things even worse.

Don’t blindly trust something as important as your retirement security to computer models. The fact is nobody has a Crystal Ball (least of all economic researchers) and that is why you have to use your brain and not rely on oversimplified rules-of-thumb. They are useful guidelines to teach important principles, but they aren’t scripture in stone.

Hopefully the 4 step model outlined here, while admittedly more complex, will help you navigate the journey and find fulfillment and financial security in your retirement.

Please let me know how this article has impacted your thoughts in the comments below…

The Complete Solution To Knowing How Much You Need To Retire - Brief & Ready For Immediate Download...

How Much Money Do I Need To Retire?

Topics Covered in the Book:

  • 3 simple models giving you an accurate, complete solution.
  • 5 common mistakes you must avoid so you don't run out of money.
  • Concise, no-fluff education so that you spend minimal time learning only what is necessary.


I have a basic problem with the 4% rule concept.  Between my wife and I, we will have 5 staggered income streams: Two Pensions one is already coming the other begins when I get 60, a supplement stream that last about 6 years starting at 56, two Social Security streams one that starts at 67 and the other that starts at 70.  So any computational tool would have to consider the value and timing of the streams.  I would like a spending program that lets me spend more earlier and taper off as I get older.    I can easily set up a stream that pulls more than 4% now but when I finally collect social security at 70 the amount I expect to get I get a jump in salary. Our Social Security strategy is to maximize the payout for long term security. Still there is a small chance of running out of retirement accounts before age 70.  So what is needed to supplement the plan is a 'Safe' plan to re-evaluate my withdrawal rate and recalculate especially when the market is down to prevent running out.    This should also allow me to have a higher initial withdrawal rate.  To the extreme, this could really be a dynamic retirement draw down plan.  Personally, I don't mind leaving my kids a little money, but I'm not interested in leaving them multi-millions (in today's values) when doing so takes a way from my wife and I reasonable standard of living.  Now I have found two tools, Wealth Ruler at TD Ameritrade the other called Flexible Retirement Planner both allow the streams of income to be inputted along with the retirement savings.  Neither allow you to work on a 'Safe' plan strategy as protection in a really bad economic/market season.

This was a good article and very interesting comments.  It took me about 3 hrs to read with comments.


Hi, Todd. With respect to your discussion on PE ratios, I am I right in assuming that the dollars you can take out yearly are the same with high or low ratios since stock values are higher with higher PE ratios even though you're withdrawing a smaller percentage of your portfolio and vice versa? Thanks.


I noticed there wasn't any discussion of higher medical costs and taxes (due to the high deficit, high spend/entitlement mentality). in the models, are higher medical costs assumed to be covered in the monthly spending estimate (thru health insurance/LTC insurance)?  but what about taxes?  they will be higher in the future and the increase may out pace inflation increases.



There are some really applicable rationales summarized here. Thanks for that. The key takeaway for those planning for a retirement far in the future is to keep informed about the latest thinking about safe withdrawal rates so they may adjust savings strategies accordingly. For those much nearer to retirement, it is vital that you not only have a plan but are able to adjust your plan in accordance with current events with regards investment returns from bonds, CDs, stocks, real estate or selling assets on ebay. In today's economy, rising healthcare costs, rising taxes on property or income can change annually and you have to be nimble enough to adjust. France and Great Britain just implemented a wealth tax on those hoarders of assets, for example. The US may not be far behind. The US continues to consider confiscation (er, replacment) of IRA and 401(k) and providing you with an annuity of their choosing in it's place. What does your strategy for dealing with a 2% annuity or 5% wealth tax look like?


I am guilty as charged and have been using many of the withdrwawl models/assumptions you note here.  I find your willingness to share (in great detail no less) your knowledge truly inspiring.  I have more work to do.



Jim Pratt

Todd, I been retired since I was 42 (65 now) from my real estate investments. The 4% rule means nothing to me except that my properties and rents keeps up with inflation, I have 100% controll over my income. I feel sorry for retirees trying to live off their retirement plans with the volatilities of world markets. I have no where near the skills and knowledge that you have, it"s KISS for me all the way. Your articles are very informative and from what I can tell, pretty accurate. Even though I would like to get out of the rental market, finding an investment(s) that would give me consistent returns of 10% (I get over 18% on my rentals) would prove difficult. The reason I read your articles is to gain more financial knowledge because I want more freedom with NO money worries. Life is good, I just want more!


Really! Whatever! This looks like a great article to get me to buy your book, and transfer wealth from my pocket to yours. What a great retirement strategy! We should all do that.... You got it all wrong!

Susan M Kuhn
Susan M Kuhn

Epic, Todd!  I appreciate your digging deeply -- and assuming that your readers really want to work at their financial understanding.  You continue to be my go to source for personal finance. 

Garry Davis
Garry Davis


Great information, as usual. As you are always saying, there is no short cut in this game. Success only comes with hard work. Thanks for sharing your knowledge and perspective.

Rick Francis
Rick Francis


I found your article very interesting- it strikes me that a more realistic calculation could be done.
For example make a system that adjusts systematically according to a portfolio’s real rate of return. If last year’s real return is zero then the drawl rate goes to the minimum possible. If the real return is positive take some percentage of those returns. I suspect if you limited yourself to taking X% of the additional returns in good years you would never run out of money if you started with enough or you could take 0% in the bad years. It seems to me that X would be a function of the initial amount at retirement, the minimum yearly expenses, and the fees and asset allocation. A Monte-Carlo calculation should be able to calculate X as a function of the variables just as accurately as it can calculate a SWR.
Note: A retiree could use a combination of Social Security, fixed pensions and annuities to provide their minimum yearly expenses reserving the portfolio for optional expenses like travel that could be eliminated in tough times.
-Rick Francis

Christine Yip
Christine Yip

Todd, as always, I am amazed at your ability to reduce highly complex research to readable form in an interesting way and most importantly, to offer solutions. This is a beautiful piece of tapestry on SWR stringing together and putting in perspective all the literature to-date and highlighting future patterns to come. SWR is a topic that is of increasing concern to future and current retirees not least because inflation-indexed defined benefits are now largely a thing of a bygone era. I like Steven Martin's comments :" I really don't plan to spend down my assets. I focus on dividends and the income yield that my portfolio can provide. My aim is that my annual income growth simply exceeds annual inflation. A portfolio should be thought of as real property, something that generates annual cash returns, like a rented building. I wouldn't be selling part of the building each year." I think if we were all to spend only real inflation-adjusted income from our retirement capital, then the real value of the capital stays intact even if the spending horizon is forever. Next, Steven mentioned about being tax effective in retirement , which like the management fees that you have mentioned, are urgent "leakages" that need plugging. Also, Todd, it would be great if you can direct future research into COMBINING what a retiree can expect to spend with US$1m or US$2 m or whatever, when COMBINED with other transfer payments such as social security in America or what we call the aged pension here in Australia (which is means-tested but excludes the family home, which perhaps explains why Australian properties are one of the least affordable in the English-speaking world.) Many thanks, and keep doing the great work you are engaged in. God bless you!


"Market overvaluations" sounds an awful lot like "market timing." Truth be told, I cringed when I read the words "Rob Bennett" (four pages into the article). Nevertheless, I agree with much of your article and with at least three of the four steps in your conclusion. (How did four steps sneak into a "three step process" ?)

Manish Malhotra
Manish Malhotra

Todd, this is a beautiful summary of the state of current research in safe withdrawal rate - thanks for putting it together. I agree with your conclusion of using diverse set of ways to construct retirement income including factoring real yields and inflation linked fixed life annuities.

Constructing constant retirement income from sources of volatile returns is obviously not a good idea. I want to mention one simple strategy of generating safe lifelong income using some products you mentioned above and then extend it to a general strategy template.

3.3% withdrawal rate using TIPS for first 20 years and buying life annuity after 20 years.

An investor retiring at the end of this year can move 2012 income needs into his bank account and use a ladder of 19 TIPS maturing from 2013 and 2020 to provide him the inflation adjusted retirement income. To generate real $1 from 2012 to 2020 would require $19.5 based on real yields from Oct 25th WSJ (some approximation of yield for years with missing TIPS)

Buy one more TIPS maturing in 2032 paying 0.7% real yield to fund an inflation linked life annuity with 8.4% payout after commissions (today's quote for 85 year old male and 80 year old female with 75% survivor benefit. Risk is that real yields 20 years down the line could be lower and hence payout could be lower. An advisor may make an opportunistic purchase earlier or later to manage the risk.). That would require setting aside $10.3 today. At 85 years of age the mortatilty credits are too strong for the retiree couple to continue wagering on their longevity.

With total investment of $29.8 ($19.5 in 20 TIPS ladder and $10.3 in another TIPS to purchase annuity) today, you can get $1 of real retirement income, effectively a 3.3% withdrawal rate.

Any variant of such theme of covering short term risk through individual bonds held to maturity, starting at the beginning of retirement and extending these individual bond ladders at opportune market times, followed by use of life annuity to cover tail risk of the retirement can be practiced. These strategies are likely to give much higher confidence in generating withdrawal rate of 3% to 4% - after fees and commissions. These strategies are definitely better than using stocks that are beginning to have higher volatility and expected lower returns than in the past.

Manish Malhotra, CEO, Income Discovery


Wow, great article. What I didn't see is consideration of quality of life at different times duing retirement.
Would a person forgo the last 15 years in retirement for the first 15 years? I plan to travel when I am retired but will quit that later. I plan to be involved with children's families in the first part of retirement and not as much later. If I'm drooling in a wheelchair in a state run nursing home, they can take all my dwindiling retirement accounts!

Any comments on that aspect?

Thanks, got to your article via "The Daily Crux"


Apologies if someone in this discussion has raised this issue.
There seems to be one flaw in these demographic assumptions. Everyone appears to be looking at life expectancy AT BIRTH rather than life expectancy at retirement age. The increase in the former is greatly impacted by the reduction in both infant mortality and fatal childhood diseases of old. This has no bearing on how long a senior can expect to live.
For example, at web site
data from 1850 to 2004 shows life expectancy at birth doubled from 38 to 76 which is a 38 year improvement over 155 years or roughly 4 months increase in life expectancy per year. The same table, however, shows life expectancy of a 60 year old increased from 76 to 81 - a six year increase over 155 years which amounts to half a month per year increase in life expectancy.
Data from Social Security at
Show similar numbers over a shorter historical period where life expectancy for a 65 year increased only about 3 years between 1940 and 1990 which is about a 1.5 month per year increase - with women accounting for the lion's share of that increase. In short, life expectancy for older people is increasing quite a bit slower than for the population as a whole. Say one or 2 months per year in recent decades. This should make for a significant change in the retirement calculations. Or, am I missing something here???

Fabian Hug
Fabian Hug

Intuitively, I don't accept the concept to spend a fix amount (the withdrawal rate) on a stack of money that fluctuates. You must plan you retirement on what your portfolio can bring you during good and bad years.

peter benedek
peter benedek

Yes a very good and accessible review of approaches to withdrawal rates and their problems.

The answer, of course, is a flexible spending rate. One approach(unfortunately) buried (but hopefully not lost) in a long list of potential solutions mentioned, is the one based on a withdrawal of a fixed annual percent of current retirement assets each year.

This has been both my personal and recommended approach. A further fine tuning has recently been suggested in a Vanguard research report by Jacconetti and Kinniry entitled “A more dynamic approach to spending for investors in retirement” which includes a sensible ceiling and floor on withdrawal rates.


John Gay
John Gay

@Wade, @Michael, @Rob: Does anyone have details on the specifics of Japan's last 30 years or so in this context? ie, a 50/50 portfolio (or similar) invested in Japan's stock and bond markets (in yen) at their peak (or earlier if not enough data points) allowed a ??% withdrawal rate (in the same context as this discussion). Curious to see the outcome and how much deflation offset poor stock performance, how bonds helped (or didn't), etc. @Wade I know your research includes multiple countries but wasn't sure if you have isolated Japan (forgive me if I missed it).

Jim Haygood
Jim Haygood

Saw this article at Doug Short's website. It provides an excellent overview of the progression of thinking and research about safe withdrawal rates, that's probably unavailable anywhere else. Well done!

I faced the safe withdrawal issue while serving on the board of a nonprofit which was using a 5% withdrawal guideline on what is meant to be a perpetual endowment. I argued that in an era of low dividend yields and low interest rates, even a 4% withdrawal rate might not be sustainable.

On the return side, I believe that the traditional 50/50 balanced portfolio is outmoded. My investigations focus on the long-term results of adding alternative asset classes to boost return while reducing the dramatic drawdowns of traditional portfolios, which can amplify the dreaded sequencing risk.

Again, well done! With some amplification, this essay would make a fine book!


Yes. This is a beautiful piece of work. It explains the basis for the common rules of thumb and explains their limitations in a way that most commentators avoid - maybe out of fear of confusing the reader/client.

This could form the basis for a great investment advice program - coming up with the metrics to determine overvaluation, getting lifestyle input, etc. Even in its current form is is useful to a layman like me to get a big picture view for planning purposes: it confirmed my intuition that the assumptions typically used were not applicable without some tweaking to bring them closer to reality.

This is the best work I have seen on your website. Maybe this with some of the ideas about metrics could be the basis of an eBook - but one that should be priced on the higher end.

Larry Weber
Larry Weber


I will offer a technical critique of your article after I fully absorb the information from the article and comments from other reviewers. It may take a couple of weeks to get back to you but it is worth the time and effort. However, all I can say is wow! This article is another home run. You nailed the issues on this very important topic.

I've read a lot of the information cited in your article. In addition, I've talked to Rob Bennett several times via the phone and e-mail about the importance of his work and his out of the box thinking on this issue. I also discussed the safe withdrawal rate assumptions behind John Walter Russell's calculator with him just prior to his death. Rob encouraged me very openly to discuss the calculation assumptions with John. It is my view that Russel was ahead of his time. He was passionate about his work and cared deeply about improving safe retirement rate assumptions. He was also very much a Robert Shiller fan for good reason. It is good to see his work being discussed and refined in your blog.

I think this article will renew the interest in the safe retirement withdrawal debate. Good job!

Rob Bennett
Rob Bennett

given the general conservative time frame most planners use, it creates some "scary" results – you essentially tell clients they can spend more because there's a chance that they'll be overspending but "die" before it matters,

I think the answer is to let clients see the trade-offs that result from following various different assumptions, Michael.

The Retirement Risk Evaluator lets users plug in different numbers for how much they want remaining in their portfolio at the end of 30 years so that they can determine for themselves what the cost is for adding slack to a retirement plan. Say that you have a retirement beginning at a P/E10 level of 15 (fair value), that a 2 percent real return is available outside of stocks and that your stock allocation is 80 percent. If you are willing to have your portfolio value reduced to zero at the end of 30 years (as is assumed in the Old School studies), the SWR is 5.08.

But the calculator also permits you to see how much you would have to reduce spending to be sure of having a Year-30 Portfolio Balance of 20 percent of the initial balance or 50 percent of the initial balance or 100 percent of the initial balance. Those numbers are: (1) 4.91; (2) 4.67; and (3) 4.26.

Say that the portfolio started with $1 million. A 5.08 withdrawal rate translates into $50,000 of spending each year. A 4.67 withdrawal rate translates into spending of 47,000 each year. It is almost impossible to imagine that a portfolio that is virtually certain to retain 50 percent of its starting-point value after 30 years is going to run out of money in the lifetime of any client, no matter how fortune he happened to be in the inheriting good genes department. So what we are really learning here is that, by giving up $3,000 of spending per year, you take a retirement plan that is safe for 30 years and transform it into a retirement plan that is permanently safe.

I find this sort of thing exciting. I like being able to take concepts and reduce them into dollars-and-cents figures. This is the aspect of the Old School studies that got me interested in SWR research in the first place.

I see retirement planning as a game of making trade-offs. If you want to have this, you have to be willing to give up that. I think the future of retirement planning is to develop tools that permit clients to think through these trade-offs on their own, to decide for themselves which "goodie" (early start-date for the retirement, high safety level for the retirement, extended survival date for the retirement, etc) is most important and then to construct a plan that enhances her ability to achieve that goodie to the greatest extent possible while also achieving the other goodies to at least a reasonable extent.

It's wonderful to be hearing your voice up close and personal again, Michael. You were writing realistic stuff about safe withdrawal rates before it was cool to write realistic stuff about safe withdrawal rates!


Financialmentor moderator

@Handman If I understand your question correctly, the answer is you are withdrawing a fixed amount based on a percentage of your assets at the time of retirement then adjusted for inflation subsequently. This is without regard to market valuations at the time of the withdrawal. Hope that helps clarify.

Financialmentor moderator

 @jimsavantme Your question is quite broad in scope. In general I walk a fine balance between "trying to be prepared for possible futures" vs. messing up all my current plans with endless conjectures of what may come to pass... or may not.


I'm not going to say the idea you shared is likely or unlikely, but rather there is really nothing to plan for. If it happens then it is done and you move forward accordingly. It sucks. I disagree 100% and hope it never happens. But with that said, I only have one life and my happiness is much more correlated to my ability to accept reality as it is than to fight it. I suppose my strategy, should this come to pass, would be to figure out how to find greater happiness with fewer assets and lower investment income. Not great, but it beats every other alternative.


Some would disagree and argue for secreting assets offshore and not declaring them on IRS forms or other legally required documentation. If that is what makes them happy then go for it. It is not what makes me happy. I like the peace of mind that comes with being 100% legal and above board.


In the end, all of this is about what serves you to lead a life of maximum happiness and fulfillment. Choose accordingly.

Financialmentor moderator

 @Diamondjim Sounds like you are doing great to me, Jim! Congratulations! There is nothing wrong KISS and don't let anyone tell you different.

Financialmentor moderator

 @financialstrategist I welcome contradictory comments that add value to the discussion because I'm not the sole source of financial ideas around here, but baseless accusations serve nobody. You are welcome to expand on your opinion that I "got it all wrong" in a constructive educational way by providing solid support for an alternative viewpoint but any future comments that just troll the dialogue and say nothing of value will be deleted. Thanks.

Financialmentor moderator

 @Susan M Kuhn Thank you, Susan!


Yes, I write with the crazy assumption that people actually care enough to spend the few minutes it takes to understand the subject. All the experts tell me I'm wrong and I'm supposed to write 500-800 word articles because nobody will read detailed information like this. Foolish me, but I figure if you are going to spend a lifetime dealing with money that it is worth 10 minutes to understand what is arguably one of the most important topics affecting your financial security.The subject simply can't be covered properly in 800 words.


You are exactly who I'm writing for, and I appreciate your support and taking the time to share your thoughts! Thanks for joining the conversation!

Todd Tresidder
Todd Tresidder

@Rick - Yes, I'm confident more 3rd Generation research will be coming down the line on this topic providing alternative withdrawal scenarios. Several innovative researchers are pursuing this subject including Wade Pfau and Michael Kitces (who are following this comment thread). I will update this thread with links to future research as it is presented...

Todd Tresidder
Todd Tresidder

@Christine - Thank you! That is the ultimate compliment to say that I have "reduced highly complex research to readable form in an interesting way...". Thanks. That was my goal with this article - to make the research accessible and actionable so that it could help a lot of people. I'm really glad it came through.

Todd Tresidder
Todd Tresidder

@Steve - Thanks for catching the 3 vs. 4 error in the article. I broke one of the original steps into two separate steps and overlooked the inconsistency during editing. It's corrected now. Thanks for pointing it out.

Regarding the Rob Bennett reference, I wanted to give credit where credit is due and Rob certainly earned it for his steadfast contribution (despite controversy). This comment walks a thin line toward disparaging another person so I do want to emphasize and redirect all conversation toward expanding the knowledge about safe withdrawal rates and away from personal issues.

Again, thanks for your support of this article. I'm glad you got value from it and I appreciate your joining the conversation.

Rob Bennett
Rob Bennett

Market overvaluations" sounds an awful lot like "market timing."

Changing your stock allocation in response to big swings in valuations for the purpose of keeping your risk level roughly constant is market timing, Steve. It's long-term market timing. The reason for the errors in the Old School studies is that there was a time when people thought either that long-term market timing doesn't work or is not needed to achieve long-term success. Shiller's research has shown that it is. That's why the New School studies report that the SWR changes with changes in valuation levels.


Manish Malhotra
Manish Malhotra

Correction on the calendar years mentioned for first 20 years: it should be 2012 to 2031.


@Charles I guess another thing to look for in life expectancy is if it is heading to a specific value.  That is to ask: Is the growth rate decaying to a point where it will be zero at a certain age? Or is the growth rate a consistent value so to imply life expectancy can increase without any known boundaries.  

Wade Pfau
Wade Pfau

@Charles I don't know who you mean by everyone, but people doing research about safe withdrawal rates are looking at life expectancy from age 65 or some other chosen retirement age, not life expectancy at birth. You are surely right that life expectancy at birth is increasing much faster than life expectancy at 65 due in large part to vast reductions in infant mortality.

Todd Tresidder
Todd Tresidder

@Peter - That is the ultimate compliment when you consider my review both "good" and "accessible" given how technical the subject really is. I tried really hard to make it both interesting and accessible to "normal" readers while respecting the academic expertise involved. I appreciate any statement implying I at least partially pulled it off. Thanks.

Rob Bennett
Rob Bennett


That's a great question.

Wade and Michael are better bets at being able to help you out re Japan. I don't have the background needed to comment intelligently on this question.

My recollection is that Shiller's book ("Irrational Exuberance") has a chapter that deals with non-U.S. markets.

My general sense is that the U.S. market has been more stable for a longer period of time than most other markets. So it is easier with the U.S. market to isolate the valuations factor and thereby to draw at least tentative conclusions re its role. That doesn't mean that people should not be looking at non-U.S. markets too. There is an obvious benefit to looking at all the data that is available to us and to doing the best we can to make sense of it all.


Todd Tresidder
Todd Tresidder

@Jim - I appreciate your feedback. Good to hear from you. Thanks.

Todd Tresidder
Todd Tresidder

Larry, you constantly surprise me. We've been in communication all this time and I had not idea you were connected with Rob. I look forward to seeing any thoughts you would like to add to the discussion...

Michael Kitces
Michael Kitces

I'm a huge fan of illustrating trade-offs, and doing it in real dollar terms as you discuss here. But what you illustrated is not the concern I'm referring to. I'm referring to the impact of dynamically modeling mortality.

If we looked at mortality stochastically for a 65-year-old, the withdrawal rate in your example wouldn't be 5.08%. It would be something like 5.5% or higher. Sound untenable? It is. But since there's a 60%+ chance the client will be deceased before the end of the 30-year time horizon - and therefore deceased before the unsustainable spending actually depletes the portfolio - it "works".

So in essence, compared to the current approach - where we pick a conservative but fixed time horizon that already extends to/beyond life expectancy - dynamic mortality modeling would lead us to tell clients that they can spend more, because the amount might be UNsustainable, but there's a decent chance they'll be dead before it matters. That worries me from a planning perspective. I don't love the fixed time horizon approach either, but I'm not convinced that dynamic mortality models are better, at least not yet.

- Michael


I certainly don't mean to disparage anybody. Mr. Bennett and/or his conclusions are definitely controversial. It seems that either due to your moderator or some other factor, his opponents have not brought that controversy into the comments here. Still, I do think that it is important for the readers of this article that they be made aware of the controversial nature of some of the viewpoints espoused.

Todd Tresidder
Todd Tresidder

@Rob, Steve: I actually prefer the term "risk management" instead of "market timing" for long-term valuation type approaches as discussed here. The idea is simply that an unfavorable risk/reward ratio can be determined over a fairly long time period. There is never any anticipation or expectation of an immediate market move (i.e. market timing) when one executes this strategy. The market can continue to rise for years or fall immediately regardless of valuations. You never have any idea. The future can't be forecast. However, you do have a strong statistical indication that the distribution of returns will be unfavorable. In my world, that is very different from market timing.

With that said, I really don't think it is a good idea to have these comments drift down the abyss of efficient market hypothesis and market timing. There is no question it is related like two sides of the same coin, but I would prefer to write a similarly detailed research piece to first provide a baseline for the discussion and keep it from drifting down too many rabbit holes.

Thanks for you patience.

John Gay
John Gay

@Wade, thank you. This is great. The illustration with U.S. and Japan sustainable withdrawal rates superimposed is so useful. Thank you again!

Todd Tresidder
Todd Tresidder

@Wade - That is very gracious of you to take the time to create a custom answer to a reader's question. It's also interesting how once again the data shows periods of excess valuation result in low SWR and point toward more conservative asset allocations. It will be interesting to see what the 30 year number is when available...Thanks for putting these stats together!

Rob Bennett
Rob Bennett

It is my view that Russel was ahead of his time. He was passionate about his work and cared deeply about improving safe retirement rate assumptions.

Thanks so much for saying that, Larry. None of the work that I have done over the past nine years would have been possible without the amazing contributions of John Walter Russell.

Anyone reading these words who wants to learn about where stock investing research is headed in the next few decades should check out John's site:

The best place to get started is the "Foundations" section. The "Guidelines" section is also super.

I miss John every day. He's the kindest and smartest man I've ever known.


Rob Bennett
Rob Bennett

Thanks for clarifying your point, Michael.

I don't entirely get it and I apologize for any confusion that I caused with my earlier comment. I am grateful to you for clearing things up for those who do get it.


Todd Tresidder
Todd Tresidder

@Steve - Possibly... I'm completely aware of the prior forum threads about this subject and chosen a different path. Anybody interested in that drama can go find it in those threads - not here. My goal is to serve through quality education so anything that pulls away from that mission doesn't belong here. Comments that add to the discussion and knowledge specifically about safe withdrawal rates are welcome - whether they agree with me or not. The only criteria is they are constructive and educational (IMHO).

Todd Tresidder
Todd Tresidder

@Rob - Again, I normally don't allow linking within comments, but John Walter Russell's contribution certainly merits a link and it was an oversight in the original that it wasn't included.


  1. [...] Todd Tresidder wrote an exhaustive piece on the science and myths associates with safe withdrawal rates in retirement. [...]

  2. [...] Financial Mentor Site: “The 4% Rule Could Cause You to Leave a Fortune on the Table or Run Out of Money Long Before You Die” Published in October 24th, 2011 Posted by Rob in Community, Experts, Goons, SWRs Todd Tresidder has written a super article about the evolution of our understanding of how safe withdrawal rates work at his Financial Mentor site. It is called Are Safe Withdrawal Rates Really Safe? [...]

  3. [...] Tresidder at Financial Mentor submitted an in depth examination of “Are Safe Withdrawal Rates in Retirement Really Safe?” Rob Bennett says “This article is something special. My guess is people will be taking about this [...]

  4. [...] Based on Market Valuation I recently came across an excellent article by Todd Tresidder discussing the many imperfections of the 4% withdrawal “rule” for [...]

  5. [...] of safe withdrawal rates. Todd Tresidder thoroughly examines whether we can rely on the traditional safe withdrawal rate of 4%.If you’ve recently been through a bankruptcy or know someone who has, check out this guest [...]

  6. [...] buying royalty rights from others. Did you know that you can own shares in popular songs of history?Are Safe Withdrawal Rates Really Safe? [Financial Mentor] If you’re unfamiliar with safe withdrawal rates but plan on retiring [...]

  7. Best of Blogs – Have you been to Whitehorse? | Retire Happy Blog says:

    [...] Many of us have been told about ‘safe withdrawal rates’ when it comes to retirement funds.  The Financial Mentor delved into the question Are Safe Withdrawal Rates Really Safe? [...]

  8. [...] Are ‘safe’ withdrawal rates really safe? ( [...]

  9. [...] Are Safe Withdrawal Rates Really Safe? – This article by Financial Advisor and Coach Todd Tresidder on his website provides a strong overview of the history of safe withdrawal rates, through its various iterations from the work of Bengen to recent work by Pfau. Overall, it provides a pretty strong summary of the research, and some of its valid criticisms and concerns.  [...]

  10. [...] Are Safe Withdrawal Rates Really Safe? @ Financial Mentor [...]

  11. [...] or hard to comprehend. That is why I recently compiled a readable summary of the latest findings in safe withdrawal rate research here so that you can understand these [...]

  12. [...] most of the assumptions used to model that don’t really apply in real life (for full details, you can read this long, but excellent article).  The real truth of the matter is that a ‘safe withdrawal rate’ isn’t a [...]

  13. [...] Tresidder from Financial Mentor published Are Safe Withdrawal Rates Really Safe?, saying, ‘In this epic 8000 word post Todd fully reveals the controversy surrounding safe [...]

  14. [...] Tresidder of Financial Mentor subtitles his site as “Financial Freedom For Smart People”. We’ll admit we [...]

  15. FMF March Money Madness, Round 2, Posts 13-16 | FavStocks says:

    [...] Are Safe Withdrawal Rates Really Safe? (Safe) – An epic 8000 word expose that puts the multi-year controversy behind safe withdrawal rates to rest once and for all. This subject is arguably the most important topic in retirement planning, and the analysis in this post earned top rankings in Google and praise from pension managers and top researchers alike (see comments!). Discover what ugly truths lurk behind the infamous “4% Rule” and how it might effect your financial security… [...]

  16. [...] Are Safe Withdrawal Rates Really Safe?, at Financial Mentor (A great survey of the first ten years of New School SWR [...]

  17. […] Tresidder’s writing over at The Financial Mentor.  Todd has already written up a post on the history of safe withdrawal rates, and you would be poorly served by me duplicating the tremendous effort he […]

  18. […] oft-quoted rule of thumb of a ‘safe withdrawal rate’ is 4% – a number which likely isn’t completely safe.  So, let’s try a few methods to figure out what sort of cash flow you could […]

  19. […] For 4% – Mr. Money Mustache – For 4% – JLCollinsNH – Against 4% – Financial Mentor – Against 4% – Schof Image: […]

189 Flares Twitter 141 Facebook 31 Google+ 8 LinkedIn 7 Email -- Pin It Share 2 189 Flares ×