The idea for this article comes from a question submitted by a reader. I’ve copied the question below, and omitted the names to protect the innocent. (No animals were harmed in the creation of this article.)
Let me start by saying, I totally enjoy reading your articles/post! Do you have any information on retiring with 100% of your money in the S&P 500? I’m sure there are thousands of scenarios but I’m curious as to a specific one…. Let’s say one decides to retire in a bull market at 60. (Will not retire in a bear market) The only income will be a NON tax-deferred investment account. (Not an IRA) Lets say the value is 4 million. Is it realistic to withdraw 5% ($200,000) each year? 2% of the money would come from dividends. And in years that are negative, just withdraw a couple percent less.
What is Involved?
Essentially what the reader is asking is if you have 100% of your portfolio in the S&P 500 index, in a taxable account, and use a variable withdraw strategy starting with 5%, can you reasonably expect that to last throughout retirement?
Let’s start this discussion by separating out the issues that this question presents. There are several factors that affect how this plays out. I’ll address each one as we go through the discussion.
Asset Allocation is the single largest driver of investment performance. Your asset allocation will also drive how your portfolio behaves through different phases of the market cycle. On the surface level, we typically think of asset allocation as the ratio of stock to bonds within a portfolio.
For example, a 70/30 portfolio would have 70% of the money in stocks, and 30% in bonds.
As you adjust those percentages the expected return and behavior (volatility) of the portfolio changes. The more you have in stocks, the greater you can expect your return to be over the long term. That is coupled with an increase in volatility, meaning your portfolio will fluctuate more.
As an example consider a 50/50 portfolio and a 80/20 portfolio. The 80/20 portfolio would have a higher expected return and also fluctuate in value more. During a bull market you would expect the 80/20 portfolio to perform much better than the 50/50. The downside to that is when the market does poorly, you would expect it to perform much worse.
A portfolio that is invested completely in the S&P 500 is 100% in stocks, or 100/0. On the asset allocation spectrum this would have the highest expected long-term return, but also the maximum volatility. You would very much consider this to be an aggressive portfolio. Most retirees would not be comfortable with that level of risk, and could benefit from some diversification into bonds anyway.
Asset allocation is also a significant factor in how sequence of return risk affects you…
Sequence of Return Risk
The question also addresses sequence of return, a major retirement risk. The sequence of return is exactly what it sounds like; the order in which you experience investment returns. We usually think of this in terms of years going from one to the next.
For example, let’s say your portfolio returns 8% the first year and 23% the next. That’s one sequence. If your portfolio instead returns 23% the first year and 8% the next, that’s a different sequence. Same returns of course, just in a different order.
So why does the sequence of returns matter? If you weren’t taking withdrawals or contributing to the portfolio then the sequence wouldn’t matter. But, because you ARE taking withdrawals from the portfolio, the sequence matters a great deal.
When you experience really bad markets either shortly before or shortly after retirement the negative effect is exceptionally bad. A significant market decline that occurs during this period can take YEARS off the life of your portfolio.
In the context of the question we are answering, timing retirement so you don’t retire in a bear market is an excellent way to mitigate this risk. Just realize not everyone will have that option as you may be forced into retirement due to unforeseen health problems or a layoff. You may also retire in a bull market but then be hit with a bear market the very next year.
Non Tax-Deferred Account
Another issue at play here is that the entire portfolio is in a taxable account.
It’s likely that a good amount of your retirement savings is in a tax advantaged account like an IRA or 401k. The benefit of retirement accounts is that they allow you to save for retirement, buy and sell financial assets, and then withdraw the money in some tax-advantaged way.
Generally, you’ll either pay income taxes on money contributed on the front end (Roth) or on the back-end (tax-differed). The transactions you do in-between those points will typically be shielded from taxes as long as you follow some simple rules, which most people do.
A taxable account doesn’t provide that same tax shelter. Any capital gains generated along the way, interest earned, and dividends received all create immediate tax consequences. The transactions are what create taxable events in a taxable account, unlike retirement accounts. Retirement accounts are generally only taxed on contributions or withdrawals.
Because of the complexity of taxable events within a portfolio, and the fact that individuals will buy and sell shares at different times and in different amounts, there is no way to calculate one universal value to account for portfolio taxation. Not to mention the progressive nature of both income and capital gains tax brackets. Since retirees are taxed at different rates, they would be taxed differently even if they made the same transactions within a taxable account.
The effect of this “as-you-go” taxation is called tax drag because it drags your portfolio and slows your investment growth. The simplest way to incorporate this is to subtract some amount from the growth rate of your portfolio.
For our example, I’ll use a .5% tax drag. This is a realistic value, but I want to stress that the tax drag will be different for everyone depending on how you manage your portfolio and which tax brackets you fall into. Yours could be anything from nearly 0% to a few percent. Tax management becomes extremely important when you have more of your savings in taxable accounts.
To see the effect, assume a 10% return on the portfolio and a .5% tax drag. The net growth on the portfolio is 9.5%. That doesn’t seem like much but remember the power of compound growth over such a long time period. 1/2% on a $4 million portfolio compounded for 30 years is just shy of $650,000. It’s serious.
You likely have a mix of accounts to include taxable, tax-deferred, and after-tax (Roth) accounts. Segmenting your withdrawal to partially come from each account can help smooth your tax liability throughout retirement, and is often more tax efficient than simply withdrawing all the money from one account first.
Variable Withdrawal Rate
This question also suggests a variable withdrawal rate which can be another excellent way to mitigate the effects of a bad sequence of return. While the question suggests a somewhat ad hoc method, I think a formal process for deciding adjustments beforehand is useful. If you make big decisions like that without a process there’s a much greater chance that you’ll make sub-optimal choices due to stress, fear, anxiety, or even exuberance.
I really like the Guyton-Klinger variable withdrawal strategy. With this strategy you start with a given withdrawal rate, such as the 5% mentioned, and then adjust that withdrawal when you drift outside of predetermined “guardrails” which are upper and lower limits on the current withdrawal rate.
In simple terms, you start with a given withdrawal rate and then adjust the dollar amount for inflation each year just like you would using the standard 4% rule. However, if your portfolio tanks and the CURRENT withdrawal rate shoots up to an unreasonable level you simply take a smaller withdrawal. Guyton and Klinger suggest a 10% reduction in the dollar amount you withdraw anytime your current withdrawal rate is 20% higher than the one you started with.
The upper guardrail is 6% since 20% of 5% of is 1%. Add the 1 back to 5 for 6%. The same is true for the lower guardrail. Subtract 1 from 5 to get 4%.
You would decrease your withdrawal any time the dollar amount is more than 6% of your current savings. You would also increase your withdrawal any time the dollar is less than 3%.
Applied to our example….
If we start with $4 million and take a 5% distribution we will withdraw $200,000 in the first year. Now, assume the first year of retirement we go into a severe bear market and our portfolio value falls to $3 million.
If in the next year if we were to take another $200,000 withdrawal from $3 million that would be a current withdrawal rate of ($200,000/$3 million) = 6.7%. That would mean we have gone outside the upper guardrail. So, instead of withdrawing $200,000 (which we didn’t even adjust for inflation by the way) we would REDUCE that amount by 10%, or $20,000.
Our withdrawal in the next year would be $180,000.
What I like most about this withdrawal strategy is that it gives you a formalized way to make small adjustments in your retirement along they way. Doing this could add years to your portfolio and prevent you from being forced into a position to make large and dramatic changes late in the game.
Does it Work?
Now that we have properly framed the question and addressed the key elements involved we can look at how this would have worked out using historical data. For the returns on the S&P 500 I’m going to use the Large-Cap returns for each year as reported in the Ibbotson SBBI Yearbook. This data set is widely used in academic and professional research.
We will look at how the portfolio fared during each decade, and then specifically look at the effect of retiring during a bear market. This comparison will allow us to see the benefit of NOT retiring during a bear market, or the cost of retiring in a bear market depending on your perspective.
I’ll also do each analysis using a straight 5% withdrawal vs using a variable 5% withdrawal. Again, the comparison will let us see the pros and cons of each distribution strategy.
How Long Did the Portfolio Last with Straight Withdrawals?
Going back to 1950 I assumed you begin each decade with $4,000,000 in your retirement account, and used a straight 5% with no adjustments other than for inflation. I also subtract the 1/2% each year to account for the tax drag.
Starting in 1950, you would have easily sustained your 5% withdrawal. In fact, because the first few years of the 1950’s produced such strong equity gains your portfolio growth would have far out-paced the inflation adjustment on your withdrawals.
Large-cap stock returns for 1950 – 1955 were approximately: 32%, 24%, 18%, -1%, 53%, and 32%. By the end of 1955 you would have a nearly $12 million dollar retirement account. Your withdrawal, adjusted for inflation, in 1956 would be $227, 072 and would be less than 2% of your portfolio. By 1966 your inflation-adjusted withdrawal of $269,000 would have been less than 1% of your portfolio total of over $24 million.
Assuming a 30-year retirement you would have $57 million left in your account.
Clearly, 1950 was a good year to retire.
Fast forward to 1960. While this would not have been a tragic decade to retire, you would have been walking a much tighter path. Your portfolio would have only grown by 1/2% in 1960, 27% in 1961, and then lost nearly 9% in 1962. Your portfolio would have roughly kept pace with the inflation adjustments through 1972. While that sounds good, ideally your portfolio would be growing in the early years so that you could weather the inevitable bear markets in the future.
That’s exactly what happened in the 1950’s scenario, and why it looks so good.
In this scenario though, your portfolio growth doesn’t outpace your inflation-adjustments in the early years. Then, in 1973 and 1974 you run into -15% and -26% returns. On top of that, inflation in those years was about 9% and 12% respectively.
Even though the market would have given you 37% in 1975, your portfolio wouldn’t recover. You would have ended 1975 with just over $4 million. Realize though, that by that time the inflation adjustments of the last 15 years have brought you up to a $353,000 withdrawal. Your withdrawal as a percentage of your current portfolio balance is nearly 11%.
Your portfolio would have been depleted -gone- 30 years after you retired. That’s actually pretty ideal for a lot of people. Having enough money for retirement, and then your last check bounces. Jokes on them.
The trick of course is that you can’t plan this precisely into the future. Only with hindsight. Most people would be incredibly stressed to cut it so close
What a difference a decade makes. If you had retired in 1950 vs 1960, you would have still had $57 million after 30 years, as opposed to nothing.
Same story, third verse. A little bit louder and a whole lot worse!
Now let’s roll forward another decade. You may see what’s coming if you’re well-read on sequence of return risk.
Those rough years of 1973 and 1974? Those now hit you right out of the gate. By your tenth year of retirement you would have been withdrawing about 14% of your portfolio. The portfolio would not have lasted through 1989. This would have given you almost 20 years of withdrawals.
So the same retirement withdrawal strategy, with three different start points, produced three VERY different results.
This was another good year to retire, and one in which this particular distribution plan would have worked out well for you. Thirty years later in 2009 you would have over $42 million and your withdrawal for 2010 would be slightly over $500,000 and 1% of your portfolio. At no point in your 30 years of retirement would you have even come close to depleting your savings.
How Long Did the Portfolio Last with Variable Withdrawals?
Now let’s look at starting retirement in each decade just like before. This time, however, we will follow the variable withdrawal strategy using guardrails.
There are a few important distinctions to note here.
First, we will repeat each scenario listed above. You’ll notice that in each scenario the portfolio would have lasted thirty years using a variable withdrawal rate. That’s a BIG improvement compared to the 1970 scenario.
Second, the value of your withdrawals won’t necessarily keep up with inflation. Remember though, with a straight inflation-adjusted withdrawal you run the risk of inflation out-pacing your portfolio growth. The variable withdrawal strategy specifically counteracts that by reducing your withdrawal before it’s too late. Then too, this simply may not be a bad thing for you. Many retirees will naturally spend less later in retirement as they age and slow down.
Third, your ending balance that you leave to heirs may very well be smaller. That’s not necessarily a bad thing. With a variable withdrawal strategy you may also adjust your withdrawal upward by MORE THAN inflation if your portfolio does really well. This will allow you to enjoy more of your savings in retirement, as opposed to leaving a larger inheritance.
Now, let’s take a look…
As we saw before, this was a good year to retire. As you might expect, you would have INCREASED your withdrawal by 10% sixteen times over a thirty-year retirement. You would have only decreased it by 10% four times, and that was after most of the increases. That matters because it means those decreases were on an already expanded withdrawal.
By your third year of retirement you would withdraw more dollars with the variable withdrawal strategy each year than you would with a straight inflation-adjustment.
By the late 70’s you would have been withdrawing $1 million annually and still had over $20 million in your account after thirty years.
Through a thirty-year retirement you would have withdrawn over $20 million total, compared to the $9 million total you would have withdrawn with only inflation-adjustments. That is a very significant difference in retirement lifestyle.
A variable withdrawal starting in 1960 would have left you with over $10 million after thirty years of retirement. How?
You would have hit the guardrail and reduced your withdrawal five times. Understand this does mean you have less to spend each year, but the trade-off is your portfolio lasts longer. It may also help you maintain your sanity if you want a little more cushion on your nest egg.
Let’s do a comparison to see how much your retirement budget you would have cut back. For example, take 1982. With inflation-adjustments only your withdrawal would have been about $640,000.
Under a variable withdrawal plan? About $280,00. That is just shy of a 60% difference. That isn’t trivial. Again though, this is keeping you from going completely broke during a time when there wouldn’t be much you could do about it.
With variable withdrawals you would still have had $12 million dollars at the end of 1989 instead of the zero dollars you had before. So how much did you have to cut back during retirement to make that happen?
You would have triggered the 10% cut four times.
In 1988, the last year you could take a full withdrawal under the straight inflation strategy, you would have withdrawn just over $600,000. With variable withdrawals? About $320,000.
The big plus here is that there is still a $350,000 withdrawal to be had in 1989 and money to take you through another decade or longer.
Again, a great time to retire.
This time, instead of leaving $42 million to your heirs you would take a 10% withdrawal increase through every single year of the 90’s. You would spend nearly $22.5 million over the course of retirement, up from $10.5 million compared to taking inflation-adjusted withdrawals only, and leave about $26 million to your heirs.
What About Retiring Right Before a Bear Market?
We’ve clearly seen that sequence of return risk plays a huge role in retirement. But, we’ve clearly seen it with the power of hindsight. What we can’t see is the sequence of future returns. So stepping back into the data, lets assume we lacked that same foresight then.
What would happen if you retired just before a bear market, not knowing that it was coming? We will look at the 1973-1974 period, as well as the early 200’s to see.
Retired in 1973
If you retired at the beginning of 1973 then 1973-1974 would fall at the worst possible point when the sequence of return affects you the most. Your inflation-adjusted withdrawal in 1975 would already be 11% of your portfolio value of just over $2.6 million.
You would have depleted your portfolio in 15 years.
What about a variable withdrawal? By 1979 you would have cut your withdrawal by 25% of its starting value to about $150,000. The real pay cut is even larger when you consider inflation averaged over 8% during that time. Still, you would have over $3 million in savings and the portfolio would still not have been depleted in the next two decades.
Retired in 2000
If you retired at the beginning of 2000 your first three years of retirement would have given you double-digit losses of roughly -9%, -12%, and -22%.
Inflation-adjusted withdrawals would have lasted 15 years.
Variable withdrawals would not have depleted your savings. However, you would have cut your withdrawal in half by 2011, while inflation averaged just over 4% per year.
Again, that is a substantial cut in real spending power, but it comes with the benefit of at least having some income in the future.
Historically, this plan would have worked in several cases. However, the risk of failure needs to be seriously considered. Three risks to this strategy that are the most glaring to me are:
- There are several occasions in which the plan didn’t work. In most situations high probabilities of success are considered good enough. One of the things that makes retirement withdrawal planning so important is that a failure is so severe. If you become destitute at 75, there isn’t a whole lot you can do about it. You can largely overcome this with a willingness to adjust your budget as you go and make cutbacks when necessary as we have seen.
- The unknown future – Retiring during a bull market is certainly better than retiring in a bear market. But what about just before a bear market? The current market is knowable. The future market is not. Suppose you decide to retire right now, this instant. Are the next few years going to look more like 73-74 or the 90’s? Neither you or I know. That’s not a criticism of this particular strategy. That’s a simple reality of retirement planning and needs to be addressed with any plan.
- Behavior and emotion – An all-stock portfolio is going to be very volatile. Large, double-digit declines are hard to stomach when you are now counting on the money to buy food, medicine, and electricity. Most retirees would have a hard time not bailing out during the worst possible time and losing a lot of money.
You can significantly improve this plan by cutting back on equity exposure in the years surrounding retirement. Being all-equity during this period leaves you completely exposed to sequence of return risk. That’s like stepping into the ring with a world-class boxer, calling him a name, and then holding your chin out. He may not hit you, but you’re done if he does.