Retirement Withdrawal Strategy

Converting Traditional IRA's and 401k's to a ROTH IRA may make sense from a perspective that is often not considered.

If married, there is a likelihood that one spouse may not survive the other leaving the survivor with a "single status" tax bracket someday. Since the single tax bracket rate is higher, there is a possibility that the survivor will be paying more in taxes - even if on less income.

It may also mean that ROTH conversions that would have been advantageous when married no longer make sense. Staying in the 15% tax bracket if married is easier than trying to stay in the 15% bracket as a single filer.

ROTH IRA

If you are in the 25% Federal tax bracket, it may make sense to contribute to your employer's plan first (403b, 401k, 457, TSP ,etc.) to take advantage of this known tax savings, including state tax rates.

But it is important at some point in your accumulation goals towards retirement to have funds in a ROTH IRA. These ROTH IRA funds are, today, tax-deferred and tax-free at withdrawal. 

As Larry Burkett, a financial planner and author, stated over 30 years ago, just remember that those IRA's belong to the government even though they have your name on them. A future Congress can change the rules on us.

Congress could decide to implement a VAT tax or a Federal sales tax or other consumption-type of tax (or even income means-testing), then income distributions from those ROTHS in retirement would get taxed when spent.

Today, the ROTH still escapes the 1040 altogether and therefore there is no income tax. That means that those withdrawals do not impact the taxability of social security benefits either. This is another good reason to build flexibility into your retirement plans.

However, the income thresholds for when social security is taxable were set in the 1990's and never indexed to inflation. 25+ years later, many retirees are going to be faced with up to 85% of their social security checks taxed whether they have ROTH IRA's or not.

In 1993 the thresholds were set and have not changed:

Up to 50% of social security is taxed if single with $25,000 of income; $32,000 if married
Up to 85% of social security is taxed if single with $34,000 of income; $44,000 if married

Ray Lucia

Ray Lucia, author of Buckets of Money and several other books, has been battling the SEC over his misleading presentations on this strategy.

I recommend this particular book (for its general concepts/strategy) on my blog site and feel a need to let any readers here know about these proceedings.

On September 5, 2012, "...The Securities and Exchange Commission today charged a nationally syndicated radio personality and financial advice author for spreading misleading information about his “Buckets of Money” strategy at a series of investment seminars that he and his company hosted for potential clients...."

If you would like to read more about this at the SEC site, here is the link:

http://www.sec.gov/News/PressRelease/Detail/PressRelease/1365171484416#.UvYkeXi9K0d

More importantly, on December 6, 2013, the U.S. Securities and Exchange Commission posted an Initial Decision on Remand that begins as follows:

"...This Initial Decision on Remand supplements the July 8, 2013, Initial Decision in this proceeding, confirms that Respondent Raymond J. Lucia Companies, Inc. (RJLC), violated Sections 206(1), 206(2), and 206(4) of the Investment Advisers Act of 1940 (Advisers Act) by misrepresenting the validity of purported backtesting in seminars for prospective investors..." 

If you would like to read the entire case, here is the link:

http://www.sec.gov/alj/aljdec/2013/id540ce.pdf

In a recent (2014) Journal of Financial Planning article, a survey was cited that about 28% of financial advisors use a time-based segmentation strategy - similar to Ray Lucia's concepts. That is, of placing several years of investments in safe investments (cash, for example), then another set of investments to cover a few more years. Finally, a portion is invested in stocks that would, if needed, be allowed to grow through more difficult times while the safer investments are used up.

The strategy of time-based segmentation is not in question and I, personally, like the approach as it is easy to understand and allows the retiree, especially, a little breathing room when markets are down.

Withdrawal Strategies

Jonathan Guyton, CFP®, in a Journal of Financial Planning October 2011 article entitled Mirror, Mirror, on the Wall highlighted a survey that looked at income distribution strategies. He cites three:

(1) Structured Systematic Withdrawals - distributions from the total portfolio which is then rebalanced

(2) Time-Based Segmentation - basically a bucket approach where the low-risk (cash equivalents) are drawn down during markets when the equity portion has a low value (Mr. Guyton actually describes this system a little differently where the nearest-term horizon needs are taken from the lowest-risk pool). 

(3) Essential-Versus-Discretionary Income - higher risk securities (stocks, for example) fund the discretionary desires and lower risk securities (cash/fixed income/annuities) fund the essential needs

#2, I prefer, because it provides a cushion to "...outlast most equity bear markets - most, but not all..." as Jonathan Guyton notes. There is a risk that the safe bucket gets exhausted though.

In his conclusion, he acknowledges that "...a thoughtful, policy-based rebalancing method would easily mitigate this risk..."

 

Monte Carlo Methods

Monte Carlo statistical method is a computer simulation of hundreds or even thousands of sequences of returns with variables for inflation rates, interest rates, growth rates and other metrics.

As mentioned by William F. Bengen in his book, "Conserving Client Portfolios During Retirement" about Monte Carlo simulations he states, "...they have been randomly generated from a set of mathematical rules and do not correspond to any actual historical data..." and this leads actually to "...lower success rates for withdrawal rates below about 5.25 percent, and higher probability scores for withdrawal rates above 5.25 percent..."

Since many advisors use withdrawals rates of 4-5%, then using Monte Carlo simulations may suggest that you have a starting "nest egg number" that is higher than you really may need in order to reach a 95% success rate where you do not run out of money.

The software is just not capable of providing the best answer - yet. But as Mr. Bengen states in his book, "...I have learned never to underestimate the ingenuity of man when he attacks a problem of importance to him..." 

Longevity Risk

Another retirement risk that competes with sequence risk is longevity risk - the risk of outliving your money. For sequence risk, being conservative helps minimize this risk but for longevity risk, being more aggressive helps minimize this risk.

The answer: there is a balance needed in your investments.

And there are also probabilities to consider, too. The likelihood that you will need income from your investments when you start retirement are higher than the probability that you will live to 100 or beyond.

Balance is an important element of balancing risk in your portfolio.

Sequence Risk

Sequence risk doesn't matter as much when working and building a portfolio as it does when you start to take money out of your investments during retirement. When adding money, it is called dollar-cost averaging and there is always the risk that stocks will continue to go up. Then the positive sequence of returns may not be as beneficial as compared to a lump-sum investment. If you have the money and the choice, this sequence risk is something to consider.

When withdrawing money, a negative sequence of returns may damage your portfolio value as you take money from a smaller pool of funds (think: a bear market like late 2007- early 2009). Minimize this risk by having "some" years of your retirement income needs in safe money (money markets, CD's) to last 2-5 years, so market risk investments don't need to be touched.

The Effect of Diversification on Withdrawals from Retirement Funds

Based on some number-crunching by Craig L. Israelsen, Phd, a $500,000 portfolio withstood 5% withdrawals ($25,000 at the start of the period and then increased annually by 3%) for the tested 25-year distributions during retirement.

Seventeen periods were analyzed: 1970-1994 through 1986-2010

Of-course "when you start" in comparison to when major bear markets surface has a big effect on whether you end up with $2 million (1970 start) or $7.8 million (1975 start, just after a major bear market ended). What a range, but you still end up with more than the $500,000 when you started.

The key though is his last sentence, "...diversification is a lifelong investing imperative..."

He says this because keeping 60% in stocks and 40% in bonds makes a tremendous difference over playing it safe with 100% in bonds (millions in almost every case), and investing equally in at least seven asset classes results in even higher ending values. Higher values become important over time because, otherwise, with no growth in the portfolio, those withdrawals become much higher than 5% as your years in retirement continue.

Probability of Failure in Retirement Distributions

Often financial planners will use a fixed period (25-30 years) for determining the length of time that retirement distributions must last. Based on a Journal of Financial Planning reported study by Blanchett in the December 2008 issue, the withdrawal rate and life expectancy have dramatic effects on the amount of money you can take out of your investments during retirement.



All studies are based on 60% stock and 40% fixed income portfolio, so if you plan to have less than 60% of your portfolio in stocks at retirement then these numbers may not be valid.



If you take a 2% withdrawal rate, there is a 0% chance of running out of money from 10-40 years based on historical data (and we have plenty of rolling periods to look at that included the Great Depression and world wars and other difficult periods of history).



A 4% withdrawal rate, succeeded 100% of the time up to only 15 years. 92.5% of the time up to 35 years based on historical data.



But to reduce the probability of failure you could use joint life expectancy rather than a fixed period of time (these studies used age 100 or 35 years from age 65). Using a fixed period "...overstates the probability of the portfolio failing..." during a couple's lifetime (which is actuarily shorter than a fixed period to age 100, for example).



The withdrawal rates increase by 1% to 2% (4% becomes 5-6%) under the assumption that either one or both spouses are still alive before the portfolio runs out of money.

If a planner is paid for assets under management, then it is in "the planner's best interest" to use a longer distribution period but NAPFA planners are fiduciaries and will have "your best interests" in play before their own. A planner that has no assets under management (like myself) can be objective since our fees are fixed.

Ineffective Strategies for Retirement

Ineffective, or at least, inefficient strategies to withdraw money from your investable assets during retirement are keeping 100% of your "nest egg" in stocks (obviously) but also keeping 100% of your "nest egg" in CD'S, money markets, bonds (even if TIPS and I-Bonds and other inflation-protected securities are used) and other fixed income vehicles.

Study after study has proven that an allocation to stocks of at least 25% and, preferably, 40% or more to stocks are required when withdrawing an income from your portfolio.

The less the percentage in stocks, the less the safe withdrawal rate becomes going to as low as 2.4% if you want no stock exposure and want your money to last 30 years. If you want that money to last 40 years then 1.8% becomes the safe withdrawal rate.

The Infamous 4% Withdrawal Rate

According to studies performed by Michael Kitces and reported in the Financial Planning Journal:

A safe withdrawal rate for a 30-year period (meaning you don't run out of money) may be:

4.5% in an over-valued market (think year 2000)

5.0% in a fairly-value market (much of the time)

6.5% in a severely under-valued market (after a market crash like March 2009)

In addition, the age when you start withdrawals also has an effect:

Age 65, you may be able to withdraw 4-4.5%

Age 75, you may be able to withdraw 6-6.5%

However, these studies are based on historical data and, much more importantly, on keeping a hefty portion of your portfolio in stocks (most studies assume an allocation of 50%/50% or even 60% stocks and 40% fixed income). If you think you can withdraw 5-6% per year keeping most of your investable assets in CD's and money markets you may not be able to keep up with inflation.

In 10-15 years, when you may need to withdraw twice as much as you are at the start of retirement now you just may not have the flexibility to do so.

How Best to Withdraw Your Nest Egg

The Buckets of Money strategy authored and designed by Ray Lucia, CFP (r) is further substantiated by scientific studies that show that withdrawing from the lowest expected return assets (that is, the "safe" money that you have in CD's, money markets and short-term bonds) is the best methodolgy.

From Mr. Weigand and Mr. Irons in a study in the November 2008 Journal of Financial Planning, "...a strategy of consuming bond wealth first would have beaten the 50/50 rebalancing strategy and the stocks-first strategy about 90 percent of the time..."

"...The longevity advantage of bonds first portfolios increases when the spread of stocks' long-term earnings yield over bond yields is larger in the first year of retirememt..."

Today that yield spread is less than 1 percent (historic lows). Their findings also suggest that lifestyle strategies that decrease stock proportions as investors grow older may not be appropriate. This strategy does lower risk, however, so it should be reviewed closely before you plan to retire.

How Much Do You Need in Retirement?

"Rules-of-Thumb" can help simplify a complex issue. The best answer to how much you will need in retirement is obtained by working with a financial planner.

The Rule of 72 states that: An amount will double by taking 72 divided by the number of years and then will equal the approximate interest rate. What?

So, for example, at 3% your "number" doubles in 24 years. (72/3=24 and 72/24=3%)

This may be good if you are watching your money grow but may be bad if you are spending your money. For example:

In retirement, if you need your investments (from your IRA's and 401k, TSP, 403b, 457 or whatever) to supplement your pension and/or social security, then whatever number you choose, it may need to be double that in 24 years.

If you are 50-60 and need $20,000 per year, then when you are 74-84 you may need as much as $40,000 per year. Obtaining a 4% withdrawal rate from $500,000 at age 50-60 is reasonable but are you prepared to take an 8% (double) withdrawal rate from whatever is left in 24 years? This is why moving all of your funds to safety of principal is not a strategy that works over the long retirement periods many of us will experience.

Distribution Withdrawal Rates

A safe withdrawal rate is often cited as 4% of your investable asset's balance. It helps answer the question: "Will we outlive our money?"



However, another question needs to be answered: "Will we be able to enjoy our money?"



4% is a rate beginning at normal retirement age around the mid-sixties (62-67). It is then indexed to inflation so that this withdrawal rate at age 75 would be closer to 5.5%. Advisors who manage your assets, however, have an interest in you withdrawing less and may encourage you to withdraw less. May be wise, maybe not. If you have safe money set aside, then the withdrawals may be able to continue increasing. Don't accept "rules-of-thumb".



More so, recent studies show that after a severe market downturn, the safe starting rate may be as high as 6.5%. How would that affect your lifestyle in retirement?



In order to determine the right rate you have to decide what income need is required for you and what inheritance you desire to leave for heirs, if any. In addition, the asset mix that you choose (how much in stocks versus how much in bonds) will determine your expected future return rate on your investments and this affects the withdrawal rate, too.

Most importantly, those withdrawals should come from the lowest expected return asset categories so that you can allow your long-term investments to be just that - long-term, where you are reinvesting dividends and interest income.

Longevity Insurance

Longevity Insurance - Business Week Article

... you'll be able to withdraw a greater percentage of your savings earlier in retirement than would otherwise be prudent—some 7% a year, versus 4%, says Shane Chalke, CEO of Seven Squared Financial of Middleburg, Va., which is developing annuity products.

Longevity insurance also lessens one of the biggest problems associated with conventional annuities: the fact that, once you sign over your money, you (or your heirs) can't get it back, even if you die before collecting a dime.

Make sure you're comparing annuities on an apples-to-apples basis.

"Buy when you retire,"

Figure out how much of your essential expenses you can cover with Social Security, pensions, and other forms of guaranteed income—and consider buying coverage for the rest. Don't put too much into this basket. "Typically, if you allocate 10% to 15% of your portfolio to this, it will give you about two-thirds of the benefit you'd get if you were to annuitize your entire portfolio," says Scott. All told, it's a good deal, provided you live long enough to collect.

Layered Cake

Maybe Mr. William (Bill) Bengen read my letter in the FPA Journal. When asked by the reporters at MarketWatch newsletter about his concept of "layered cakes" and withdrawing a higher percentage than 4% (like 4.12% or 4.56%) in retirement, he replied:

"Using a year-by-year forecast, you build a model of retirement going out to the end of lives," Bengen said. "It's nice to come up with a number like 4.5%, but what does it mean if you're not looking at the rest of the client's situation?"

BINGO ! (Exactly what I was saying)

My letter to the editor on Retirement Planning Distributions

October 11, 2006

Dear Editors of the FPA Journal,

I respect the stochastic (Monte Carlo) approach as a tool in financial planning and respect the many fine contributions over the past 13 plus years of Mr. Bengen and as the author of the article entitled “Baking a Withdrawal Plan ‘Layered Cake’ for Your Retirement Clients”. Here comes the “but” though and I mean this respectfully. Mr. William P. Bengen’s approach to a client’s retirement portfolio withdrawal rates is not like “baking a layered cake” but rather like “reaching for a pie in the sky”.

Even, Mr. Bengen, admits this when he states in one of his examples that “…an advisor who actively managed portfolios and exceeded the returns…by two full percentage points” and this over a 30-year time horizon is, in his own words in the article “…a hypothetical construct of-course.” And, again, in his conclusions states that “…any attempt to project a high degree of precision in these calculations promises more than any professional can reasonably expect to deliver”.

I agree with those statements. What is the point then of illustrating bar charts and ‘layered cakes’ with withdrawal rates to two (2) decimals? Is there a difference between a 4.15% withdrawal rate and a 4.42% withdrawal rate? Really? No, I disagree here. It is dangerous to use Monte Carlo simulations for anything more than a tool to approach a concept and very misleading to use this stochastic method as a practical solution to the problem of withdrawal rates in retirement.

Monte Carlo has several limitations that I rarely see mentioned and more so when applied to the distribution phase than to the accumulation phase of one’s life. If, for example, a client has been given an 80-95% success rate for reaching a particular number to start retirement, then overshooting the mark may not necessarily be so bad if there is balance in one’s life (lifestyle spending choices versus savings/investing). Besides, there is still retirement to enjoy some of those over-savings. However, in the distribution phase of retirement, an 80-95% success rate may mean that the client ends up with the same amount of money or even more money than they started with after the retirement period of 25-30 years. If the client wanted to be close to zero, then this is an 80-95% success rate for the planner who is managing the money but may have meant serious cutbacks in lifestyle for the retiree. This success rate is misleading at best.

Another point for all planners to keep in mind who use Monte Carlo simulations is that the input variables are themselves suspect. Returns, correlations, standard deviations, beta, interest rates and all of the other inputs are based on historical data and cannot, at this point in its development (and maybe never), be able to project how these variables will change in the future. 10,000 iterations is meaningless, too, because many of those iterations are going to provide back-to-back years of returns that are highly improbable. For example, who would expect that a client would retire and then suffer through a 1929 depression-like return, followed by a 1973-1974 recession, followed by 1982 double-digit inflation rates and another recession, followed by the 1991 recession returns and finally the wonderful 2000-2002 market experience? If the first eight years of retirement follow that iteration (and I am sure that one of the 10,000 follows something similar), then the only approach that makes practical sense is one like Mr. Evensky and his firm that recommend years of safety.

I have listened and read and studied many of these theories on how to manage a retirement portfolio during the distribution phase and am convinced that Mr. Ray Lucia, CFP® in his book, “Buckets of Money”, provides the most clear-cut, practical solution. More importantly, the client would be able to understand his concept, believe in it and therefore be more likely to implement it successfully. Rather than go into the details here I encourage FPA to research his theory and possibly present his approach in a future edition of the FPA Journal.

Following Mr. Ray Lucia’s advice rather than Mr. William Bengen’s, the planner/advisor would never have to, in Mr. Bengen’s own words in his August 2006 article, “…explain to the client that in the event of a major bear market early in retirement, Draconian measures may be required to salvage his withdrawal plan…”. No, instead, the client would be able to sleep easier at night knowing that the stock/bond portions of the portfolio had time horizons that were reasonable and appropriate and that the cash and short-term portions of the portfolio were sufficient to meet the client’s immediate income needs.

It is time to return to the needs of our clients and truly help them with a financial planning process that is not found in a “pie in the sky” but where the “rubber meets the road”.

Sincerely and respectfully
submitted,

Phil Bour, CFP®
South Riding, VA
BourGroup

Withdrawal Rate Ideas for Retirement

The Yale spending formula is as follows. The allowable spending in any fiscal year is equal to:• Seventy percent of the allowable spending in the prior fiscal year, increased by the rate of inflation, as measured by the Consumer Price Index, for the 12 months prior to the start of the fiscal year; and• Thirty percent of the long-term spending rate of 4.5 percent (a total of 1.35 percent) applied to the four-quarter market average of the endowments, for the period ending December 31 prior to the start of the fiscal year (almost all universities and colleges have a June 30 fiscal year).

Monte Carlo potential problems or shortfalls

More on Monte Carlo pitfalls:

No attempt is made to correlate a given year's inflation rate with prior years' inflation rates or with corresponding portfolio returns. Some may view this as a significant weakness of the simulation, however, the historical record indicates that correlations in this area tend to be unsteady at best.

The correlation between the inflation rate and portfolio returns is not modeled.

Another pitfall:

The withdrawal policy, or real spending policy, does not allow the simulation to account for the natural human tendency to make adjustments when things aren't going as well as expected, or when things are going better than planned.

A successful retirement is defined as one in which you don't run out of money and you have enough to pay your expenses. In very general and rough terms, most retirement planners would consider a probability of success that was below 80% to be a failed plan.

Sometimes portfolio survival is prioritized above annual spending needs. In such a case, the programs drive to preserve the portfolio leads to retirees starving!

One thing to keep in mind is that these balances represent the average balance for all simulation runs. This can be a bit misleading because half the time the retiree would have a smaller balance, and half the time they might have a larger balance.

Again, keep in mind that the balance and spending percentage amounts are averages and don’t represent any real or concrete data.

Can you tell that I do not like Monte Carlo simulations for practical retirement withdrawal strategies? Forget those colorful graphics - they are "near" meaningless.