You see a properly diversified portfolio would mean that ALL of your stocks and ALL of your bonds have simultaneously gone to zero value! How likely is that? There is debate in the profession about rising equity “glide paths” or declining equity glide paths. Confusion may arise due to what event is the glide path designed for? Getting TO retirement (target date funds)? Or Getting THROUGH retirement (withdrawal research). The discussion here is about the later … i.e., what might the allocation look like as one ages through retirement.
Rather than constrain the spending of the many, via Safe Withdrawal Rates (SWR) (what happens when conservative assumptions such as 4% and age 95 are used*) so that those few who may outlive age 95 still have money, the Dynamic Updating approach looks at spending a prudent amount during younger retiree ages, with a rolling eye each year on the shifting possibility of outliving revised expected ages as well, also with the goal of retaining portfolio value to sustain spending into ages possibly much older than 95 (when such older ages become a higher likelihood).
I’m in the school of thought that the amount of equity (stocks) in one’s retirement income portfolio may be higher while a “younger” retiree, and then shallows out to less stock as one ages with less time remaining, and thus more exposure to stock market volatility with less time to recover. Thus the visual may look like the above photo where the blue would represent stock index funds and the black represent short term bond index funds. In other words, I’m in the school of thought that the equity glide path declines (less stocks with age).** The volatile (blue) part of the portfolio is from stocks, while the more stable (black) part of the portfolio comes from short term bonds.
I’m going to use the science of oceanography with the ocean and it’s waves as a metaphor. The above photo is what happens to waves as they move in from the deep water to land.
Taking a closer look at the exposure to volatility, much of the stock portfolio represented by the blue section of the diagram at the right, is not subject the up and down of wave action. Below the surface, much of the water is undisturbed (below the “wave height” line at right) – as is much of a properly structured portfolio. The up and down part of a portfolio may be seen by looking at standard deviation probability of how far down a portfolio value may drop. In other words – what is the “wave height” of a portfolio?
In deep water the wave effect is felt deeper than wave height, yet again, not to the complete depth, but near the surface to a measurable depth. Standard deviation provides a measure of the depth, and probability of exposure to that depth, when it comes to portfolio values. In this metaphor, and Dynamic Updating viewpoint, deep water waves represent younger retirees who have more time remaining before they “reach shore.” In other words, a declining equity glide path that is suggested by running simulations based on aging into shorter time periods remaining.***
Thus, most of the portfolio value below the dollar value amount calculated by standard deviation is “untouched” by the market wave actions – the ups and downs of the markets combined effect on a portfolio.
What amount of a portfolio may be unaffected by market “wave” action? Each example for each $100,000 of portfolio value …
Defensive Portfolio 75% short term bonds, 25% stocks: 4.44% real return, 6.29% standard deviation … 68 percent of the time the waves go down 6.29% = $6,290 which means $93,710 of the portfolio is unaffected 68 percent of the time; 95% of the time the waves could go down 12.58% = $12,580 or $87,420 of the portfolio value is unaffected. 99.7% of the time the waves might go down 18.87% = $18,870 or $81,130 of the portfolio is unaffected by the “storm.” The SWR approach is to constrain spending closer to the $81,000 portfolio value in effect because SWR looks at the lowest trough of the waves; while the Dynamic Updating approach looks at spending off of the more often experienced values above the $87,420 to the full $100,000 with a flexibility of reducing (not stopping) spending IF needed to the $87,420 level temporarily until the storm passes and spending resumed above that normal level again. Also, spending does NOT need to constantly adjusted as some envision when thinking about flexible spending. ****
Conservative Portfolio 60% short term bonds, 40% stocks: 5.53% real return, 8.65% standard deviation … 8.65% = $8,650 or $91,350 unaffected 68% of the time; 17.3% = $17,300 or $82700 unaffected 95% of the time; and 25.95% = $25,950 or $74,050 unaffected 99.7% of the time.
As one can see, with more volatility exposure the more the portfolio value is affected. So reaching for returns with greater stock exposure has a direct effect on how much of the portfolio may be subject to market storms. Most of the time the effect is “shallow,” however there are times when storms stir deeper. The market storm of 2008/2009 affected portfolios differently – the difference being precisely the level of volatility the portfolio was exposed too! The standard deviation numbers above INCLUDE 2008/2009.
Constraining spending based on rare events is part of the SWR approach. The Dynamic Updating approach looks at what is normally prudent from both Probability of the Portfolio (standard deviation and its’ effect on the percentage of Monte Carlo simulations¤ that fail to reach the end of the time period measured) as well as Probability of the Person (period life tables). The Dynamic Updating approach also replaces the “rule of thumb” approach prevalent in the SWR method (e.g., 4%, age 95, 60% equity portfolio, etc.) with specifics from the actual portfolio and expected longevity characteristics of real people.
In a prudent plan, one may base spending on a calculated trough (e.g., two standard deviations) to measure how much of the portfolio is available to support floor or necessary expenses. The amount of spending above that portfolio value could be called flexible spending on things that are nice to do such as travel, nights out, etc., for example.
So … in practice … one first sets time period over which the money should last based on current age and life expectancy# (the difference equals how many years for this year’s distribution period; Dynamic Updating has one do this each year because this is how one measures and monitors not outliving their money). With the number of years determined, and using Monte Carlo software (with actual portfolio characteristics), one then determines the raw withdrawal rate outcome with 10% of the simulations failing (possibility of adjustment). The raw withdrawal rate (WR%) may be adjusted to mute exponential effects (seen most easily in IRS Required Minimum Distribution tables) => adjusted WR% = WR% * (1-1/n) where n is the number of years used in this year’s simulation. Now the prudent income range (annual) may be determined by multiplying the adjusted withdrawal rate by the upper and lower portfolio values one determined as shown in the examples above.
Moral of the story: NOT ALL of ones portfolio is subject to market volatility … and therefore NOT ALL of ones’ spending is either. Value in dollar terms remains stable below the affected “water line” one may calculate via standard deviations. Thus, investors and retirees in particular, should have a better understanding of the specific portfolio characteristics and pay more attention to standard deviation than one tends to do otherwise (tendency is just a focus on returns).
*Vanguard longevity table: a couple, both age 65, have a 18% chance either may outlive age 95; female a 13% chance; or male an 6% chance to outlive age 95. As one ages, the probability goes up to outlive older ages, and thus the portfolio should be constructed, measured and monitored accordingly (in the metaphor above, one is getting closer to the beach). If one calculates a 24 year period, which would result in a higher spending rate for this couple, either has a 51% chance to outlive age 89, female a 36% chance (64% don’t outlive 89), and male a 24% chance. The question becomes when do you want the money available for spending? Spend a little more now at age 65? Or retain more of it for spending past age 89 or 95? Is there a bequest motive, and how much? (Calculations are based on mortality data from the Society of Actuaries Retirement Participant 2000 Table on their site at the time this was written).
**The rising equity glide path is simply the other side of the same withdrawal research coin – it looks at the same beach. The results from that research come from restricting spending so that the “wave” portion of the portfolio, the part of the portfolio balance that goes up and down with the markets, is NOT really part of the potential withdrawal rate. If it were included, more money could be withdrawn (the withdrawal rate could be higher than 4% in most situations). Thus, there is an “unused” portion of the balance that grows as a function of portfolio growth. Over time, this unused portion of the portfolio becomes large enough where more equity exposure could be taken. In other words, this Safe Withdrawal Rate (SWR) research starts at the beach, using the ocean metaphor above, and moves into the ocean – the reverse process as described above where the risk of volatility is viewed at the beginning of retirement from the open ocean and moving towards the beach. The SWR view says sequence risk is greatest at the start of retirement and goes away as you age because of the “unused” portfolio value for spending. The Dynamic Updating view as shown in our research is that sequence risk is always present and never goes away at any age. Regardless of what birth cohort you may be in (what year you were born), the present market and economy affects all similarly – the difference being your exposure to the markets or economic effects, whether working or invested. And no, one can not extract oneself from the system unless you can go to Mars to weather it out! Risk transfer simply transfers the risk to another party (Kitces) – but the other party is also affected by the same events relative to their own exposure to the issues at the time.
***I should point out that the rising equity glide path research also positions less equity at older ages if the data were arranged from the perspective of time remaining instead of total time of withdrawals (counting periods down from highest to lowest time: 30 years, 25 years, 20 years, etc. time remaining).
****The more a retiree requires inflexible spending the more constrained they will need to be with inflexible income sources – sources that are not subject to fluctuating values. Those sources typically expose retirees to inflations’ loss of purchasing power risk (there’s always a risk trade off with anything¤¤), unless indexed for inflation increases such as Social Security and some pensions, and thus don’t have potential returns to provide more spending ability later for higher costs due to inflation’s effect. Spending from a “Distribution Reservoir” which contains 3 to 5 years of spending money (cash and short term US and International bonds) helps dampen out any need to adjust spending quickly due to market movements because those movements change direction and depth often; thus this builds in a smoother period of time to evaluate and adjust as needed.
¤Monte Carlo simulations are NOT a retirement model, they are a single point solution derived from simulating out years to arrive at the solution for that point in time only. Using the 24 year period from the above longevity example for a 65 year old couple, the 23rd year simulation is just 1 year long. However, an 88 year old couple (23 years after age 65) have a 47% likelihood of either of them outliving another 7 years; and again at age 95 (88+7) a 54% likelihood of either outliving another 4 years (now to age 99; and this process continues on until demise – and by design and management, there is still money left). THE RETIREMENT MODEL should not arbitrarily end as simulations need to; the model should incorporate expectations for those later ages because, as our research showed (and common sense), spending early affects spending ability later. And, as long as someone is alive, they never reach their expected longevity age (it is always an older age than present age). The distinction between point solution and lifetime model has not been incorporated very well into the discussion to date.
#Life expectancy percentiles may also be used to “pull” money for spending into earlier years, or “push” money for spending into later years, depending on each person’s overall goal of how they want to spend their money. Many times people want market returns to do the heavy lifting in retirement, when in fact it usually comes down to spending too much, or not, over rolling time periods. Dynamic updating of portfolio characteristics and time period expectation looks at how to make it more likely one doesn’t outlive their money managed within a diversified portfolio.
ªProperly diversified portfolio explained conceptually: Reducing the Risk of Black Swans: Using the Science of Investing to Capture Returns with Less Volatility by Larry Swedroe. Also, William Bernstein has a series of short books including the difference between “Deep Risk” and “Shallow Risk.” So one may conjure up fearful scenarios of tsunamis or hurricane surges. Both are local, rather than global, phenomenon. A broadly diversified global portfolio may have some “local” effect, however the rest of the portfolio should be constructed with offsetting mathematics (efficient frontier, covariance matrix, etc.); in other words not all eggs in the same basket. Risk transfer is not the same as being able to avoid risk since the very phenomenon stressing markets affects all who are in those markets proportionally to their own exposure to the stressed market at the time. We all live on the same blue marble. There are no guarantees against “tail risk” anywhere (more than 3 standard deviations from current price). In other words quoting Kitces “…“black swan” events that present tail risk at “very, very, low probabilities of failure” are tail events that can simultaneously strike portfolio-based AND insurance-based solutions.”
¤¤Putting all your money into insurance or bank companies involve company or industry specific risk, which is unsystematic (diversifiable) risk that can be diversified away through proper allocation, so all that remains is systematic risk which is what the concept of Black Swans, discussed above, is about.
Note: A great discussion about different research currently out there is summarized a paper Wade Pfau links to in his blog post here Making Sense Out of Variable Spending Strategies for Retirees March 16th 2015.
Note: A great discussion about different strategies currently out there is summarized in a post by Dirk Cotton March 17 2015 Dominated Strategies, Illogical Strategies, Problematic Strategies and Strategies That Just Make Me Queasy.
Note: Are market returns “normal” or distributed differently. In practical application, the differences in spending results suggested may be slight.