Actual Probability of Sequence of Returns Risk
Has anyone calculated the actual probability of sequence of returns risk mattering? I’m a probability guy and if there’s an 80% chance it won’t matter, then I’m not going to actually worry about it.
Join the conversation
Sign up to reply, follow discussions, and connect with the ChooseFI community.
Comments
That all depends on what you mean by "mattering". I seem to recall a Choose FI episode some years ago where Brad, Jonathan, and the guest discussed the idea of "fixed" spending versus "discretionary" spending in retirement. If you have fixed spending that requires a 4%-5% withdrawal rate, then a bad sequence could result in failure (i.e., completely running out of money). However, if your fixed expenses only represent a 2% withdrawal rate and your discretionary expenses (say, fancier than necessary car/house/food, travel, etc.) represent a 2%-3% withdrawal rate on their own, and if you are willing to cut back or eliminate the discretionary spending in response to a "bad sequence" then odds are it won't "matter" in the sense of running out of money as a likely outcome. So, the "probability" is influenced by at least that consideration.
Oh yeah, Madfientist has this calculator on their website which is called the Discretionary Withdrawal Strategy . It's here, you will probably need to make an account to get in: https://lab.madfientist.com/calculators/discretionary_withdrawal
The blogpost about it is here: www.madfientist.com
They use the fact that some of your spending in retirement may be discretionary (and thus can be temporarily reduced when the markets do bad) to allow you to increase your safe withdrawal rate. Alternatively, you could of course see it as a way to have increased chances of success at the same withdrawal rate. The more you can reduce your expenses when the market is down, the lower the chance that you'll deplete your portfolio.
Fabiooltje,
Thanks for sharing the madfientist strategy. I've been researching/experimenting with flexible withdrawal strategies and this seems like perhaps a simplified version of the risk-based guardrail strategy I've been reading about, which seems to dwell more in the CFP world than in the FI community. It's a bit of a rif or refinement on the Guyton-Klinger guardrail approach. There's a detailed article here: www.kitces.com but in general it involves establishing a withdrawal rate that approximates say, an 80% probability of success (using monte carlo or historical analysis) and then monitoring your portfolio. If it grows to a a point where you reach the upper guardrail (say, 100% PoS), you give yourself a raise to bring it back down to 80%. If the portfolio is depleted or loses value to the point where you reach the lower guardrail (say, 60% PoS), you tighten your belt a bit to get back to 80%. This approach seems to strike a good balance between the risk of underspending in the early years of retirement and ending up with a huge legacy, while still incorporating flexibility measures to prevent overspending when the market goes south.
I've done some DIY calculations using projection lab and other planners that incorporate monte carlo and historical analysis but the best tool I've found is incomelaboratory.com Unfortunately, it's aimed at the CFP world and is priced accordingly. I've been messing around with it in the 30-day trial space but will have to give it up when that runs out.
Does anyone have experience using the risk-based guardrail approach and/or know of any resources available to the FI community?
While I don't use the guardrail strategy, Kitces did have many articles on it. I believe BigERN also put is usual, pessimistic take on it out there somewhere as well (FYI, I tend to fall very close to BigERN in the analysis I buy into).
I don't think I saw anyone link this specific post from BigERN, so I'll drop it here:
The histograms and heatmaps on this post have been some of the most useful-to-me factors in deciding on my own strategy. They focus not on guardrails, but on "equity glide paths", but also include non-glidepath probabilities as well for some, using historical ensembles (rather than monte carlo simulation). You could take a look at glide paths and see if the probabilities are more amenable to you as well, especially some that ramp up equity allocation very quickly after retiring.
If you're truly OK cutting up to 50% of expenses within a month of a market downturn, I'd say you really can risk a much higher withdrawal rate if the future bad-case scenarios are at all like anything in the past. I'd be more skeptical of your ability to do that, but that's something I can do as well and I know it =).
Thank you! This is exactly what I meant. I plan to use a 5% wd rate most likely, but a good chunk of that would be spending I could easily cut. I’d also be fine with looking to add extra income if need be. I’m a former poker pro who can always make some side money if need be.
So, in your case it sounds like sequence of returns "matters" to the extend that it acts as a driver for adjusting your discretionary spending, and not as a potential cause of "failure" (i.e., running out of money because of accelerated depletion in the face of expenses that you cannot reduce or offset).
Precisely. Basically, I’m not going to plan for it by holding bonds or a lot of cash. I’m going to stick with 100% equities and use my flexible wd rate and other possible income as the ways to mitigate issues, rather than limiting growth.
I would caution 100% equities allocation during retirement. I do not believe any experts recommend that either. 100% equities is only for the aggressive growth phase. As you near retirement, you should be reallocating to hedge against the sequence of returns risk. This will stablize the discretionary income portion of withdrawal.
I feel like the safe withdrawal rate advice from the original Trinity study already takes into account the sequence of returns risk. Because the cohorts that do bad in the original SWR study, did bad precisely because they met negative returns at the beginning of their retirement period. So for a shortcut I would say: if you create a withdrawal rate such that your chances of success according to the SWR study (and updates) is 90% or over, then your chance of meeting the wrong "sequence of returns" is therefore 10% or lower.
I feel like worrying about sequence of return risk when you have for example chosen a 3% withdrawal rate (meaning a chance of success of something like 97% and that is without you adjusting your spending) ... is nitpicking.
And oh, still go and nitpick to your heart's content if it makes you feel better / more secure, but OP doesn't seem to need to do so. (Or at least not at this stage of their life).
Depends on what you mean by "mattering". Its not really a probability to begin with and not a probability of running out of money, but of having to alter your plan. You are just as likely, however, to end up with multiples of what you started with, so many calculations are too conservative unless your primary goal is to max wealth at death.
In practice it is highly dependent on what kind of portfolio you are holding, your base rate of withdrawal, and whether you are willing to be flexible in those withdrawals. There is no reason you can't be using a 5% withdrawal rate if you are properly allocated and your expenses have a discretionary component of at least 25%. (40% is the norm.)
OTOH, you can virtually always solve this problem by simply "not spending money" -- i.e., 3% or less. In that case, it does not even matter what you are holding within reason (could be 20% - 100% in basic index funds), as the only issue will be how much you die with.
Many people waste a lot of time refining "plans" and fiddling with calculators when they have already effectively solved their problem by committing to not spending much money and continuing to accumulate until death. Most popular financial gurus actually follow this strategy in their personal lives, although they get sheepish and reticent when asked about what they are really doing.
It’s nice to hear someone else expound upon this like I have in my mind. My plan is to hold 100% VTI/BTC until I die. Seq of returns or market downfalls will simply adjust my spending and/or prompt me to generate more income. I’m going to be defaulting to 5% and flexing that up/down as needed and based on how things are going. And if I run out of money at 88 because I was living it up in my 50s/60s - GOOD.
Karsten from ERN did a short blurb about the likelihood of SoRR for early retirees, and the basic idea was that it is much more likely than you'd expect, so be sure to include CAPE in your calculations and/or assumptions.
The likelihood of hitting your number is most likely when the market has run up for a while, and if you retire when you hit your number, the likelihood the market is overpriced and due for a correction is far higher, put another way: the likelihood of being impacted by SoRR is far higher. If you want to greatly reduce SoRR, retire when the market is underpriced (i.e., low CAPE), and now is almost definitely not that. Calculators that include CAPE (like ERN's) factor in the higher SoRR.
His analogy was driving times. If you budget for the average (when 80% of the time the roads aren't congested) but actually drive when most people are on the road (during rush hours), you will probably be very late.
TL/DR; the likelihood of being impacted by SoRR is closer to 80% than 20%, unless you get a windfall independent of the overall market.
I hate this logic. It all sounds like crystal ball market timing nonsense. Nobody knows what the markets will do tomorrow. In the long run most people expect them to keep going up. CAPE ratios have been bad for a while and the market has just gone up and up and up. Just because the market is at an all time high doesn't mean it's any more or less likely to go down or up. The only numbers you should be using to plan are the base rates.
Agreed. 80% chance that the market goes down is laughably incorrect.
I'm currently of a similar mindset that I won't worry about it too much. I'd rather retire early with the understanding that if I hit a bad sequence of returns I can always work part time to earn some more money if needed. I prefer to plan to retire early with a backup plan than work for another 20 years just to be extra safe
Can anyone speak to the notion of your FI threshold is by date and not by number?
17 years to FI, should be 17 years, if it's "averages" were tracking.
Your asset allocation has a pretty big effect on the risk of things going sideways from a bad start. Have a look at this chart and you can play with the asset allocations to see how much SDS or Start Date Sensitivity your decumulation portfolio might have. You may not be able to predict what early sequence or returns will be. But you can reasonably predict how your asset allocation will behave if things do go poorly.
Spoiler alert.... 100% stocks has a really high SDS and the risk parity style portfolios have a much lower SDS
Portfolio Charts | Start Date Sensitivity – Portfolio Charts
Thank you! I’m fine with high SDS if it also means highest % of growth because I can simply decrease spending or increase income.
I think there's probably a sweet spot where growth is still fairly high and risk is lowered. I believe this is the goal of risk parity style portfolios and the efficient frontier.
Also, I agree being flexible is great and using guardrails or adjusting withdrawals based on market performance is a great concept. Portfolio charts has this other tool for simulating rules for adjusting withdrawals based on market returns over time. It's great to play with this and develop your own set of rules portfoliocharts.com
Lastly, I would say that human nature makes it very easy to increase your spending and expand your lifestyle. Reducing spending on the other hand can be very painful. It's probably a bit foolish to plan on cutting your spending dramatically if/when things aren't going well in the markets.
I’m new at this decumulating phase, but does it make sense to address risk by recalculating your 4%? For example, my husband and I are planning to recalculate our 4% every three years. We might get a raise, or we might have to cut back on some expenses until the next recalculation.
I plan on recalculating my 5% annually.
The original 4% rule includes increasing your spending every year at the rate of inflation. Only the 1st year would be at 4%. Very few people are probably following this strict interpretation of the 4% rule as written by Bill Bengen
But yes I think having some rules to adjusting spending based on the market returns is probably a good idea. This tool on portfolio charts is great for simulating how spending adjustments might work out in a variety of scenarios. portfoliocharts.com
One small argument to not worry too much about it, is that first years after reaching FI, when we are prone to the sequence of return risk, are also the years when it is relatively easiest to go back to work, just call it "sabbatical" and continue your career if investing goes very wrong.
You could also consider an amortization based withdrawal strategy, which accounts for both withdrawal timeline and your current portfolio balance.
https://www.bogleheads.org/wiki/Amortization_based_withdrawal
Really? 1 in 5 seems likely enough to have some non-zero level of concerns.