Welcome back to another guest post. Dr. David Graham, over at FIPhysician has been on a roll. His spike in productivity has been the perfect “hedge” against my drop in productivity while traveling this summer, so when he offered me to write a follow-up on his very well-received guest post a few weeks ago, I was all for it. This current post is about adding a “glidepath” to your retirement portfolio and how and why this would change the success prospects over a 60-year retirement horizon. Over to you, Dr. Graham…
In my last post, I show a 4% Safe Withdrawal Rate (SWR) is actually NOT safe over 60-years (assumptions, assumptions). A more conservative 3.25% SWR does ok. On the other hand, if the asset allocation is increased from 60/40 to 90/10 stock to bond ratio, a 4% SWR thrives again. ERN advises, however, that a 90/10 portfolio sets you up for even more Sequence of Return Risk (SORR). SORR describes the long-term detrimental effects initial negative market returns have on overall portfolio success. Even if the stock market eventually recovers, selling part of your equity portfolio at rock-bottom prices can lead to premature failure of the withdrawal strategy.
What protects from SORR yet permits a higher SWR? A rising equity glidepath is one possibility. Let’s look at the details…
A glidepath primer (by Big ERN)
Obviously, I have written extensively on glidepaths. Two long posts in my SWR Series, Part 19 and Part 20. In a nutshell, in retirement you want to raise your equity share. Now that’s intriguing because the conventional wisdom has been that over your working years, you would reduce your equity share as you approach retirement. Shouldn’t you just continue the equity path down? No! The reason is Sequence Risk!
A rising equity glidepath means that you reduce (not eliminate, only reduce!) the necessity of selling equities while they are trading at (hopefully) temporarily depressed prices, say, during the Global Financial Crisis. But the glidepath keeps giving even after the worst of the crisis is over. In a way, the glidepath is actually even more useful after(!) the stock market reaches the trough, e.g., March 2009. A rising equity glidepath lets you keep riding the nice positive equity momentum (aka bull market) coming out of the crisis. With fixed stock/bond weights, you’d constantly rebalance your portfolio, i.e., sell equities that are rallying and buying bonds that are performing far worse. A rising equity glidepath reduces (not necessarily eliminates) this rebalancing that’s clearly a losing proposition during a bull market. You’d mostly sell bonds and let the equities ride! In summary, the reason why I think that this is so useful: The peak to trough in 2007-2009 was only 17 months (I know seemed like forever back then!), but the bull market so far has been going on for 10+ years and the glidepath keeps giving!
But now, back to Dr. Graham…
A Rising Equity Glidepath
Let’s look at combating SORR with a rising equity glidepath. Most glidepaths—those in target-date funds—decrease the stock percentage to or even throughout retirement. That seems inefficient because SORR is most problematic 5 years before and 10 years after retirement, so there is no reason to keep the stock percentage low after most (all?) sequence risk has passed.
A rising equity glidepath lowers stock percentage before retirement and ratchets it up thereafter. Can a rising equity glidepath increase your SWR? Let’s look at an early retiree who plans for a 60-year retirement and wants to protect that nest egg against running out of money. We will use the same basic assumptions from the post a few weeks ago…
A Reminder from the Last Post
As a reminder, see above a hypothetical scenario from my previous post with a 60/40 portfolio. In green, note a 4% SWR over 60 years. A Monte Carlo simulation (not displayed here) predicts a success rate of 18% and this portfolio expires after 44 years. In blue, a 3.25% SWR grows over time to reach $1.6M. Monte Carlo predicts a success rate of 44%.
Using a Rising Equity Glidepath
See the effect of a rising equity glidepath rather than a fixed 60/40 portfolio above. In green, a 4% SWR almost make it to 60 years. Monte Carlo odds are now higher at 29%. In blue, note the 3.25% SWR actually has a rising slope compared to the fading slope in figure 1. The portfolio size grows over time using a rising equity glidepath vs. a static 60/40 allocation. Monte Carlo odds are now 50%.
ERN: Isn’t this really fascinating? A small change to your passive asset allocation makes a huge difference. There’s no stock-picking, no market timing, just a passive asset allocation path, and the 4% Rule can last another decade and a half. It’s consistent with my findings from the SWR Series where I exclusively looked at the failsafe withdrawal rates, but it’s also nice to display the change in the retirement length for a given withdrawal rate.
Illustration of the 4% SWR with a Rising Equity Glidepath
Figure 3 demonstrates the percentage of portfolio withdrawn over time assuming a 4% SWR indexed for inflation. With a rising equity glidepath, the percentage of portfolio utilized every year stays under 5% for the first 15 years. Then, it slowly increases to critical levels after 50 years. Note that the 4% SWR fails in figure 2 at year 59.
Assumptions, Assumptions. Again!
For this scenario, ERN’s data is utilized to model a rising equity glidepath.
The portfolio transitions to 60/40 five years before retirement. After retirement, the portfolio rapidly increases to 90/10 over the next 10 years.
ERN concludes a rapid increase in stock allocation—over 10 years—performs better than increasing over 20 or 30 years. In addition, an aggressive portfolio performs better than less aggressive ones.
ERN: Again, the reason is that you want a glidepath that covers not just the bear market (often only 2 years or so, actually only 17 months in 2007-2009) but the bear market plus the subsequent bull market. You probably don’t need a glidepath over multiple business and market cycles, so 10 years sounds like the perfect sweet spot.
Instead of landing at 60/40 or even 100/0 after SORR, a 90/10 asset allocation is used from years 10 to 60 after retirement.
What about SORR?
Let’s look closely at sequence risk. We need to consider SORR given that 90/10 portfolios fail frequently due to negative market returns. What if we simulate SORR using actual returns from 2000-2010? After that first decade, assume that stocks return 9% nominal and bonds 4% (also nominal) for the purposes of this scenario (for a blended 8.5% return of a 90/10 portfolio).
SORR and Aggressive Portfolios
Above, we see the effect SORR has on 90/10 portfolios. In green, a 4% SWR rapidly extinguishes the portfolio after 25 years. Even a 3.25% SWR is no match for SORR in this setting, as seen in blue. The initial bad decade—and ongoing withdrawals from the portfolio—decimates almost half of the value. Over time, and despite decent assumed returns, 2% inflation catches up and extinguishes the portfolio after 55 years.
ERN: Even with the 8.5% blended portfolio return (0.9×9%+0.1×4%), you can’t sustain the 4% withdrawals plus inflation adjustments. That’s because the portfolio was so decimated during the two bear market. Recall, the 4% WR refers to the 4% relative to the initial capital. The effective withdrawal rate was above 8% at the bottom of the second recession! This is again the definition of SoRR! Some folks in the FIRE world always get this wrong; they think that it’s enough for the market to recover, but we have a much steeper hill to climb! The market has to recover the bear market drop, plus the withdrawals plus the inflation adjustments! That can take a long time! Or it may never happen and the portfolio runs out of money!
SORR and Rising Equity Glidepaths
Using the rising equity glidepath described above, a 4% SWR is still not safe. Seen in green, despite a rising equity glidepath, a 4% SWR survives 30 years but not 60 years.
Conversely, in blue, note the performance of a 3.5% SWR. A 3.25% SWR is too successful; instead, look at 3.5% SWR. The 3.5% SWR with a rising equity glidepath compounds after the initial bad decade and passes with flying colors.
A couple points to consider:
First, a rising equity glidepath improves performance assuming consistently low future returns. From my prior post, increasing a 60/40 portfolio to 90/10 improves the odds of success of a 4% SWR, but risks poor sequence returns.
Demonstrated here, both 4% and 3.25% SWRs are decimated by SORR. A rising equity glidepath, however, can help. It rescues a 3.5% SWR (more so 3.25% SWR), but not a 4% SWR.
FIRE vs. Traditional Retirement
Take care when discussing SORR and a rising equity glidepath. Traditional retirees—with 30 years left—are different than 60-year FIRE portfolios.
On the flip side, there is nothing conventional about FIRE. Proponents on the path to FIRE certainly are more financially literate than the average retiree. Critics of rising equity glidepaths point out it is difficult to get an 80-year old to invest 90% in stocks. Conversely, it is easy to talk a 40-year-old into aggressive investing, but try to talk that same 40-year-old down to 60% equities when they retire early!
Kitces mentions this subject. He starts with rising equity glidepaths but more recently recommends bond tents as volatility dampeners to protects against SORR. It may be that he changed his tune—going from rising equity to bond tents—so that he doesn’t have to sell old folks on stocks, rather sell younger folks on bonds.
Traditional retirement is a different animal. This review of rising equity glideslopes is not for them.
Practical Advice for a Soon-to-be Early Retiree
So, let’s say you will retire in 5 years at the age of 40. Now is the time to adjust your portfolio down to approximately 60/40.
This is especially true now, given high current stock market valuation.
In fact, ERN discusses that a rising equity glidepath is not useful during periods of low market valuations (e.g., low CAPE ratio). When valuations are high, however, (e.g., CAPE >20) there is a small benefit to rising equity glidepath—depending on the slope and beginning and ending allocation.
Then, once retired, ratchet stock allocation back up to 80% or higher once the risk of SORR has passed. ERN’s prescription—do this rapidly over 10 years—and, the higher stock allocation the better.
What is a safe withdrawal rate Over 60 Years?
So, what is a SWR for a 60-year retirement? ERN suggests less than 4% even with a rising equity glidepath. In truth, 3.25% is safer, but you might get to 3.5% if you protect your downside—SORR—and are aggressive again after the peril has passed.
Back to ERN for some final thoughts
Thanks again for taking the time to write this very important post. A glidepath is probably my personal favorite tool to cushion Sequence Risk. But wait? Aren’t we all supposed to be passive investors? Isn’t investing supposed to be simple? Well, saving for retirement is indeed simple (see JL Collins’ book), but as I outlined in my SWR Series Part 27, withdrawing money is a slightly more complicated mathematical and financial problem. There is some rationale for (ever so slightly) tinkering with your portfolio while withdrawing money in retirement. Apart from tailoring your SWR to equity valuations (see SWR Part 18) a glidepath is also recommended. It’s only slightly more “active” but it tends to raise your failsafe withdrawal rate by enough that people should really take notice. And by the way, is the glidepath really that active anyway? Remember, using fixed equity/bond weights would also potentially entail rebalancing your portfolio during the bear and bull market when equity and bond weights shift away from their target weights due to diverging returns! So, what’s really active and what’s passive here?
If a glidepath is so successful, then what’s the downside? Glad you asked! For full disclosure here are two (slight) caveats I can think of:
- The glidepath works like an insurance policy. If you indeed have a bear market right around your retirement date and then a subsequent bull market then you will certainly do better than the static allocation with a high equity weight. The downside is that when you don’t have a bear market but rather a continued bull market then – you guessed it – you will do worse than under the high static equity weight. But that’s when you don’t worry about running out of money. You can afford to do a little bit worse because that’s the scenario when the portfolio will likely grow beyond your wildest imaginations. Just like car insurance is “wasted” if you don’t have an accident. But even if car insurance has a negative expected value you still find it a good “investment.” It pays off when you need it the most and costs a little bit when it doesn’t hurt so much.
- The efficacy of the glidepath depends on how bonds perform during the bear market. If bonds actually rally during the bear market then the glidepath works best. We got accustomed to this negative correlation during the last few recessions, but keep in mind that this is not guaranteed. Case in point, the 1970s and early 80s. Stocks and bonds both dropped during the inflationary supply shock recessions. But even then a glidepath worked, just not as well as in a demand shock recession, provided that the bond portfolio drops less than the equity portfolio (which was the case even in that bad inflationary period).
Even with those two caveats, I’d still consider the glidepath a nice insurance policy against Sequence Risk!
Thanks for stopping by today! Please leave your comments and suggestions below!
Picture credit: Pixabay.com
77 thoughts on “Can a Rising Equity Glidepath Save the 4% Safe Withdrawal Rate Over a 60 Year Retirement? (Guest Post by Dr. David Graham)”
Excellent post! I have 2 comments:
1. What is the difference between a bond tent and a rising equity glidepath? My understanding is a rising equity glidepath is just the back half of a bond tent.
2. It would be interesting to see simulations where the first 10 years are 2000-2010, then the rest is either historical or Monte Carlo returns. It is difficult to draw too much confidence when using constant returns (though I guess at that point it is just a standard safe withdrawal rate analysis with a 90/10 portfolio and however much you are left with in 2010).
1: you’re completely right. The glidepath is just the right leg of the bond tent.
2: That would be an interesting experiment. Though, you might get some really nasty low SWR because you might get some MC draws with a renewed equity drop right at the beginning of year 11.
Wondering why the rising equity glidepath isn’t recommended for a traditional 30-35 year retirement? Did I read that correctly? Is there an alternative recommendation for SORR?
Also, are you saying that an early retiree should begin the 10 year rapid move to a 90/10 allocation once they retire, even if in the throes of a bear market at that time?
Its not that it isnt recommended for a traditional retirement, it is that it is not NEEDED for a traditional retirement. See my post on rising equity glidepath in traditional retirement here: https://www.fiphysician.com/what-is-the-best-asset-allocation-for-retirement/
The effectiness of the GP was better for the long horizon, so this is consistent with my research in Parts 19/20 of the SWR Series.
I certainly recommend the rising GP for the traditional retirees. Wade Pafu and Michael Kitces do so as well.
About the timing: Exactly because you’re in the bear market you should start the GP. The bigger question is, should you already shift the equity weight before the crash or should you wait for the drop. See the discussion about the active vs. passive GPs in parts 19/20 of the SWR Series.
I wonder what your thoughts are on the current economic cycle with 0% or negative bond yields and significant debt burdens (corporate and government). Bond allocation in this environment will (at best) only preserve wealth. There is no chance of any income generated (and, I would argue an unappreciated risk of default).
Would like to know what your thoughts are on what exact bonds to buy in this scenario with them seemingly just as risky as stocks. What alternative investments might be allocated in place of bonds?
I don’t hold any bonds, but I am in the accumulation phase. For a bond alternative – it will depend on what purpose you need that allocation to serve. For yield, I would get preferred shared, but they would be terrible as balance against declining stocks. If you can get 6% on FDIC insured accounts (you can get that with a little work on around 200k), that beats bonds. Real estate, gold, or other assets work too. Again, it depends on the purpose.
Would you be willing to share the basics of achieving 6% on $200k in FDIC insured accounts?
Visit doctorofcredit – you will find many bank signup bonuses and deposit bonuses. There was a 1% match on up to 100k in transfers from Ally if deposit held for 90 days, just ended similar deal from Marcus. That’s a 6% yield with little work.
There are 4%+ rewards checking accounts that do require a little work, but they are always available (as opposed to promotions that come and go).
A couple can get 6% on up 200k relatively easily. For anything above that, preferred shares pay 6%, but with much higher risk.
6%? Through signup bonuses?
Doesn’t seem like it would be 6%. Wouldn’t it be 3-3.5%%? 2.2% or so for the interest rate plus the 1% bonus? This would be the same whether one or two people did it because % interest isn’t additive when pooling principle balances. So, having a spouse also do 100k with the bonus does not get you to the 6% you claim. 3% guaranteed return isn’t bad as a way to weather the storm of a recession but if it were 6%, your SWR would also be 6% since the return is guaranteed.
In my SWR Series I use the 10-year U.S. Treausry Bond. An ETF would be the iShares IEF.
You can use alternatives with higher yields, but caution: they also have higher risk and higher correlation with stocks (hence, less diversification). But you might also look into AGG (government + investment grade corporate) and LQD (only investment grade corporate).
I’m troubled by that too. The only hope is that in the event of another bad recession, yields will go down even more (see 10y yields in Germany, Switzerland, etc.) and through the duration effect bonds will gain. Then rebalance, sell bonds and move into stocks. So, even with low yields, there is some rationale for bonds.
I personally don’t fear default in US government bonds.
Did you read the recent op-ed by Ray Dalio. I think he makes a convincing argument against bonds ( I am not as convinced by his argument to hold gold).
He specifically uses this duration effect in an argument not to hold your breath for a repeat of that. Going from 6+ to 0% was nice… going from 2% to 0%? not enough of a spread.
Nice post. I have never implemented this myself, but 5-10% gold had a good SoRR performance in all market crashes.
If you believe the coming paradigm shift and crash go hand-in-hand with a spike in interest rates then short-term (money market etc.) works well.
Excuse my ignorance but I am a bit confused about when you know you are safe from SORR. For example, let’s say you have just retired, you choose an SWR of 4% and follow one of the strategies to protect against SORR. One of two things is going to happen:
1) You get a SORR event.
In this case you have protected yourself against SORR and hopefully all is good. I think I am good with this scenario.
2) You do not get an SORR event.
This is the bit I am struggling with. Presumably your fund grows and if you calculated a new SWR after a number of years it would be lower than the 4% SWR you started with. But when can you say “great I no longer need worry about SORR” Is it when the new SWR is (say) 3.25% and you have a high equity weighting? And what if the SWR falls below 3.25% Do you start travelling 1st class?!
Strictly speaking you’re never completely safe from SoRR. Even with a glidepath.
But ideally, in case 2, the portfolio grew enough that your effective WR is low enough. Then sure, you should ratchet up your withdrawals. Or keep the extra money as a reserve to leave a bigger estate to your heirs.
Dr. David Graham has definitely been on a roll! Thanks for the insightful post.
I fit the description of someone who may retire in within 5 years… I really have to consider getting closer to a 60/40 allocation vs my 95/5 if you look at stocks/bonds. Though, I do have a bit too much cash now, so it’s more like a 60/30/5 (stocks/cash/bonds).
What would you do in my position given that cash component? Throw it into bonds?
Appreciate any insight!
Hey Finance Clever I’m right there with you so looking forward to seeing the response to this question. Typically I’ve been a 90/10 investor but we are planning to reach FIRE in about a year so I’ve been holding on to quite a bit of cash too as a cushion and we’re around 63/25/12 stocks/cash/bonds. We’re also super conservative though and sitting closer to a 2.5% SWR but I’m still nervous of SORR with the market the way it is. My plan is to slowly deplete some of the cash once we FIRE to not have to withdrawal from our investments for the first few years and/or to throw it at the market depending if stocks go on a super sale (and maybe work part time for the first year or two to be even more conservative!)
At 2.5% SWR you should be pretty safe. That would have survived any 60y horizon even at the worst possible historical starting points.
Cash doesn’t seem so bad, given the inverted yield curve right now. Bonds are still a better insurance becuase they rally due to the duration effect in case the stock market crashes.
So, if you keep at least some bonds for diversification you should be fine. 60/40/0 or 60/35/5 wouldn’t make a huge difference.
Hi Big ERN. Thanks as always for the great content. I hate to be the crank on this, but could you explain how this duration effect works to the investors benefit when 10 year treasuries are at 2%? Seems to me that the upside gain is very limited as a practical matter by the low interest rate that you’re starting from. Realistically, interest rates in the US are probably not going to go below zero (lowest yield on 10 year treasuries looks like 1.366% in 2016). So let’s run a thought experiment using that as a historical floor. Potential gain from duration is pretty much capped at say 0.63% (2.0 less 1.366). But on the other side of the wager, there is a whole lot of potential downside loss. If that Yield goes up to say 4%, it’s going to hurt. And 4% is still a relatively low rate from a historical perspective.
We assume that a recession will result in stock losses being offset somewhat by bond gains. I agree that’s possible, perhaps even likely to happen. I also remember the late 1970s when that wasn’t the case.
Yes, the duration effect is limited. I’d think that in the case of a really bad recession, the Fed will try to push the 10y yield to below 1%. Maybe even to 0%. With a duration of 8, you could get a double-digit bond return just from that. Not that great but it beats the -50% in stocks.
Not sure if that bit of upside is worth the downside risk of a repeat of the 1970s. So, I can see the appeal of the cash instead of bonds. 🙂
I think this explains bond convexity very well….
Nice link! Thanks!
I am on track for 4-ish years from now. However, rather than selling appreciated equities and paying the taxes at this point, I decided instead to accumulate cash (instead of continuing to invest future savings into equities). While the cash return is somewhat negligible, bonds aren’t a whole lot better these days. And, consistent with some comments above, I don’t see a whole lot of short-term upside to bonds – even if there is a SORR event in the near future. So I’d rather just keep the cash for the flexibility. My current plan is a cash/equity split at the start (presently something like 90/10, but with the cash portion increasing through retirement date, and ending up more like 70/30). Then I just use the cash to fund the first several years of retirement – effectively functioning as a glidepath, while leaving the equities alone during that time.
Curious to hear any views on this type of approach.
I like your approach and it’s similar to what I’m thinking about. When I hear talk about bonds and how the returns are negatively correlated with stocks my immediate reaction is: OK precisely what returns are we talking about? The 2% nominal on 10 year treasuries? That’s not a return, that’s chump-change! And it appears that bonds aren’t even all that negatively correlated. So if I’m going to get crummy returns and little correlation advantage, then why not just keep investments in cash in case something better comes along?
You’re definitely taking the contrarian approach by going cash. And I’m a proud cash coward. So you get two thumbs up from me.
While we’re talking about correlation, I saw a very interesting article on Charles Schwab’s website about how international stock market returns are less correlated with the US than they’ve been in quite some time. So why have bonds if you can offset with other stocks. Maybe if we’re really good and say pretty please, Big ERN will comment about international stock correlation.
Again: bonds/cash are used to simply cushion the drop in equities.
Also the appeal of bonds is that in case of another recession, the Fed will be aggressive again and drop the yield. You’re earning from the duration effect. Not much potential, of course with yields so low already.
Also: I see ZERO potential to protect from the next big stock drop through international stocks. If we have a bear market here, all your international stocks will get hammered. Happened in every single past event:
We are essentially doing the same thing! We are one year away and have shifted from 90/10 stocks/bonds to 63/25/12 stocks/cash/bonds and plan to use the cash to start without touching the equities. I thought I was crazy thinking like this but glad to see others thinking along the same lines.
With bond yields so low and only limited potential for bonds to appreciate if yields go down, I can clearly see your point. See my caveat #2 in my final thoughts.
So, funding the first few years with a cash reserve instead of bonds is clearly a viable option.
I am in a similar situation as 40ORFIGHT commented above – I retired at 52 but am not even holding 60% equities. I am inclined to hold my current high cash position until we see a downturn but of course that is a bit of a fools game given that no one knows when that may happen. However, I am very risk averse (a condition I am trying to overcome :)). At the time of a downturn, I can move more aggressively into equites such that I am basically creating the glide path discussed over the next 10 years. Thoughts?
Uh oh. Now you’ve done it. Spoken heresy. And by my count there seem to be at least 3 cash rebels here tonite. See what happens when you run a cutting edge finance blog? You can’t even go on vacation without the posters running amok!
Market timing is a fool’s game. Except when it isn’t.
Just hard to get my head around putting a lot of new $ to work with these high valuations. Agree with the comments about bond returns. Would prefer to just hold cash. Even selling premium in options in this environment is a challenge but does enable some opportunities.
You’re not the only one.
Big ERN posted a link some time ago to the star capital website out of Germany. It shows comparative ratios between world equities markets. If you have the chance, take a look and see if anything leaps off the page.
(The anti spam software seems to not be liking me tonite. Sorry for not posting a link)
We’re joining this party too. I’ve commented above but we are reaching FIRE in about a year and have shifted from 90/10 stocks/bonds to 63/25/12 stocks/cash/bonds with the cash sitting in a high interest savings account earning 2.8%. Our plan is to use that for the first few years to not have to tap into our equities if there’s a market tank. But we’re also super conservative and looking at a 2.5% SWR so we know we’re crazy 😉
Crazy like a fox. I’d be more than happy to retire at a 2.5% WR – it’s just that pesky problem of continuing to work until that point…
Depends on where you are in the accumulation phase. If you’re starting out I wouldn’t try to time the market. Example: I got my first job in 2000 right around the market peak. I got a big pay raise in 2008 right around that peak. And I stubbornly invested in stocks, regularly! Worked out extremely well for me.
Just to be clear, I’m only talking about cash instead of bonds for purposes of the glidepath (and for a relatively short time-horizon). Overall, I’m still very much in favor or equities for the long haul – and would plan to return to almost pure equities after drawing down the cash cushion at the start.
I too am interested to see what our experts have to say about this approach. A downside is losing the gains on bonds during any SRR event (which would be plowed back into equities at low prices) – but I’m still not convinced that will be the scenario given our current bond environment.
I agree that having some cash provides flexibility. And it probably also leads to some inevitable market timing attempts. Frankly, it would be hard for me not to want to dump some of the cash into equities down 40% like the recession.
I think it’s safe to assume that current company is OK with equities. At least in theory. I personally think equities are a good investment. Just so long as they’re priced to allow for a reasonable rate of return in the long term. And therein lies the problem.
As I said in my final comment #2: in the case of a 1970s supply shock with an inflation shock the bond-GP would be less effective (though likely not ineffective) against SoRR. With a cash-GP you would be better hedged against such a scenario, so no complaints from my side if folks prefer cash over bonds.
If you have the stomach to move “aggressively into equities” that would work. But ask yourself: Did you do so already in December 2018 when the S&P dropped into the 2300s? Or were you scared out of your mind and didn’t dare to move in? That’s the problem a lot of retail investors have with tactical asset allocation. 🙂
Dr. David Graham and Karsten,
Ever since reading about a glidepath, I’ve always wondered how do we implement this in practicality?
Hypothetically, lets say that one would need $1,000,000 in the bond portion as the 40% (this is AUD in case anone things I’m fatFIRE). Once you’re $1MM away from your FIRE number, do you then start to accumulate bond assets with contributions? If so, would this not delay FIRE since we wouldn’t have the benefit of equity performance to help us reach our FIRE number sooner? Is cushioning the SORR, the acceptable trade-off?
If instead, we rebalance by buying/selling assets to achieve the 60/40 portfolio then we would be subjected to capital gains taxes.
Sorry if my question isn’t entirely clear.
Depends on your flexibility in the timing of retirement:
1: you want to reach FIRE as quickly as possible and have flexibility with your date. By all means, go 100% equities (highest expected return) and shift $1MM when you get there (maybe in several smaller tranches).
2: you have a fixed retirement date and want to hedge against the SoRR of a market crash right around that time. Use the bond tent as suggested by Kitces.
If under #1 you’d face cap gains taxes you might be forced to do #2. But most of us have large chunks of money in tax-advantaged accounts where we can rebalance tax-free.
I see now where I was getting confused. The retirement accounts, i.e. Superannuation, in Australia works quite differently. I have been grappling with this fact as well since most FIRE materials are US-centric when it comes to taking tax and retirement accounts into consideration.
Aussies, can’t access their retirement funds until 60 and the rules around it change periodically making it difficult to accurately plan. I am not good enough to work out if I should be contributing to a locked fund when I plan to retire circa 40 years old. My gut feeling is there is some % of my assets that should be going to this fund but I’m unsure what amount. I do not think it’s the same mantra as maxing our your 401K.
Also, most small players only have access to managed funds that do not allow much freedom in asset allocation.
Perhaps this is one thing I need to hire a professional about? How much is your consulting fee (I am a bit afraid to find out)?
I’m not in the business (yet) to offer for-fee advice on this, so here’s my five-minute version for free:
1: You’d have to accumulate enough money in your non-retirement-accounts to bridge the the 20-year period between ages 40 and 60.
2: But money is also fungible. If you want to do the bond tent/glidepath you can do the rebalancing of the portfolio in the tax-deferred account without tax consequences due to realized capital gains.
So, maybe keep tax-efficient stocks in your regular brokerage account and then do 2 SWR analyses: 1: the overall SWR analysis for your overall portfolio over 50+ years. 2: a SWR analysis of a 20-year portfolio with 100% stocks to make sure that portion lasts for 20 years.
Thanks Karsten. You’re the gift that keeps on giving. Not only do you have the most rigorously-backed FIRE articles, you’re also so engaged with us in the comments.
Your five-minute version certainly helps a lot and gives me somewhere to start whereas before, I was lost as how to even approach it.
I developed a different approach. Instead of varying AA, I created 2 portfolios, a small portfolio and a big portfolio. The small portfolio consists of 3 years of WR and is risked at 15/85 VTi / BND. It’s expected return is 6%. The big portfolio can be risked at 60/40 at retirement and then can glide up 10 years later. If you expect to retire in 20 years and you need 120K to live for 3 years you can fund that with 4K/yr x 11 years and then let that ride to 20 years. So for 44K in a 15/85 you will have the small. If you want fund it at 80/20 and then switch to 15/85 and it will take less to fund.
The large you fund as you would otherwise fund your retirement so if you choose 1M fund to 1M. Upon retirement then you would have 1M big and 120K small. Lets say it’s Dec 1999 you choose to retire and the market dies for 3 years. Do you spend the big or the little? You close the big portfolio to all withdrawal and just assiduously re-balance while the market is down. You live off the small. Since the small is only 15% stocks it’s most;y bonds and you have 93% of your money to live on. If you reduce the Portfolio by 7% then you have 37K/yr to live on for 3 years. Most recessions are under 3 years so it’s likely you won’t spend all the money. The closed Big portfolio suffers sequence risk from the market drop but NOT sequence of return risk which includes withdrawal aka selling low. Re-balancing supercharges the portfolio recovery when the market turns.
The 60/40 portfolio returns to pre-crash highs faster than a more risky portfolio. In 2008 a 100% stock portfolio took 5 years to recover its pre-crash highs. A 60/40 took 3 years and by 5 years was 15% ahead so reduced risk peri-retirement is important protection. The small portfolio absolves the big from early SORR. The trajectory of the big following the crash is strongly up. The small portfolio effectively re-indexes the retirement date from Dec 1999 to 2003. I ran this scenario on several crashes including 1973. 2000 and 2008 and every portfolio recovered with an up trajectory compared to the bad SORR cohort. So re-indexing the big portfolio to a different sequence is another technique. Let’s say you pass 10 years and never need to use the small. In 10 years the small would grow to 200K. You can then combine small and big to a single portfolio and glide up to whatever you like. You only need to protect the portfolio once, so unlike a bucket theory you never refill the bucket. You either use it once or dump the bucket back into the big if you don’t use it so it supercharges your return 10 years into retirement going forward. Each succeeding year post retirement the risk of SORR drops slightly so if the crash comes 5 years in the effect is smaller than 1 year in. A 50 year retirement can likely withstand a little higher WR than a 60 year retirement so likely 3.5% over 50 years as opposed to a constant 3.25% would work using the big/small approach.
As the Veterinarian on FIRE would say, There are many ways to skin a cat! That two-account approach is very similar in spririt to an “active glidepath” I mentioned in Parts 19/20.
I looked at some additional features: What if Sequence of Return Risk starts later in retirement? How does a 10 year glidepath compare to a 20 year? How do both compare to a 60/40 portfolio? What about with high or low return assumptions? https://www.fiphysician.com/modeling-rising-equity-glidepaths-over-a-60-year-retirement/
I looked at that too and I found SORR was length of retirement dependent. A normal 30 year retirement suggests 5 years prior and 10 years post, for 15 years out of 35 year time period to worry about SOR early in retirement. I found by extending the retirement period a ratio of sensitivity was maintained. So for a 60 year retirement around 20 years of SOR exposure was significant. If you use my one shot portfolio method you would own the 2 portfolios for 20 years into retirement before re-risking the less risky money. In this scenario SS also comes into play. if you retire at 40, 25-30 years in you can claim SS which reduces WR on the portfolio and a reduced WR is less sensitive to SOR so the portfolio becomes more robust and could withstand re-risking the low risk 15/85 to whatever you choose giving the portfolio a shot in the arm going forward. There are 2 ways to get rich one is adding principal and the other is adding risk.
I read a paper that looked at “correct” portfolio risk in the last half of retirement and their conclusion was the “efficient risk”, (similar to the tangent risk on the Efficient Frontier) was about 75/25. Over time that gave you the most return for the least risk. Lower risk like 60/40 gave you less return, higher risk like 100% gave you lower return because of disproportionate risk vs reward. There is always a time factor associated with risk and the time factor is the amount of time it takes to get back to zero following a crash. When you are getting back to zero you are not growing. When you are retired the end game is death and even a 60 year retirement ends in death, so there is a hard boundary to the amount of time you have to screw around with recoveries from crashes.
Hi ERN! This question may be off topic to this post…but how did you do with your options that expired Monday the 5th? I’ve been following your work for awhile now and was happy to find you since your 2 DTE trades mirror how I trade when the VIX is over 17.5. Since that big of a move is exceedingly rare, I’m curious how you managed your positions. Thanks in advance!
Two went in the money. But only by about $5 (strike=2850, index finished at 2844.74).
Funny thing is that they were the ones that I sold on Thursday 8/1. All the options I sold on Friday were OK. Again, the ones where I “broke the rules” of my strategy (rolling 1 day before the expiration of the old options) were the ones that lost money…
Thanks ERN! I was in EXACTLY the same situation. The Thursday options I broke my own rule and they were the only ones to lose. Fortunately since then I have gained back almost all of what I lost, and gained a valuable lesson from all of this. Thanks again for all you do here!
Ha, I’m not surprised I’m not the only one. Was too tempting to start writing some options that Thursday! 🙂
Wow, thank you Big ERN and Dr. David Graham for this extremely useful perspective!
If I may ask a follow up question — given your thoughts on holding bonds and a mortgage simultaneously (another fantastic article, by the way), do you think it is prudent to factor our home equity into the fixed income side of the equation as we approach our FI date and pay our house off simultaneously?
If context helps… my wife and I are exactly the archetypes you describe — 5 years away from our our anticipated FI date (assuming a 3.4% SWR, maintaining our current ~70% savings rate, and no major market corrections in the interim).
Many thanks for your careful and thoughtful work. We LOVE your blog and your broader contributions to the community.
Good point. I’d only include the assets outside the home equity when calculating the withdrawal amount = assets x SWR.
But one should include the home value/equity in the estate.
ERN, what’s your opinon on the following; to protect against Sequence Risk, would it make sense to buy long-term index put options to “protect” the first years of retirement? Eg. LEAPS for each of the first X years of retirement. With current high valuations, you can probably buy relatively far-OTM puts to minimize costs, and when markets tank, sell (not write!) put options to cover the required 3,5-4% income, instead of selling depreciated shares. The idea is similar to the Equity Glide Path: sacrifice some returns for downside protection, so I wonder how they would compare.
Great point!!! It’s something I’ve been mulling over as well. But downside protection is so expensive. It’s so expensive that I sell puts:
(also: writing a put = selling a put. Or did I miss some subtle difference?)
But you got a point! Your calendar spread, i.e., buying long-dated puts and making up the premium with selling short-dated puts (with higher yield) could work. Not in all market environments, but most of the time…
No I actually meant selling puts that you would have in portfolio (=sell to close), not writing new puts (= sell to open).
If you would buy long term OTM puts now, you could sell them back (thus, not write) later when/if markets tank, at much higher price than you bought them. If markets don’t tank, you lose time value, but that’s the cost of insurance (the Equity Glide Path strategy also sacrifices some returns).
Regarding “expensive”, you can choose the price of your downside protection by choosing an appropriate strike price. You could calculate at what price point (of an index) the SWR would eat too much of your portfolio and buy puts at that level. If they are far enough OTM (given current valuations), it could be relatively cheap? And yield far more than 4% when you need it…
Ah, OK, gotcha. In that case I don’t think I could endorse that so wholeheartedly. Buying insurance is way too expensive. In the worst case, if the market underpeformance is really long and gradual (1965-1973) you’d even make things worse: you pay for insurance, it never pays you back and make Sequence Risk worse.
This would only work in a 2007-09 sharp decline.
.. and such short and sharp declines are actually not so bad in terms of Sequence Risk, compared to the long and gradual ones. Good argument, thank you!
Have you thought about adding health insurance calculations into your results? Since health insurance is rising faster than inflation, it seems like that might be the one cost that could cause problems longer term.
Yes! I certainly take that into account in my personal calculation. I use the Google Sheet and add supplemental costs to take that into account, especially once we get older.
With bonds yielding nearly nothing, I’ve been hearing more about an algorithmic strategy where one keeps 3 years of expenses in cash. Spending comes from the 22X expenses stock portfolio UNLESS the market drops X%. If that happens, you stop selling stock and start spending down the cash portion of your portfolio until stocks rise again. You replenish the cash account over time as the bull market takes over.
Might be harder to model or do Monte Carlo simulations on, but it seems like this cash-shock-absorber strategy could further improve portfolio survivability compared to the glide path which assumes withdraws are proportional to AA.
After all, bond funds might crash harder than stocks if we have an inflationary supply shock.
Sounds very much like a cash bucket (as simulated in part 25).
It helps some but it’s not fool-proof solution. Three years of cash bucket would be way too little to make it through a lot of the past recession. See Part 12, especially this chart:
The bottom chart: the red shaded areas are the length of major drawdowns: to reach the old high plus inflation. 3 years won’t cut it! 🙂
Has anybody tested a rising equity glide path from a 30/70 equity/fixed asset allocation (AA?)
After all, that AA is what one pundit now recommends for those who have just retired.
Versus starting with a 60/40 AA as in this post?
I haven’t posted it for the large audience to view, but if you want to play around with the Google Sheet and look at the tab “Case Study” you can check how a GP with any starting/ending weights would look like. Unfortunately only for one ficed starting date.
But if you want to look at the obvious worst case scenarios:
anywhere around 1965/66
You will see how it performs! 🙂