Forum

Policies – please read!

  • Everyone can read posts, but to start a new topic or add a reply you’d need to register. To create an ID, please verify your email address. You will get a link at that email address and you can then pick a password.
  • Be courteous to others. Treat others like you like to be treated. Discuss the issues. No ad-hominem attacks!
  • It’s OK to include external links if they are relevant. But please avoid spamming! Please no affiliate links.
  • Before starting a new topic, please check if that question/topic has been discussed before already and add to that discussion instead of starting a new topic.
  • If you’ve written a cool blog post that you want to share with others please post this in the “Self-Promo” category only!
  • Please read the usual Disclaimers and the Privacy Policy!
  • The forum policies may be amended in the future!
Hey, I just “saved”...
 
Notifications
Clear all

Hey, I just “saved” the 4% Rule for 60 year Retirements* (ratchets, guardrails, variable equity allocation)

17 Posts
2 Users
2 Likes
5,249 Views
andyg42
Posts: 27
Topic starter
(@andyg42)
Eminent Member
Joined: 9 months ago

I just found the Holy Grail. I solved the Fermat’s Last Theorem of FIRE: I’ve found a way to “save” the 4% rule, even for 60 year retirements*. 🙂 

 

Exec Summary: Hacking ERN’s SWR sheet, by using a combination of simple “guardrail” rules, replacing bonds with gold as the hedge/ballast for stocks, and replacing a mechanical pre-planned 10-year glidepath with a variable equity allocation based on S&P nominal and real values, % down from high, trough values since the most recent peak, and the number of months since the most recent stock trough, I show that you can have a 4% WR* together with a real Final Value >= 100% of initial portfolio after 30 years (and thus a solution that would work for a 60 year+ retirement).

 

In fact, for a 30 year retirement, if like myself you don’t worry about the possibly of sustained deflation (such as what happened during the Great Depression, pre-“Helicopter Ben” Fed policy), you can use a 4.5% WR and have a > 80% Final Value, leaving plenty of margin for error (or legacy).

 

* Just as with any “flexible” system that can include spending cuts, there can be some number of months with spending below the 4% of initial portfolio amount. But I will show that, especially with a 4% IWR, the total number of months with such cuts is fairly low

 

 

 The approach proposed combines “guardrails” with a functional alternative to a glidepath, and allows increased spending/consumption whenever the real portfolio value hits a new high.

 

The guardrail rules do not have the problem ERN wrote about with Guyton-Klinger, as there are only a maximum of two 10% cuts ever from any peak consumption amount; further, the 2nd cut happens rarely and even more rarely for very long. In fact, you could probably make the numbers work even if the 2nd cut were eliminated (I prefer having it there because it feels more “real world” appropriate, and helps protect against a sequence actually worse than anything ever experienced)

 

 

I've done this by hacking the "Case Study" tab of ERN’s SWR sheet (the tab that has the glidepath) for a given cohort retirement month, and then tested the dozen or so worst cohort months to verify.

 

There are 3 key components 

A. “Guardrails” A simplified G-K-like rule set ("guardrails") that avoids the problems ERN wrote about to:

  1. Ratchet up spending whenever real portfolio value increases ["retire again"]
  2. Cut spending by 10% from the latest baseline if real portfolio value down 35%+ from the current peak in the first 15 years of a 30 year retirement sequence.
  3. Cut spending by an additional 10% of baseline if real portfolio value down by 50%+ from current peak in the first 20 years

 

To obtain the spectacular results I show, I am using gold rather than bonds for the non-equity portion of the portfolio. I will explain/defend this no doubt “controversial” decision separately.

 

B. Calculate what the new equity allocation should be for the following month, with a two step process.

Step 1: Do a fairly simple calculation where the next month's stock allocation is a function of (the six month moving average of) how much stocks are down from their high. Where x = 6-month-average of % down from high: the calculation is equity alloc = initial_equity-baseline + if(x< 25%, 1/2*x, x-13%)

 

Step 2: Potentially modify that above equity allocation to:

  1. Maintain momentum: if stocks close to their all-time high, don't reduce equity allocation from prior month
    • …and go 5% over baseline if also near real peak
  2. Avoid catching a falling knife: Reduce equity allocation substantially (20 or 30 percentage points below baseline!) when stocks fall 10% below the high and keep the equity allocation low for a few months (4 typically) past the most recent stocks trough
  3. Press equity bets for a long while after a trough: Don’t reduce equity allocation below prior if within 21 months of most recent trough unless stocks hit a new peak

 

Note that the remaining non-stock components of the portfolio allocation maintain their internal relative weights for the non-equity portion of the portfolio.

Also note that since the amount of equities that need to be sold some months can be substantial (the “avoid catching a falling knife” part), it will likely only make sense to implement if that move into and out of equities can be done in a tax-advantaged account. 

 

I do all the above by adding columns to the hacked Case Study tab that calculate:

  • S&P max to date
  • Real S&P value
  • Real S&P max to date
  • % S&P down from real high
  • % S&P down from nominal high
  • Trough since last S&P (nominal) peak
  • Months since last trough
  • % portfolio down from its high

 

 

To verify, I tested against the dozen worst retirement cohorts since 1926, all corresponding to stock market (nominal an/or real) peaks:

  • Sept 1929; Mar 1937; Feb 1946; Aug 1956; Nov 1965; Dec 1968; July 1973; Mar 1974; Mar 1979; Feb 1980; Dec 1980; Jan 2001

 

Using a 75-0-25 (stocks-bonds-gold) initial allocation and a 4% WR, per above all scenarios return > 100% Final Value after 30 years. All but 1937 had average consumption numbers exceeding 4%. March 1937 resulted in an average of 3.96% - still rely close.

 

September 1929 and March 1937 were definitely the most challenging periods. For the September 1929 cohort (average consumption: 4.53%), there were 46 months with consumption below the baseline, and 7 months that were 11%-20% below. The March 1937 cohort (3.96%) had fully 123 months below baseline and 17 months that were 11%-20% below. 

[Per above, I’m personally less interested in and wasn’t trying to optimize for these two “sustained deflation” cases. But even here, the results are not at all terrible.]

 

No other cohorts ever had months > 10% below baseline. January 2000 (4.41% average) had 16 months below baseline. November 1965 (4.16% - the 2nd lowest) had 9 months below. No other cohorts had more than 6 months below baseline.

 

Each scenario shows a final consumption value of at least 4.6%.  This is consistent with my separate finding that at a 4.5% WR, a 30 year retirement works with at least 80% Final Value for all except the 1929 and 1937 (sustained deflation) cohorts [which still “succeed”, but with much lower FVs).  More on this in a follow-up post.

 

 

It’s clear to me that the two “controversial” parts of my proposal will be 1) the use of gold rather than bonds as hedge against stocks, and 2) the variable equity allocation, and in particular the “avoid catching a falling knife” component, is clearly a measure of market timing. I will cover these each in separate forum posts, strongly defending the first, while only weakly defending the second. 

 

Even though it seems at least as sensible as a fixed glidepath approach, I’m particularly looking for feedback and for people to point out problems with the variable equity allocation proposal. Most obviously, if done in a non-tax-advantaged account, the occasional “churn” could result in hefty tax bills.

 

 

Unfortunately, the hacked sheet itself is not yet quite fit for consumption by others. Big ERN, as the creator of the original, could probably figure it out, but he might be the only one. I will work to clean up the hacked tab so that it’s somewhat close to the same readability as ERN’s original and then post it here.

 

Some other commenter on one of ERN’s posts noted that the SWR sheet had cost him/her hours of sleep. That’s definitely been true for me the last few weeks - and I’ve loved it!

16 Replies
andyg42
Posts: 27
Topic starter
(@andyg42)
Eminent Member
Joined: 9 months ago

Why Gold Not Bonds

 

Why am I using gold rather than bonds for the non-equity portion of the portfolio? Here are 5 reasons:

 

  1. Bonds as a hedge against stocks are only important to handle deflation. While none of us knows exactly what will happen in the future, the one thing I am simply not at all worried about is the possibility of sustained deflation, such as what happened during the Great Depression. While I’m not the biggest fan of the Fed, they certainly know how to – and that they need to – expand the money supply to ensure we don’t have sustained deflation. This is the famous “Helicopter Ben” point (the original idea is from Milton Friedman):
  2. Other than for sustained deflation, gold makes a much better hedge against stocks. In the last 70 years, the two decades that stocks did poorly - the 1970s and the 2000s - are the two decades gold performed very well. 
  3. It works! When used as the ballast for equities, even for static allocations, ERN’s SWR sheet shows this (look at the numbers for 1950-) for periods other than those with sustained deflation. And in doing my research with the hacked sheet, gold does decently enough when combined with the variable equity allocation even for the Great Depression period, while being far superior for most years since 1945. It actually does quite well for the 1929 cohort (not as well for the 1937 cohort). By contrast, using bonds as the hedge has big problems whether using static allocation or my variable one, if trying to use a 4% or greater WR. Bonds are disastrous for mid-1960s retirees, of course. But they’re also not great for January 2000 or August 1956 cohorts either.
  4. Previous ERN posts show that bonds are actually not that great an equities hedge while being risky themselves, especially in the current low rate environment (Part 2, Part 3, and here), while gold is quite helpful (Part 34) as a hedge for stocks
  5. Despite running most of his research simulations using bonds as the hedge for stocks, ERN himself doesn’t actually hold any bonds in his own asset allocation (he uses options trading and real estate instead)

 

I understand that most studies use bonds as the hedge against stocks. And I understand that most studies use data since 1926, and that includes the Great Depression, where bonds are an ideal hedge and gold was not as good.  But just as folks (ERN included) usually don’t focus on the 1871-1925 period because it’s no longer representative of how our world works, I submit that thanks to improved central bank understanding and policies, when it comes to the issue of deflation 1929-1945 is no longer representative either.

 

Reply
earlyretirementnow.com
Posts: 324
(@earlyretirementnowcom)
Member
Joined: 7 years ago

Wow, that's awesome!

Do you have some detailed simulations for the 9/1929 and 11/1965 cohorts and how they would have fared?

Reply
1 Reply
andyg42
(@andyg42)
Joined: 9 months ago

Eminent Member
Posts: 27

@earlyretirementnowcom For a 75-0-25 stocks-bonds-gold initial allocation with the variable equity allocation as described above

For 4.0% initial WR, $1M portfolio, 360 months (consumption $ figures annualized):

1929-9:

Final Value: $1.028M

Months below baseline consumption: 46

Months 11%-20% below baseline: 7

Average consumption: $45,267

Final consumption: $49,055

Lowest consumption: $32,000

 

1965-11:

Final Value: $1.054M

Months below baseline consumption: 9

Months 11%-20% below baseline: 0

Average consumption: $41,565

Final consumption: $45,954

Lowest consumption: $36,000

 

To be fair, 1937-3 fares worst:

Final Value: $1.024M

Months below baseline consumption: 123

Months 11%-20% below baseline: 17

Average consumption: $39,581

Final consumption: $47,987

Lowest consumption: $32,000

 

 

 

 

Reply
andyg42
Posts: 27
Topic starter
(@andyg42)
Eminent Member
Joined: 9 months ago

Ok, time to "show my work".

Hacked spreadsheet attached, .xlsx format

I cloned the Case Study tab - see the tab called AG-Case-Study

 

The variables that can be entered are in purple on orange background

 

The relevant output is in green in cells P14:P20.

 

I'll post some accumulated results for the given most difficult cohorts using 4.0% and 4.5% initial WRs in a bit

 

Reply
1 Reply
andyg42
(@andyg42)
Joined: 9 months ago

Eminent Member
Posts: 27

Ok, attached is some results for 4% and 4.5% WRs using the hacked sheet attached above on the dozen or so worst retirement cohorts.

For each tab, top shows Final Values after 360 months, while for those variants using the guardrails to ratchet/cut spending, the bottom covers average spending, final, worst case, as well as number of months above, at, below, and > 10% below the original baseline (real) spending

 

I did some rechecking of the numbers, but as they were all entered manually based on reading off the hacked sheet, and done over the course of a few days, I don't guarantee accuracy. If you spot a problem, please let me know

Reply
andyg42
Posts: 27
Topic starter
(@andyg42)
Eminent Member
Joined: 9 months ago

Is Variable Equity Allocation “Ok”?

 

I was all set to write something longer about why doing market timing in the context of SWRs is “ok”, when I came across this and this and this response in the comments section of an ERN post on Market Timing and Risk Management, Part 2 – Momentum, which address the issue here better and more concisely than anything I could come up with on my own:

  • “most people would be willing to forego some upside in exchange for hedging the sequence risk. In other words, the SWR will be lower most of the time, but only during times when it was really high. But you will likely lift some of the extreme low outliers during the Great Depression, 1970s/80s and 2000s.”
  • “I’m willing to underperform when the stock market rallies if this provides insurance for the big disaster scenarios when the 4% rule failed!”
  • “Momentum underperforms in the periods when the market is rallying (60/40 would even more so). But a lot of people are willing to forego a little bit on the upside in exchange for the lower vol and lower drawdowns. That’s the whole purpose of risk management.”

 

And who exactly was it who made these comments? Some random crank? No, in fact it was ERN himself!

 

Truth is, while I have devoured every word of the SWR series, and most of the comments sections as well, I had never seen that post on Momentum until this weekend.  But the logic is impeccable! The whole point of mitigating SORR is handling those worst few retirement cohorts.

 

I actually sort of “back-doored” myself into coming up with my crude timing system, in my attempt to do something that was better than a static glidepath. Because it was/is very clear to me that using a static 60 to 100% equities glidepath, and then holding equities forever at 100% of your portfolio for the remaining 20 to 50+ years of retirement is in no sense the “safe” thing to do. 

 

While it “works” in the sense of not depleting the portfolio because you’re probably drawing <3% annually after 10 years for most cohorts using constant consumption, it’s pretty obvious that unless your goal is to leave the highest average legacy amount when you die, maintaining equities at 100% indefinitely only makes sense if stocks are still 30%+ underwater. Otherwise, you continue to experience SOR risk.

 

Per the numbers and hacked spreadsheet attached above, it’s pretty clear that doing this kind of market timing works really well in addressing SORR and enabling meaningfully higher SWRs.

 

 

Which is why my remaining question is: how is it that ERN hasn’t addressed this within the SWR series yet? 🙂

Reply
Page 1 / 3
Share: