NesteggCycle: Engineered Crash Protection P2

Nestegg Cycle © The Art of Engineered Crash Protection
Part 2: Can we Do Better than Guyton's Guard Rails?

Posting Date: July 19, 2026 By: Fred P Davidson, Founder of NesteggCycle.com

SEE ALSO: Engineered_Crash_Protection Part 1: Lost Decades

INTRODUCTION:
Consider a retirement planned to run 40 years, paying an inflation-adjusted 55,000 dollars per year.
We start with one million dollars and expect to earn 7.00% NET of inflation.

Unfortunately, we encounter a perfectly STAGNANT, FLAT market for the first five years.
We earn nothing during this time, but market values are neither up nor down. How does this play out?

IF we had perfect foresight, we'd know that this scenario is completely survivable at full payouts:
In the 5 initial flat years, we pull 5 times 55,000 = 275,000 out of the million, leaving 725,000.
The remaining 725,000 regains its 7.00% net return and successfully pays 55,000 per year for the remaining 35 years, leaving a legacy of about 137,000 at the end, after paying out 2.2 million dollars in 40 years.

But what about management of these payouts in real time without the benefit of perfect foresight?
Now we have to treat an unexpected shortfall in our balance or earnings with much greater caution,
since we want to steer clear of catastrophic early collapse of principal.

It should be pointed out that the Guyton-Klinger Guard Rails rules of 2006
often perform well enough, and that we are looking for the outliers.

I have simulated the above 40 year retirement under (my simplification of) these Guard Rails (GR) rules,
and what we see is a surprisingly poor job, given what we've just shown is possible in this case.

Assuming 3% inflation (see below, "Other important stipulations")
the payout drops about 3% year by year for 5 years after the initial 55,000, plateauing at 47443.48
for years #6 thru #33, and yes, that is 28 years,
then steps up to 52187.83 for years #34 thru #37, and finishes the final 3 years paying 57406.61.

We've paid out only 1.97 million (less than 90% of above), and left a legacy of 1.29 million (over 9 times above).

Only four of the 40 years boasted payouts at or above the "par" amount of 55,000 per year, while there were 28 years where the payout was only 86.3% of "par".

So the question becomes: CAN we devise an algorithm which will do better than that?
As we work through this posting, you will come to see that YES we CAN!

NOTE that we use a specialized definition of a "crash":
This note is paraphrased from my 2024 first posting about crash-points:
Portfolio Stress Testing
Our specialized crash occurs suddenly and completely on the morning of Day One of your retirement.
It is a synthetic construct, simpler than a real crash. A flat line, there is no jagged-line volatility.
This simplifies exploration of effects that might occur.
Our crash minimally means the LOSS of your net YIELD: You stop gaining against inflation.

This means we can have a "crash to 100%" at which you still have the whole million, but no net earnings.
The "crash to 100%" test provides a simple yet strong first impression of the durability of this spending scenario!

A "crash to 80%" would additionally mean that the million suddenly dropped to 800000 in value, on Day One.

Our crash recovers instantaneously, exactly on a retirement anniversary, by applying the inverse fraction to the remaining balance. So in the case of a crash to 80% (0.80), the balance remaining after payouts at end of the crash, gets multiplied by (1 / 0.80), which is 1.25.
After that, the portfolio yield (as entered by you) resumes immediately and until the end of the term.

DISCUSSION: The "Three Coulds"

To get things started, let us "instrument" each row of a 40 year Guyton GR rules simulation,
so that in addition to the current Guard-Rails Payout, Balance, and NEXT Guard-Rails Payout,
we also show three annuity calculations, projecting possible ideal-condition payouts as follows:

Could1 = Annual payout at 7.00% net yield, exactly exhausting the balance at end of the retirement term.
Could2 = Like Could1, but reduced enough to extend payouts for 6 years beyond end of the retirement term.
Could3 = Like Could2, but reduced further to extend payouts for 12 years beyond end of the retirement term.

NOTE: Early into writing this posting (April 2026), it looked to me that we might need to add as many as 12 years to get a sensible payout, but subsequent analysis (late June 2026) suggests that about 4 years is good enough as the add-on for the "Could" calculation, and indeed, comes much closer to the initially postulated "perfect foresight" situation above.

Four years provides a modestly larger payout but a much smaller Legacy compared with 12 years.
IF you are looking for more Legacy,
stay with a larger number of years added on to your "Could" calculation.

Other important stipulations here:
The initial balance is one million dollars;
During the first 5 years, we are gently "crashed to 100%" such that the 7% net yield is suspended, but the market prices of securties remain level;
Inflation is a constant 3% per year;
Since we are presenting inflation-adjusted amounts, and since GR suspends the COLA when there's no income, we report this as a payout reduction:
new_payout = previous_payout / 1.03 = previous_payout times 0.970874;

	
	This Run Crash Length entered as: < 5 >
	This Run Crash_To Percentage entered as: < 100 >
	
	        SCENARIO -- years_crashed:  5 /  years_recovered: 35
	
  Yr   GR-Payout     Balance      GR-Next     Could1      Could2      Could3

   0   55,000.00     1000000.00   55,000.00   75,009.14   73,259.96   72,139.01

   1   55,000.00      945000.00   53,398.06   71,240.49   69,457.09   68,317.50   NO COLA
   2   53,398.06      891601.94   51,842.78   67,579.02   65,762.50   64,605.34   NO COLA
   3   51,842.78      839759.17   50,332.79   64,020.59   62,172.32   60,998.87   NO COLA
   4   50,332.79      789426.38   48,866.79   60,561.09   58,682.73   57,494.47   NO COLA
   5   48,866.79      740559.59   47,443.48   57,196.35   55,289.90   54,088.58   NO COLA

        SUM of IN-CRASH payouts: 259,440.41

Some observations from this table:
Under ideal conditions (NO crash, get 7% net every year), as much as 75,009/year could be paid out initially,
reduced to 72,139/year to extend payouts for extra 12 years;

In fact, if the first 5 years were all "average" instead of "crashed", the safe (Could3) payout would rise to just over 79,000 per year, while GR rules would hold the payout to 55,000 for about 7 more years after that, then specifying a 10% raise to 60,500 per year.

But see how quickly the possibilities deteriorate once this "crash" starts;
After 3 crashed years, the limit is down to 64,021 (Could1) annually, with the safer (Could2, Could3) payouts reduced to 62,172 and 60,999.
After 5 crashed years, the limit is down to 57,196 (Could1) annually, with the safer (Could2, Could3) payouts reduced to 55,290 and 54,089;
The initial reserves have been almost completely depleted;

The Guard-Rails rules have correctly sensed pressure to reduce the payouts, but have been too harsh, taking us down to 47,443/year after 5 gently crashed years;

What do we mean by the Initial Reserves?
This will be fleshed out in a future posting but for now, note that under average non-crashed conditions, 40 years of 55,000 dollar payouts could be provided with a principal amount of 733243.99 earning 7.00% net; but we've stipulated a starting amount of 1 million, so this excess of 266756.01 is an initial reserve, giving us some flexibility against contingencies.

Let's explore the post-crash Recovery:
Here we've added a new column, "C3/GR", on the right: the value is Could3 divided by the GR-Payout, so as to highlight
the growing discrepancy between the GR-rules Payout, versus the "Could3" rational payout.

As there are many years with unchanged GR-Payout, we omit some rows for a more concise presentation.

	
  Yr   GR-Payout     Balance      GR-Next      Could1       Could2       Could3       C3/GR
  
   6   47,443.48      744955.28   47,443.48    57,955.09    55,878.45    54,575.40    1.15032

  16   47,443.48      809939.37   47,443.48    70,617.82    65,270.10    62,134.75    1.30966

  26   47,443.48      937772.91   47,443.48   107,229.59    88,519.13    79,299.04    1.67144
   
  33   47,443.48     1095281.46   52,187.83   203,233.00   131,051.36   105,971.78    2.23364
  34   52,187.83     1119763.33   52,187.83   234,921.64   140,980.43   111,318.59    2.13304

  37   52,187.83     1203979.57   57,406.61   458,778.42   184,794.57   132,190.49    2.53298
  38   57,406.61     1230851.53   57,406.61   680,773.87   206,127.95   140,741.64    2.45166

  40   57,406.61     1290370.22   57,406.61       1.00     270,714.25   162,460.18    2.82999

GR_mini_crasher_01 retro-fit:  SUM of payouts: 1,968,829.11
BALANCE at END:  1,290,370.22

Some observations about this post-crash recovery:
Most strikingly, the GR-Payout remains constant at 47443/year for 28 years, during which time the Balance grows from 744955 to 1095281, and the possible "Could3" value almost doubles from 54575 to 105971/year.
Absent any non-GR overrides, the retiree endures extensive underpayments, while his/her legacy grows to a usually unwanted, huge amount.

WHY Does This Happen under GR Rules? (1 - General)

The short story is that the structure of the Guyton GR rules creates a "wide moat"
requiring many, many years of natural growth of the balance, before crossing the threshold to the first 10% pay raise.

In the table below:
Column "YR_C" is the number of years Crashed to 100%.
Column "MOAT" is the amount by which the balance must GROW before first 10% raise.
Column "YR_2_R" is the number of years to the first 10% raise under GR rules.
Column "Could+12" is the annuity calculation of payouts sustainable for 12 extra years.
Column "Could+4" is the annuity calculation of payouts sustainable for 4 extra years.

Consider:

	
		Post_Crash
	YR_C	GR_Payout	Balance		MOAT		YR_2_R	Could+12	Could+4		   Notes
							
	0	55000.00	1000000.00	250000.00	12	72139.01	73757.69	   
	1	53398.06	945000.00	268592.27	14	68317.50	69963.92	   
	2	51842.78	891601.94	286643.06	16	64605.34	66278.04	   
	3	50332.79	839759.17	304167.88	18	60998.87	62696.10	   
	4	48866.79	789426.38	321182.48	23	57494.47	59214.19	   
	5	47443.48	740559.59	337701.32	28	54088.58	55828.39	   Note 1
	
	NOTES:
		1:  Initial reserves almost depleted. May need small CUT below par 55K to keep remaining payouts safe.

WHY Does This Happen under GR Rules? (2 - Assumptions)

Let's bring in a bit more detail about the exact rules being used in this simulation.
These will be slightly simplified from Guyton's description, while attempting to maintain the spirit.

We'll also clarify some assumptions built into the philosophy of my approach.
Then we'll apply it all to one case: the most extreme of the above series, 5 years "crashed to 100%".

Our first Guyton GR rule is: "No Cost of Living Increase in years where there is no income".
Now my overall presentation here assumes that the payouts shown are adjusted for inflation.
I also assume we mean no NET income above inflation, but that the "crashed" balance IS STILL staying even with inflation.
This may be at odds with your normal understanding of a "crash", but is indeed what I mean by a "crash to 100%".
Thus if the retiree gets a check for "55,000" in the first year, he's seeing that same "55,000" printed on the checks in the remaining years of the crash -- BUT -- the purchasing power of that "55,000" is reduced each year -- "deflated" by our assumed 3% inflation rate.

Guyton's Capital Preservation Rule:
If the payout rate comes to exceed the initial payout rate by 20%, we take a 10% payout cut.
Our initial payout rate is 5.5%. A 20% increase would take us to 6.6%.
Our payout rate after 5 crashed and deflated years is: 47443.48 / 740559.59 = 6.406%
We're CLOSE to, but NOT exceeding 6.6% here, so there is NO 10% pay cut taken.
NOT Shown, but of possible interest: a sixth crashed year would bring on the first 10% pay cut.

Guyton's Prosperity Rule:
If the payout rate falls BELOW the initial payout rate by 20%, we take a 10% payout RAISE.
Our initial payout rate is 5.5%. A 20% decrease would take us to 4.4%.

WHY Does This Happen under GR Rules? (3 - Worked Example of the Moat Calculation)

Here are the worked details for the 5 year crash: the bottom-most entry in the previous section.

We arrive at recovery with our payout rate CLOSE to but not exceeding the Capital Preservation Rule threshold.
This is where things get interesting.
The payout will not be adjusted until the now-restored 7% annual earnings, LESS the 47443.48 annual payouts,
cause the principal balance to grow to the point where the Prosperity Rule kicks in: where the payout rate drops to 4.4%

But where is that? It happens when the balance reaches 47443.48 / 0.044 = 1078260.91
So our "moat" is this target balance, minus the current balance of 740559.59, giving 337701.32
We've rounded this to 337701 in the discussion that follows.

I call this a "moat" because of the time and difficulty to cross it!
In the first year of restored earnings, they will average 7% of the 740559.59 balance.
That's 51839.17 but we're also paying out 47443.48, so our balance only grows by the difference: 4395.69

At this rate, it would take about 337701 / 4396 = 76.8 years to cross that moat to the first pay raise!

Thanks to compounding, the process does accelerate, and by the final year before crossing the moat, we've reached a balance of: 1067967.24 on the way to 1095281.46, a net gain of 27314.22
Had this final gain been in effect the whole time, we'd have crossed the moat in 337701 / 27314 = 12.4 years.

The actual moat crossing time after 5 crashed years was intermediate: 28 years until the first 10% raise.

FINALLY: Nimble Guard Rails with No Moat

It became clear during this work, that an updated annuity calculation based on current years remaining,
was the key to much more realistic and responsive resetting of the payouts in response to changing conditions.

I tried setting up a run-time competition between the original Guyton rules versus the annuity calculation, with the latter "stepping in" as an override IF and WHEN that made sense.

But it turned out that it ALWAYS made sense to go with the annuity calculation for the next payout reset.

Here is the proposed main GR rule:
On each day that nestegg shares are to be sold to provide payouts, perform this annuity calculation:

SOLVE for payout given the current values of these three parameters:
1 - Nestegg Balance
2 - Expected longterm average earnings, NET of inflation
3 - Remaining payout years of retirement, PLUS a buffer of 4 to 12 years.
Example: You've decided to use 6 years as your buffer, and there are 17 years remaining in the term.
Plug in as the number of Years: 17 plus 6 = 23.

I also want to propose a NEW guard rail rule as a defense against squandering of built-in reserves in the early years of retirement:

During the first 5 to 8 years: CAP the payout at the initial "par" amount (here, 55K/yr).
That is: allow payout cuts if needed, but postpone payout raises above "par" until after these first years.

Providing stability in the Spirit of Guyton:
It would be unexpected to get the exact same annuity result each time.
If greater payout stability is desired, impose your own 10% threshold before accepting a payout change.

CAUTION: Any deviation from the actual recommendation will incur a lost opportunity cost.
This is the inevitable trade-off for seeking more-stable payouts.
Empirically though, during this research, I settled on a threshold of 3%, which avoided trivial payout variations while remaining close to tracking the best possible outcome.

An important departure from Guyton is that, under rare, severe market conditions,
this algorithm can recommend a payout cut much more DRASTIC than 10% .
When this happens, it is important to accept the reality of the recommendation to avoid severe portfolio damage!

SUMMARY

We have provided a very simple and effective reality-check to the Guyton-Klinger "Guard Rails" Rules.
This becomes very significant in times of market stress!

USEFUL LINKS

For basic background about Crash-Points, please see: THIS HELP PAGE

For deeper background about Crash-Points, including derivation of the equations,
our specialized definition of a Crash, our important case of the "Crash to 100%",
and our DISCLAIMERS, please see: THIS ARTICLE

Links to Crash-Point (CP) "SOLVER" tools:
SOLVE_target_CP1_payout
SOLVE_target_CP2_reserve

For our entire body of material on this website, please see: ARTICLES AND CALCULATORS

FUTURE TOPICS for this Engineered Crash Protection series:

We hope to produce postings covering such matters as:
Have I reached Immunity against a Lost Decade yet?
Reserve Amount and Annuity Principal as Building Blocks of a Retirement Scenario

Nestegg Cycle © The Art of Engineered Crash Protection Part 2: Can we Do Better than Guyton's Guard Rails?