(Resurrecting a long dormant thread....)
One of the objections that longinvest, siamond, and others had against Prime Harvesting was the "memory effect". That is, the way Prime Harvesting works is dependent on the exact value of your portfolio on the day you retire. If you retire a few days earlier or later then the strategy can behave very differently.
While I don't think it is totally necessary to not have memory effects, I still think it is a fair point and introduces a bit of wonkiness. Something I'd been thinking about for a long while was whether one could use actuarial calculations (like PMT, PV, etc) in order to drive the "harvesting" instead of just looking at inflation-adjusted stock prices.
This is far from definitive but I figured I'd share what I have anyway, in case anyone else in this thread was interested.
Because this is how the universe works...at virtually the same time I did this, Wade Pfau published something very similar.
The basic idea is: we need to decide when we "harvest". That is, when to sell some stocks and buy bonds with the proceeds. McClung says "whenever your stocks are 25% above their inflation-adjusted starting value".
Let's try another way of looking at things: we know how much we're withdrawing today. We know (approximately) how long we will live. We can use a Present Value calculation to decide how big our portfolio "should" be in order to sustain that withdrawal.
We have two things we need to decide:
1. What "rate" should we use in the PV calculation? If the number is too low then we will never harvest. If the rate is too high then we'll harvest too much, end up mostly in bonds, and be reliant on unrealistic growth from stocks to support our life style. I'll use 5.4%, which is the global average of equities from 1900-present according to the Credit Suisse Global Returns Yearbook.
2. How much should you harvest? Sell 20% of stocks? 25%? Whatever amount is above the PV calculation threshold? For no good reason (other than it is what McClung also chose) I'll go with 25% of stocks.
Here it is against US data from 1871-present (it is labelled "ActuarialHarvesting" in all of the charts below)
And against UK data from 1900-present
And against Japanese data from 1957-present
So at first blush, this looks like a reasonable re-definition of Prime Harvesting to get rid of the memory effect.
For those interested in more detail I wrote two blogs:
https://medium.com/@justusjp/actuarial- ... 17d3a044ce
https://medium.com/@justusjp/actuarial- ... b48c115640
And here is Pfau's series of articles on a similar topic
Part 1: https://www.advisorperspectives.com/art ... ent-income
Part 2: https://www.advisorperspectives.com/art ... ond-ladder
Part 3: https://www.advisorperspectives.com/art ... r-strategy
Also, bobcat2 has written a few posts here on bogleheads about the "funded ratio" which contains some vaguely related ideas.