Prime Harvesting vs Rebalancing (graphs!)
The recent book Living Off Your Money by Michael McClung is excellent, if you’re a retirement research junkie. He provides concrete recommendations on how to invest in retirement, how much money to withdraw each year, and how to withdraw it. He provides lots of analysis, charts, and graphs.
But no book is ever going to be comprehensive, especially with so many variables. Which provides an opportunity for me to dig into some details and see how it holds up.
One of McClung’s recommendations is something called “Prime Harvesting”. Every year that you are retired you need to take some money out of your portfolio. How you take it out is called an “income harvesting strategy”. Imagine you decided you need to take out $40,000 this year. Do you sell $40,000 of bonds? $40,000 of stocks? Whichever one did best last year? Whichever one did worst? Some other combination?
(Determining how much you take out is a different strategy: your withdrawal strategy.)
The “default” strategy most people would use is Annual Rebalancing. At some point in the past you decided you would have a portfolio that was 60% stocks and 40% bonds. Every year you rebalance your portfolio to maintain that ratio. When you take a withdrawal, you also maintain that ratio. This means when you need $40,000 you would sell $24,000 of stocks and $16,000 of bonds.
Prime Harvesting is an alternative approach that the author claims is more efficient; you end up with more money in your pocket under most scenarios.
Prime Harvesting works by selling bonds first. If your stocks have appreciated by 20% (inflation-adjusted), then you convert some stocks to bonds. This has caused some people to ask, “Doesn’t this whole thing fall apart if we have a Zero Interest Rate Policy (ZIRP)? Wouldn’t you end up just selling all your bonds and end up with 100% stocks?”
The answer is yes, but it also doesn’t look like there are any problems with that. Let’s look at a bunch of charts!
You don’t have to wait for some Possible ZIRP Future to see a scenario like that play out. Let’s turn back the clock…A 1910-retiree who has a 60/40 portfolio would see their stock percentage fluctuate like this over a 40-year retirement:
You can see that we gradually sell off our bonds and eventually reach 100% stocks. The market improves and we finally start selling off some stocks for bonds. Then towards the end of retirement, we head back to 100% stocks.
Having 100% stocks for a retiree is not the normal recommendation…but are there any actual adverse effects compared to Annual Rebalancing? Let’s look at the annual inflation-adjusted income for Prime Harvesting and Rebalancing:
Based on income, it is hard to see how being 100% stocks hurt this 1910-retiree. Actually, we can see that Prime Harvesting resulted in substantially more income starting around Year 16.
This was just a snapshot of one year, 1910. What does the overall record look like, when we compare Prime Harvesting to Annual Rebalancing? Let’s compare “inflation-adjusted lifetime income” generated by the two strategies for every retirement year from 1871 to 2015, assuming a 40-year retirement:
When looking at “inflation-adjusted lifetime income”, it looks like there is almost no (historical) scenario where Annual Rebalancing did better than Prime Harvesting. But there are a few…
Let’s zoom in on one of those years where Annual Rebalancing did substantially better than Prime Harvesting. The eight best years for Annual Rebalancing were 1949, 1950, 1948, 1954, 1953, 1951, 1995, and 1885. The first five of those are all clustered around the same time period 1949–1954.
Let’s look at some of those in a bit more detail, tackling them in reverse order….
There’s a repeating pattern here. Annual Rebalancing has won overall in these years but when you look at the details the win is less impressive. It is usually a case of either:
- Annual Rebalancing returning $2,000 to $3,000 extra a year for a number of years (e.g. Chart 5), or,
- Annual Rebalancing not reacting to drops from market highs as quickly as Prime Harvesting does. Eventually the two strategies end up with similar incomes but there can be a largish gap for a few years. In all of these cases, however, Prime Harvesting is still returning more than the retiree started out with.
Overall, if those are the worst case scenarios for Prime Harvesting relative to, then I’m comfortable taking that slight risk with Prime Harvesting.
Let’s compare the worst case and the best case for Prime Harvesting. We saw the worst case (for a 50/50 portfolio) already but here it is again:
And now the best case against a 60/40 portfolio.
The risk is small and the upside is quite large—up to 60% more income per year for the last 17 years of retirement.
All of that is based on historical market information. But some people are worried that we’re entering an era of unprecedented low market returns. What happens if we are in a NIRP (Negative Interest Rate Policy) world for the next 20 years? Does Annual Rebalancing become better then?
Let’s look at what happens if we have 30 years of 2% real returns for stocks and -2% real returns for bonds:
You’re kinda screwed in both worlds but Annual Rebalancing loses out to Prime Harvesting at every point on the spectrum.
The real answer to a low-returns world like that is your withdrawal strategy but it is good to see that Prime Harvesting doesn’t exacerbate the situation.
Let’s try to really stack the deck in favor of Annual Rebalancing:
- 10 years of ZIRP returns. This forces Prime Harvesting to sell all of the bonds.
- Then, a big (-25%) drop in stocks. Annual Rebalancing gets to “buy stocks cheap” and Prime Harvesting is out of luck, since it is already at 100% stocks.
- Then, normalish returns onward (8% stocks, 4% bonds, 3% inflation)
Even here….Annual Rebalancing seems to trail.
It is hard to find any scenario where Annual Rebalancing is a substantial improvement over Prime Harvesting.
The one area where Annual Rebalancing does consistently win is “percentage of bonds owned”. Prime Harvesting doesn’t really care how many bonds you own. It is happy to sell them all off and leave you as an 75-year old retiree in 100% stocks. Whereas Annual Rebalancing will diligently rebalance you back to a 60/40 ratio.
Next time I’ll look at bond ratios and overall portfolio values for the two strategies to see whether Annual Rebalancing might have a higher “sleep well at night” factor, despite lower annual income.
Notes: Market data is based on simba’s backtesting spreadsheet from bogleheads.org (which is in turn based on Schiller’s data). The portfolio uses the S&P 500 and intermediate-term government bonds for the two parts of the portfolio. Income harvesting just tells you HOW to make your withdrawal, it doesn’t tell you how much to take out. The “EM” variable withdrawal strategy from the book was used in all examples above.