The spleen is an organ that sits below left diaphragm in the abdomen. It’s generally thought of as an organ that filters bacteria and old blood cells. You can live without a spleen but you live better with one. Not only does it filter but it stores considerable blood, and it’s contractile. In the case of shock, a spleen through contraction can auto-transfuse it’s stored blood into the general circulation improving blood pressure and oxygen carrying capacity, just in the nick of time. Of course massive hemorrhage can overwhelm the spleens storage ability, but in terms of the so called golden hour where shock turns from a reversible phenomena to terminal that extra auto transfusion can make the difference.
I’ve written a good deal about my theory of a large portfolio risked at around 60/40 or 70/30 and a smaller portfolio risked as the tangent portfolio on the efficient frontier. I’ve written extensively about the risk and inefficiency involved in portfolios off the efficient frontier. The inefficiency is described using Monte Carlo analysis as an increasing portfolio failure rate and a failure rate that starts sooner into retirement. In my reading it appears poor SOR in the first 1/3 to 1/2 of retirement dominates the sequence and available funds in late retirement. The classic graph from the homepage of FIREcalc is:
This graph is of 3 identical 750K retirements, red starting in 1973, blue starting in 1974, and green starting in 1975.
This is GDP growth during 73-75 and the red line is average GDP from 1947-2009. GPD growth therefore were contracting by as much as -7.5% since normal is around +3.5%. This is the bad Juju of SOR. This chart describes 9 quarters of bad SOR Q2-73 to Q2-75, which is consistent with the red 1973 portfolio above, 6 quarters of bad SOR consistent with the 1974 portfolio’s outcome, and 2 quarters of bad SOR consistent with the 1975 portfolios outcome. A 30 year portfolio consists of 120 quarters so only a tiny fraction i.e. 9/120 quarters or 7.5% of a 30 year retirement was sufficient to run the ’73 portfolio out of money in just 20 years. ’74 survived 30 years but ended at half of it’s initial value, and ’75 thrived despite some bad SOR to start to close at more than double it’s starting value. To be fair the withdrawal load on these portfolios was 35K/yr or 4.6% so a smaller withdrawal could have lead to ’73’s survival as well. But what if you had a financial spleen? Could a reservoir of efficiently stored dough auto-transfused at the right time make a difference?
An alternative to a smaller WR would be to hold 3 years of WR in a differently risked portfolio, namely a efficient frontier tangent portfolio. So in this case for the 750K portfolio, 105K would be stashed in a 20/80 tangent portfolio and the remaining 645K left in a riskier say 60/40 or 70/30 portfolio that also rides on the efficient frontier. The idea would be in times that are good to remove money from the $645K portfolio and when times were bad to remove money from the $105K portfolio. To do this effectively you need to know when to pull the money from which account. A 35K withdrawal on a 645K portfolio constitutes a 5.4% WR so in years where you make greater than 5.4% take your money from the higher risk portfolio. In years when you make less take your money from the splenic portfolio.
In a 50% market downturn you would loose half your money if 100% in stocks. You would loose 33% if in a 60/40. You would loose 14% in a 20/80 since it’s mostly bonds.
The rule is to sell high, and in a crash the portfolio with the 14% drop is the high portfolio so sell from that portfolio. In normal times the 60/40 would be high and you would sell from that portfolio. The result of this is each portfolio has a chance to automatically donate at the proper time and recover in the proper time, according to it’s relative value. In the case of a crash the high risk portfolio is closed and re-balanced and no money extracted. Money is extracted from the low risk. In the case of good times money is extracted from high risk and low risk is allowed to grow unmolested. The expected rate of return on a 20/80 is 6.8% which isn’t chicken feed. The expected return on the 60/40 is 9.17% So a small percent of your money (14%) is growing at 6.8% and a large percent (86%) is growing at 9.17%, but in the down turn the portfolio with the larger risk suffers the worse insult and can least afford liquidation and is protected from the ravages of bad SOR. The protection isn’t perfect and given a long enough drain in bad times the low risk fund will run out of money, but it’s in bad times early in retirement that you you want the protection. It’s less necessary as time goes on. If you don’t use the smaller portfolio once it begins to accumulate interest over inflation, you can start to dollar cost averaging the excess from the low risk to the higher risk. If low risk runs out of money early do not refill it, as it served it’s purpose to protect the high risk portfolio. As retirement proceeds you begin to move out of the period of portfolio failure due to bad SOR, say year 10 – 15 in a 30 year retirement. As long as the low risk portfolio insurance exists you can start to glide up your total asset allocation by say 1%/yr as you re-balance every year with the added excess and dollar cost average from the low risk fund.
This technique solves the issue with the usual bucket theory problems where the cash bucket is constantly draining money from the riskier buckets. Instead you are using a buy low sell high strategy. It also solves the problem of a bucket of cash which returns virtually nothing. The low risk portfolio is in fact risked most efficiently and though returns aren’t stellar they are quite consistent because their risk is low, adding their own value to the portfolio in terms of stability. The other thing I like about this is it mechanically forces you to buy low and sell high. In the case of a crash by not selling low in the riskier portfolio you are effectively buying low relatively speaking. You sell high and buy low when you re-balance from the bond money you have been stashing from selling high on the way up. That excess money sold high is then reconverted to cheap stocks when they are at the best value. If the low risk portfolio is transfusing it’s excess gains into the higher risk portfolio it is another form of buying low. This portfolio is cutting it to the bone with a 4.6 WR. The S&P 500 has only returned 5.5% on the average over the past 19 years. I would feel much better about a 4% WR or a little less but none the less this takes advantage of a kind of risk shifting over time to protect at least to some extent against SORR. Add a little SS to the mix and you’re golden. I don’t think I would get hyper anal about precisely a 5.4% return as the trigger but in a year like last year where the market was down 5% that a net of 10% below where you need to be and I’d think about pulling the trigger. I haven’t done an in depth analysis especially on when to pull the trigger so I won’t recommend this but I wanted to get it in the archive while I had it in mind.
6 Replies to “Splenic Portfolio Theory”
Interesting concept, thanks for planting a seed. I like the idea of auto transfusion. Not sure if it meets the requirement for simplicity enough that my wife could easily carry it out in my absence (she’s plenty bright, just disinterested – common ailment).
I’m still considering the details. This is a proof of concept piece that needs some refinement. It actually is quite simple and mechanical which is it’s beauty, but it exists in an environment of government regulation and tax law that makes things complicated. I’ll work on a piece that gives specific scenarios such as when CD kicks the bucket what does the disinterested ol lady do?
This notion of financial “simplicity” and DIY is a fallacy in the FIRE concept. It is what pisses me off most about boggleheads and peeps like MMM and WCI. Because of the legal environment, retirement is anything BUT simple, but it is predictable and a reasonable algorithm can be created. If you DIY you simply can not afford to be disinterested. It’s like driving your car at 100MPH while screwing around on your phone. It will work fine just up to the moment it doesn’t. This is exactly why I pay an adviser who completely understands my plan and is able to step in and take over execution. When you are talking a multi-million dollar future and it’s proper disposition you can’t afford screw ups in execution and the little dab of money that security costs is trivial compared to the headache and potential loss caused by your cheapness. The government has zero interest in your welfare and it has even less interest in her welfare. Morally you can not distill this down to 10 bullet points while pretending to be the cleverest guy in the Dr’s lounge. The stakes are far too high. You get what you pay for. Pay nothing, you got some joker’s bullet points to hang your future on. Some joker who’s main interest is in furthering his blog’s income.
Did you come up with this cool nomenclature for the 2nd portfolio? I do like the auto transfusion concept, a financial transfusion. And is this pretty much a similar concept to the lifeboat strategy that you mentioned in a guest post on my site or is there something different?
It’s a more refined version of the portfolio insurance strategy on your site, but still some refining to do>
I like the auto-transfusion idea. My pre-Gasem retirement planning had an aspect of this concept as a Plan B if Plan A turned out to be less robust than expected. My pre-tax account was planned for post 70 living and is likely overfunded based on RMD calculator projections. My taxable account is roughly the same size as pre-tax but has 7 years equivalent worth of cash/bonds at our typical spending level which I plan to start removing and using once I retire, i.e. have spent my personal capital. Roth accounts are only 15% of NW currently but trying to boost as much as possible through conversions from pre-tax plus BD Roth IRA and Roth 401k from current work. I’ve always considered Roth to be the account of last resort for use if things go wrong with my plan. But my taxable should also have extra funds available to handle unallocated and unforeseen expenses during retirement. I like knowing I have a spare tire in case the original springs a leak and goes flat. You’ve convinced me that minimizing risk is the preferred goal in deaccumulation rather than maximizing portfolio performance. As I boost my bond holdings, it should preferentially occur in the pre-tax account to slow the growth there and let the Roth be the vehicle to maximize performance. Splitting the portfolio into different risk strata and pulling from the most efficient part is an interesting idea. To make this work I think you have to get the pre-tax RMD low enough to where it doesn’t force you to take too much from the wrong pool. If I am able to convert enough pre-tax to Roth by the time I’m 70 this type of plan should provide an extra layer of protection and allay the fear of dying broke.
I’m not so much an advocate of minimizing risk as diversifying risk and then asking what economic conditions will force failure, then asking is there a way to mitigate that failure without screwing up some other risk mitigation strategy. So right now I’m Roth converting to mitigate my tax risk in the coming decades. My feeling is the government owns part of my money by my choice of investing in tax deferred money and given the lack of responsibility of the political class it’s better to get the pain over with now. I’m investing in Roth because it’s a ready receptacle for post IRA money and will provide over time an increasing resource to finance the inevitable known unknown leading to my death and later my wife’s death. It’s an imperfect plan but far better than nothing or paying some insurance company for some crappy annuity. A 2 tier low/high risk portfolio gives you the choice on from where to take your WR money, and some simple trading rules keep you buying low and selling high, which is the bedrock strategy of Capitalism. I have another risk ploy for bad times in that I own some gold stocks and gold mine stocks which I bought cheap and in the down times gold and gold mines tend to soar so this gives me another resource to sell high in the bad times and use that money to live on, or use the proceeds to buy stocks low, so if I buy gold for 1000 and it goes to 2000 in a crash I’ll sell 1000 and plow that 1000 into beaten down stocks ready for the rebound. The problem is people have been trained that you “can’t time the market” and then use the example of going from 100% stocks to 100% cash, and it’s true you can’t effectively time 100% to 100% but you can effectively time going from 20/80 to 70/30 based on a trading rule which takes you and your paranoia out of the equation. It’s also true you only have to protect what’s most at risk when it’s mostly at risk. So in a downturn a high percent equity exposure places you most at risk because of the volatility so a low risk alternative that also produces income and can be sacrificed instead of sacrificing the high risk equity which is the economic motor and in the good times a good motor is just what you need. Of the last 30 years only 6 years were under water. If you divide the 30 in half (the first half is most important to SOR and portfolio survival), of the first 15 years only 4 were under water. 3 of the 4 were clumped together, the dot com bust, so having some little pot of relatively risk free money available to cover a dot com situation in the first 15 years is just what you need. All of those choices mitigate risk but also to some small extent mitigate return as well. Once you get past the tangent portfolio every bit more return you try to achieve carries a a risk that costs more than the return is worth. So the tangent portfolio for total stocks/total bonds is 20/80 and predicts 6.8% return and 4.28% risk. Going to 60/40 predicts 9.12/9.33 so an extra 25% return costs you 55% more risk. Going to a BH3 predicts a 24% improvement in return vs 65% more risk since the BH3 is not efficient. In good times the extra returns pays off but in the bad times the extra risk eats your lunch. In 2008 it took the S&P 500 5 years (2013) to recover to it’s 2007 high. It took a 60/40 about 3 years (2011) since the 60/40 only crashed 33% instead of 50%. 5 years is 17% of a 30 year portfolio, a long time to be pulling money from an injured account. Since the problem is a sequential problem having different assets that behave differently at different periods in the sequence gives you something which is high (relatively speaking) and the rule is sell what is high. Can the system be overcome? Of course. Given a bad enough sequence everything can be overcome, but then poor return also can damage your prospects, so you can’t guarantee anything but you can set yourself up for a higher probability of success by lowering your probability of failure, and that’s the name of the retirement game, don’t die broke while living a nice life in the mean time.