As expected, after a year for the financial history books, my previously posted Black Tar Portfolio, the asset allocation scheme that both never lost money and did so with the highest annualized return needs to be revised going forward to meet the same criteria after incorporating 2008’s returns into the historical simulation.
First let’s review the previous Black Tar portfolio, tested over 1972-2007 (thanks to assetplay.net‘s wonderful, free historical database).
63% US Short Term Treasuries
17% Emerging Markets
08% International Value
For the 36 years tested, the worst yearly return was 0% and the average annualized return was 11.43%. But 2008 changed all that, as the portfolio returned -13.31%. I think most would agree though that this is still quite respectable when compared to the performance of other popular lazy portfolios.
Shortly after the New Year began, I incorporated the 2008 returns and let the computer began the search for the new Black Tar Portfolio. For 1972-2008 it looks like the portfolio with the highest annualized return that has never had a down year is:
36% US Long Term Treasuries (such as VUSTX)
36% US Short Term Treasuries (such as VFISX)
12% International Value (such as VTRIX)
09% Pacific (such as VPACX)
04% Commodities (such as PCRDX)
03% US Small Cap Value (such as VISVX)
This particular portfolio had the following characteristics over 1972-2008:
|Best Return (1986)||30.33%|
|Worst Return (2008)||0%|
|Average (Arithmetic) Return||9.87%|
|Annualized (Geometric) Return||9.63%|
|Sample Standard Deviation||7.42%|
Observations about the new portfolio? Well, as predicted earlier, the more years one adds to the historical backtest, the more likely that the constraint of never losing money pushes you deeper into treasuries. That is the case with the new portfolio, as it has moved from 63% to 72% treasuries.
And I found it interesting that the treasury allocations came out to be an equal 50/50 split between short and long terms, especially since the backtest engine shows that 100% in intermediate treasuries instead has a higher compound annual growth rate and lower standard deviation. Obviously there’s some sort of correlation mojo going on with the other asset classes that you miss if you only examine the treasuries in isolation…
Also a natural question to ask when you see 9% more of the allocation going to treasuries is, what asset class was that 9% taken from? You could say that it came almost entirely from commodities, as it has gone from a weight of 12% down to 4%. By comparison, only 1% was removed from stocks, taking the stock allocation from 25% down to 24%.
The composition of that 24% stocks is a bit different than before. International value plays a larger role now, moving from 8% to 12%. Emerging markets has disappeared entirely (sniff, sniff), and we now see the appearance of a 9% weight in MSCI Pacific Index (mostly Japan), and 3% in US Small Cap Value (perhaps a distant relative of emerging markets).
Time-permitting, I’m hoping to next examine how the Black Tar Portfolio morphs if I change the criteria of never losing money into something like never losing money 95% of the time. Or I could also examine performance on a different time scale, like requiring there to never be a net loss over 2 years instead of 1.
Oh, who am I kidding? Something in my DNA needs and craves a historically proven asset allocation scheme that has a large percentage riding on emerging markets (e.g. Black Tar 2008). I’ll probably keep simulating until I find one again…
UPDATE: the Black Tar 2009 portfolio returned…
4.7% in 2009
8.0% in 2010
7.9% in 2011
6.1% in 2012