TAA Model for December, 2010
This is a new monthly feature at the MarketSci Blog.
Our Tactical Asset Allocation (TAA) model selects up to four assets from a diversified basket of asset classes on the final trading day of each month. Below is our allocation for today’s close. Click to read more about the TAA model.
I eat my own cooking, so I’ve devoted a healthy share of my own net worth to the TAA model (read why). On the last trading day of each month I’ll be sharing my new allocation (see above) and real-time performance (see below), excluding trading frictions and taxes.
The model performed in line with the benchmark in November, returning -0.27% versus -0.25% as of 11/29. There will be no changes to the portfolio for December.
As the performance graph below shows, the model has (so far) been a near mirror image of the benchmark. That’s because the portfolio has looked so similar to the benchmark, with 36% invested in Treasuries and 39% in stocks/real estate.
That will not be the case over the long-term and in the coming months I would expect the model and the benchmark to, for better or worse, deviate significantly.
[linearly-scaled, growth of $10,000]
Happy Trading,
ms
. . . . .
To stay up to date with what’s happening at the MarketSci Blog, we recommend subscribing to our RSS Feed or Email Feed.
Filed under: Tactical Asset Allocation | 7 Comments
Hi Michael
The new TAA allocation avoids leverage? What is the theory/calculation you are using as to when to use leverage, can you discuss your thoughts on that important subject. As you know, Faber does not really say much about it except to say it helps. So once you decide the SMA test and relative strength test to pick your classes and relative positions, the question of whether to go over 100% with leverage or not needs some formulaic process, otherwise self-sabotage can occur if left as a subjective item.
Also how would the cost of leverage (broker rate) get factored into this, e.g. would the expected returns have to be over some hurdle to consider leverage.
Hello Kamal – a couple of months ago when I was still fleshing out the model I had seriously considered running a second leveraged variation of the model. However after further testing I just don’t think it will make sense over the long run given historical margin rates (and the subsequent impact on risk-adjusted returns).
It might make sense at this moment (given how low those rates are), but my intention (for myself as much as for anyone else) is to create a “generational strategy” that could run, unmodified, for a very long time.
I think there’s a place for leverage (ex. we use it often in YK), but not here.
michael
Hi Michael,
I forget if you covered this in the past posts, but I assume you looked at varying time periods — one month, one-week, etc. — to find the best combination of performance and ease. I was wondering if you looked at how the system fared if you rebalanced twice a month — once in the middle and once in the end? If so, did it alter performance in an appreciable way? With only four positions max at a time and ways for people to trade it cost-efficiently, I wonder if this would be a benefit.
Thanks for all you do on this site to help us little guy investors/traders out.
Jeff
Hello Jeff – good question.
It’s difficult to do a solid test using weekly/daily data because long-term historical data for some asset classes (ex. real estate) isn’t available. Based on the data we do have, I (to my surprise) don’t think trading in shorter time frames helps performance (especially when accounting for transaction costs and the hassle factor).
Here’s a tangentially-related blog post that discusses:
http://marketsci.wordpress.com/2010/07/13/the-golden-cross-daily-vs-weekly-vs-monthly-2/
michael
Hi Michael,
Given your response to Jeffry’s question, this may be difficult to test historically, but based on your monthly seasonality studies, have you tried to run the TAA model buying positions 3 to 4 days before the end of the month? I realize the TAA is in the market throughout the entire month and doesn’t sell during the “unseasonal” times but thought it may have some merrit to move the asset classes with the greatest momentum right before a seasonal up-move.
Also, if I understand the TAA model correctly, you tie the volitility of the portfolio to the S&P 500. So I assume that when the S&P is experiencing increased volitility, the portfolio would do the same. If that is correct, what are your thoughts about doing the exact opposite? Increase portfolio volitility during low S&P volitility and reduce it during periods of high S&P volitility? It seems like this may help allocate risk a the most appropriate times given market conditions.
All the best,
Brad
Hello Brad – two great questions.
First, based on the testing we can do (given the limitations on data), I don’t think there’s enough of a performance difference to justify trading on any particular day, so month-end just keeps things simple. If we were talking about pure trend-following that was moving in and out of an asset and cash I think the day of the month would matter, but because of the nature of TAA, I don’t see it as much of an issue.
On the volatility question, I’m still mulling the best solution (the model is still a work-in-progress). Part of me says that something like what I did in this post might make sense:
http://marketsci.wordpress.com/2010/03/05/volatility-adjusted-buy-hold/
This is similar I think to what you’re suggesting. More to follow.
michael