I’ve had the chance to take a close look at the auto-trading done in our OP II model on Collective2.com Auto-trades seem to working really well with little or no change between model trade prices and auto-trade prices. Hence, we’re reopening OP II to new subscribers and will continue to track the continuity in prices periodically. Click on the link below to view the model and performance.
I’d like to thank the members of Collective2.com for their faith and trust in us as all of the available slots in Optimized Partners II model have been taken!
But the investment model used in the Optimized Partners II system is still available to clients of Optimized Partners, we must limit its availability to Collective2.com members.
We do have one spot available for our 30-stock model OP III
A trading update will be available soon.
Recently we’ve been researching and testing a mid to large cap stock ranking system that has been in the public domain for several years and known as the “Magic Formula”. The Magic Formula is based on Joel Greenblatt’s book “The Little Book That Beats The Market” and he has followed up his book with a free website which features the Magic Formula. Joel Greenblatt is the founder and manager of Gotham Partners.
The Magic Formula book was a bestseller in 2005 and its written in plain english with simple math computations, its very easy to understand for most people. It is described by Joel Greenblatt as “a long term strategy designed to help investors buy a group of above average companies but only when they’re available at below average prices.”
This may be a spoiler alert but the Magic Formula isn’t really “magic” at all, but in our opinion a moderately successful quantitative investing model.
There are two key factors in the Magic Formula:
1. Determine the company’s earnings yield = EBIT /Enterprise Value
2. Determine the company’s return on capital = EBIT / (net fixed assets + working capital)
Exclude utility and financial stocks. Buy 20 to 30 companies and hold for one year.
Our interest in Greenblatt’s formula is that its effective on mid to large size companies which is the most difficult sector to generate consistent Alpha (for example, the rate of return above the S&P 500). Since the MF formula is suited to mid and large caps its able to handle large portfolios easily and could be a solution for clients whose accounts grow to size too large for small cap models. For example, in this situation we could design a portfolio where 70% could be allocated to our small cap models and the remaining 30% devoted to the MF design. Thus, allowing the client to maintain a large small cap presence.
By our estimation, we followed the guidelines set forth on the MF website and constructed a 11-year simulation using the MF ranking system to create a portfolio of 20 stocks and held them for 12 months.
Test Period: 1/1/1999 to 12/26/09
MF Total Return: 175% (includes commissions and slippage)
Benchmark Return: -9.28%
MF Annualized Return 9.6%
MF Sharpe Ratio .2
MF Max drawdown 53%
Composition allocation: Large Cap 65% and 25% for Mid Cap and 9% Micro caps.
Opinion: Well, its a start but our calculation doesn’t come close to Greenblatt’s 18.57% stated rate of return. The risk of buying and holding for one year is too much for us, not even factoring in coping with two substantial bear markets in the time period. My guess is his returns do not include commissions or accounting for slippage or even management expenses which is not realistic.
Based on our experience, regardless of Greenblatt’s good intentions of trying to simplify the quantitative investment process, perhaps only a very small minority of investors would endure the past 11 years of his study.
Plus, due to the large stock allocation social screening would significantly impair returns.
But it does hold promise since even with this unoptimized test the MF formula is able to generate real Alpha during a brutal period when the S&P 500 return was negative, which leads me to believe that with refinement (as shown below) there is promise.
Optimized Partners revision of the “Magic Formula” 1999 to 2009
We have made some changes that just seem like the right thing to do.
1. Rather than an annual rebalance we employ a weekly rebalance.
2. We have added a 25% trailing stop loss to reduce downside risk and keep individual stock losses to a modest level.
3. We’ve added a macro market exposure mechanism devised by Bob Dieli of nospinforecast.com known as the Aggregate Spread. We do use this with some subjective oversight.
4. On the Buy side of the Magic Formula to further filter down companies worth buying we’ve added three accounting factors and a sector weight maximum of 33%. The accounting factors are related to profit margins, sales changes per quarter and the 5-year compounded growth of sales.
5. On the Sale side we noticed that the Magic Formula has no triggers for when to sell stocks other than to wait a year. Maybe we’re control freaks but that seems counter intuitive to us since there is a great chance many of the stocks purchased will lag for long periods of time and impair annual returns. So we added a factor of selling stocks after 45 days of holding if they don’t at least provide a return greater than the benchmark S&P 500.
Test Period: 1/2/1999 to 12/26/2009
OP/MF Total Return: 351%*
Benchmark return: -9.28%
OP/MF Annual Return: 14.48%*
OP/MF Sharpe ratio: .69
OP/MF Max drawdown: 29%
*Assumes 0% interest or rate of return when not invested in stocks. If an investor had swapped stocks for Treasury bonds during the out of market periods the total and annual rates of return would likely be significantly higher.
Opinion: We’re closer to the authors return figures but still not there. And by now you may be wondering why this blog post is of any significant importance. The 1999 to 2009 was really a brutal period that saw not just one bear market but two. Value based small cap models were able to dodge much of the 2000 to 2003 losses because by year 2000 there was a massive bubble not just in tech stocks but also in large cap growth stocks, a period we’re unlikely to see again in our lifetimes.
Optimized Partners revision of Magic Formula trailing 10 years
My interest was really piqued when I looked back at the trailing 10 year period dating to February 2004, with having to deal with the large cap tech bubble our rendition of the Magic Formula will raise some eyebrows. We must confess the changes we made to our version of the MF were made with the last ten years in mind, not the 1999 to 2009 period.
Test Period: 1/2/1999 to 12/26/2009
OP/MF Total Return: 1029%*
Benchmark Return 60.93%
OP/MF Annual Return: 27.45%*
OP/MF Sharpe ratio: 1.3
OP/MF Max drawdown: 18.8%
*Returns do not include management expenses but do include commissions and accounting for slippage. Returns do not include interest or capital gains that could be earned by shifting to Treasury bonds when out of the stock market.
Opinion: This is why we’re excited about the Magic Formula and the potential it possesses. Very few ranking systems can generate any significant Alpha in large and mid caps but employing our revisions to Greenblatt’s pre-existing model holds a great deal of potential.
This appears to be a keeper.
Brad Pappas
The theme we’ve been recommending for our client portfolios for 2014 has been buying closed end bond funds over stocks. So far so good.
We believed the risk/reward for US equities was very unfavorable for the first 9 months of the year and we began to seriously lighten up on our stock holdings in December.
As of Friday, we now read that Goldman Sachs has downgraded US stocks and sees the potential for a 10% pullback in the next three months.
Goldman’s perspective is that the stock market is richly valued “by almost every metric” but there is no mention of the extreme positive sentiment for stocks and that the mid January to mid February period tends be negative for stocks.
We agree with GS that there will be a better entry point for stocks later in the year as the longer term view of stocks remains favorable.
Brad
P.S. In our last blog post we mentioned that we thought we were wrong in our bullish call on US long term Treasuries based on the ADP data. We had sold the TMF and swapped into the bearish TBT. Well, it appears we may be right all along and this morning we took a small loss on the TBT position and went back into the TMF.
Long TMF
I relent on my bullish Treasury bond call which was based on the extreme negative sentiment readings. We owned a modest position in the TMF which was a positive bet on the long term US Treasury bond but after today’s ADP private non farm payroll numbers the past of least resistance is likely down for long term Treasuries, despite the extreme negative sentiment.
In place of the TMF we will be swapping into the TBT which will rise in price should long term Treasury prices fall. The TBT will also serve as a hedge to our relatively large closed end municipal bond positions. Since a rise in interest rates could hurt muni prices and hopefully some of that loss could be made up in the TBT.
Long TMF and TBT
A tip of the hat to @mktoutperform
Tuesday the 17th is the start of December’s seasonal strength
