They attempt to find the most ‘clonable’ and best hedge fund managers. And then to copy their holdings. So, this is a portfolio of stocks held by people paid the most in the industry to pick stocks, based on low turnover and other elements that make what they do ‘copyable.’ Backtesting shows this works and produces consistent alpha.
But…Fifty stocks chosen by the ‘no rank’ method in the screener from the universe of the Russell3000, equal weighted with div’s, also rebalance semi-annually - is up just over 66% in that time.
So…in the past 18 months, we’d have been better off just buying equal weighted stocks than copying the holdings of the highest paid, best, lowest turnover and most ‘copyable’ hedge fund managers in the world.
I’m open to other ‘stats’ or fun tidbits if people think about this question and have fun or interesting ways of looking at it.
Slightly better results, but still only basically tied an equal weighted (and likely lost after taxes). So…Is there alpha here? How long will it take to know?
Of course there’s alpha; we can see that with an 18 month backtest of a simple ETF screen: Ticker(“ALFA”). The annualized alpha was 9.47%. GURU was a lot better over that time frame: 18.33%.
FWIW, a screen of all US equity ETFS excluding leveraged, short, and leveraged short, and excluding SPY (the benchmark), delivered 18 month alpha of 9.9%
Tom, not to sound obvious but aren’t we always a quarter late with these hedge funds replica ETFs?
Btw there’s a good website for tracking top managers picks (gurufocus.com). Their format makes it very convenient to access a company’s fundamental (more so than the P123 panels which are a little cumbersome IMO). What’s interesting is in each company summary page, you can see the aggregated amount of buyers or sellers among those managers quarter by quarter. It’s a (small) reassurance when you buy a company and a couple months later you find many managers were doing the same thing at the time. Check it out, it’s worth the look.
I actually subscribed and tested gurufocus.com. It has an extremely effective useful front end that gives excellent relative valuation metrics.
Then i decided to select the top guru at end of year 2013 and see how his stock picks are performing. I did this because my concern was when i am buying the guru is selling. Guru performance was below spy. Sorry i dont have spreadsheet but i got out of trial period.
Finding alpha in this market IMO will be more difficult the next three years. And i am concerned once taper reality sets in more the market may get ahead of early 2015 and present some challenges second half 2014. So i am considering a core two prong strategy: 1) Focusing on US small cap relative strength + strong balance sheets 2) Dividend aristocrats applying a solid port123 ranking system with sector sensitivity. Aristocrats are stocks that are multi-national and highly exposed to the growth of the global consumer.
The start of this thesis is take advantage of US growth over next several years with small cap. Capture yield, alpha and lower risk with large cap aristocrats.
open to critique and any feedback
joe
The clear point on ‘alpha’ is that it’s benchmark dependent. I don’t think a market cap weighted ETF is the proper benchmark for any ‘active alpha’ portfolio strategy. And neither ETF beat a ‘no rank’ selection of the same number of stocks (rebalanced at the same period) from the Russell 3000 over the period. Thus, the only ‘alpha’ they had was potentially due to equal weighting. They had (or zero) negative alpha vs. the equal weighted bench.
This points out that letting users create ‘custom benchmarks’ for sim’s / ports for the alpha and performance stats calculations would be valuable for port. monitoring. Not a hi-priority I guess, but that’s the core issue.
Tom, as per the da Vinci quote you put into your profile signature, “Simplicity is the ultimate sophistication.”
If a standard benchmark doesn’t work for you, then create your own whether it’s a no-rank selection of a certain number of R3000 stocks or whatever.
Sure it would be nice to be able to do it with a couple of mouse clicks on p123, but no matter which of the ton of requests we satisfy next, there will always be many members complaining about all the thing we didn’t do while we worked on custom benchmarks. Custom benchmark creation is definitely one of the many things on my preferred to-do list. But I’m not going to just sit around waiting for it. There’s a reason why God (or was it Gates, or, rather, the guy who invented VisiCalc) created Excel. So when I really need to do any work with a custom benchmark, I just roll up my sleeves and do it (as I do right now for a nano-cap sub $10 priced stock index I created for a model i have for which I don’t think the R2000 is a suitable benchmark).