Author: mungofitch

Subject: Re: silver parachute

Date: 5/24/2016

Recommendations: 13

Since ETF`s structure are a little different than individual stocks, could you select the best fundamental screening filter(or combination) with long term potential from the following list ? As a starter to some backtesting of course. Ah, if only I knew. Screening for such things mechanically is hard, especially at the fund level. I might look at the long run average returns of different industries and pick some of the better ones as industry ETFs. And look at ETFs by average ROE...if there is an overlap among the top few, that would be good. Five year book/share growth rate is usually a pretty good predictor. For example, historically some good winning industries have been drugs, medical devices, hotels and gaming, cosmetics, beer, rails, pharmacy services, oil. Tobacco was best by a wide margin for a long time, if you have the stomach for it. Similarly, if using backtests to do it, I`d use backtests of rankings of industries by their average value of some metrics, then use that test to decide which industry ETFs to buy. Much longer backtest than anything ETF based. For individual stock screening some of my favourite criteria would be five year average ROE (available at the FT.com screener), and price to five year average EPS. I`d eliminate a number of industries, though, as their economics are unlike those of simple product and service forms for whom EPS works. As a general rule I`d skip banks, miners and commodities, utilities. Always good to steal ideas, too. Warren Buffett isn`t always right, but he`s right more often than I am. He says if he had to put all his money into one stock (other than Berkshire), it would be Wells Fargo. So I`d have a big chunk of them! One problem is that our backtesters are not well suited to testing long holds. What works at the short hold is price prediction; what works at the long hold is value prediction. I would *love* to spend a week with a backtester that could test five year holds and see what works best. But it would need some features to decide what to do with money in corporate events like mergers and buyouts, as holding cash or an acquirer till the end of the five year period would not be appropriate. In theory, best should be something like "highest return on on capital without undue leverage, highest return on incrementally invested capital, and selling at a reasonable multiple of cyclically adjusted earnings". Jim