Saturday, July 9, 2011

MC02 Value screen

I have been having discussions with Richard Beddard about value-based portfolios, whether they worked, and under what circumstances. My view is that deeper value shares only provide truly superior performance, as a class, coming out of bear markets. I don't think this is where we find the markets today, so I consider mechanical value-based strategies to be risky propositions at this point in time. That's not going to stop me in what I am about to do, though.

I also take inspiration from Joel Greenblatt and his Magical Formula. I am a little skeptical of his formula, though, as people have reported results that are less than the returns that he claims to have. It is also too difficult to replicate his precise calculations, not least because he hasn't laid them out in precise detail.

Another source that I am taking inspiration from is Stingy Investor, who is doing exceptionally well with a Ben Graham formula. The portfolio has thrashed, and I do mean thrashed, the S&P500 over a decade, having returned 18.1% annually. Good enough for you? The formula is very restrictive, and tends to throw up very few stocks.

I have decided to broaden and simplify their approach into just two basic criteria: balance sheet safety, and cheapness. Here are the exact criteria I used:
  • market cap > £200m - for adequate size and low spreads
  • z-score > 3 - the balance sheet safety measure
  • PTBV (Price to Tangible Book Value) > 0 - I don't want any company with negative tangible equity
  • PER > 0 - I want to ensure some earnings
  • operating margin > 0 - this just ensures that earnings aren't made positive by exceptional gains
I rank the results by ascending PTBV. It is easy to run this screen using Sharelock Holmes. I then run through the results, from top to bottom, selecting 10 shares. I want to diversify by sector. Seeings as the results tend to be clustered in the same sectors, I allow only two companies per sector. I also apply some very lightweight rationalising to my selection. For example, I will allow two miners in, but not if they're both platinum miners; and I consider oil producers to be the same as miners. I have also excluded REITs, because they tend to trade on low book values anyway, and I think they are similar-ish to housebuilders. Similarly, I have lumped "construction and materials" in with housebuilders. One food producer that cropped up on my list was ACHL (Asian Citrus Holdings), which sells crops in Asia. It has a dual-listing on Hong Kong, which many readers might know is a giant red flag to me. Nevertheless, I have included it in the list. The next share on my screen is AEP (Anglo-Eastern Plantations), which develops in Indonesia and Malaysia. So I have given that one a miss, as it is too similar to ACHL.

I have not investigated the companies in any depth, and have not applied any insight as to how they are likely to perform, their relative merits, and so on. I have simply tried to diversify sectors and applied a minor amount of common sense.

With that in mind, here's the list I came up with:


The prices (SP) quoted are the ask price in pence as of yesterday closing, obtained from Interative Investor. The FTSE All Share stands at 3122. See you in a year's time.


John (UK Value Investor) said...

Sounds like me when I first got into value investing. It's a valid strategy for sure, if you can stick with it long term. My general opinion of investing now, having been doing it as a stock picker for over 3 years, is that the most important thing to do, once you've got a system you're happy with and that you think will work, is to STICK WITH IT!!!

That really is the hard bit. Most investors have much more brainpower than is required to be a good investor so all that spare horsepower goes into reviewing other systems and thinking that the grass is greener.

I hope this system works for you and that in 5 years time you'll still be talking about tangible assets.

Yorkiem said...

Love the moniker 'stingy investor'! A name like that should keep you focused on your strategy.

I like the idea and I suppose that we're all chasing the same things: superior companies at discounted prices.

Have you found an easy or relaible way or back-testing your strategies? I know that Richard tried a few months ago, but that involved him being tied to his spreadsheets for days by all accounts.

Anonymous said...

Thanks for the reminder to keep things simple. My discount screen has been throwing up the same weak companies (which I eliminated during deeper research) for months now.

In desperation to broaden my portfolio I have been getting progressively bogged down trying to find other systems that work (thanks John for putting it so well).

I would love to have a long-term return of 18% pa - heading off to Stingy Investor to read up about it.

Richard Beddard said...

OK, I'm officially confused. This is deep value isn't it? And I approve! Particularly appreciate the operating margin > 0 innovation which is new to me.

Mark Carter said...

I hope this system works for you and that in 5 years time you'll still be talking about tangible assets.

Fingers crossed. Actually, the reason I used a book value measure rather than PER is that Penman's table seems to indicate that low P/B performes slightly better than low P/E. I was also mindful of cyclic anomolies with PEs - i.e. cyclic companies are usually cheap when their PEs are very high, paradoxically. That's why I went with a book value measure.

Have you found an easy or relaible way or back-testing your strategies?

Nope. I think such a thing would be very difficult to do without an expensive database. A lot of work, too!

I would love to have a long-term return of 18%

I hear you on that one ;) I have attempted to capture a lot of their criteria in a single statistic: the z-score. I think that their debt/equity ratio and interest coverage should be amply reflected in the z-score. The current ratio is a very difficult one to achieve, and I think it might be needlessly picky. I am happy with the z-score as a proxy for 3 of their safety features.

I am less happy with their "price to sales" ratio, which I think has some theoretical weaknesses. So I have just gone for the simple value metric of PTBV.

This is deep value isn't it?

Yes, deep value. I am essentially using only two criteria: balance sheet safety (as measured by z-score), and cheapness (low PTBV). The other criteria just try to weed out anomolous statistics; they're not part of the "philisophy" of the screen.

Particularly appreciate the operating margin > 0 innovation which is new to me.

This is just one of those "anomoly" checkers - say for when you have negative profit before exceptionals, but the exceptionals are positive and put the earnings as a whole into positive territory. So there's less to this extra little check i do than meets the eye.

Richard said...

F_Score includes current ratio and debt/total assets but only checks to see if they are improving, not absolute levels.

If you're going to use single year PE/PB I read somewhere combining both is better than either individually!

However, I too am for keeping things simple and using as few variables in my screens as possible.