The cheapest six stocks on the market
Posted on February 14, 2007 by Richard Beddard
Filed Under Investing, Naked PE |
What is a cheap stock? A low price is not sufficient on its own to define a cheap stock because you don’t know what you’re getting in return. £5,000 might be cheap for a second hand Jag, but not if it’s the welded wreckage of two car crashes and a spray job.
Investors usually relate the price of a company’s shares to some measure of its value. The oldest trick in the book is to divide a company’s earnings into its profits. The resulting number is a ratio, the price to earnings (PE) ratio. A stock with a low PE is cheap because its price is low relative to its earnings. Sadly, this trick is not as effective as you might think.
There are many reasons for the failure of the PE, among them the fact that overzealous accountants can manipulate earnings, but the basic reason is that the price of a company changes with investor’s perceptions of what future earnings might be. If they’re right, the company has a low PE ratio simply because it really is a rusty Jag. Invest in it, and you won’t go far.
Keith Anderson, a lecturer at Durham Business School, has spent the last three years studying the PE ratio. I wrote about some of his findings in an article last year. Analysing decades of data, he declared the PE virtually useless at predicting returns. Studies show that low PE shares may beat the market by 3 or 4%, but once you take trading costs into account you might do better buying an index tracker. Low PE shares aren’t necessarily cheap at the price.
Dr Anderson’s mission is to improve the PE. In a sense, he’s doing what diligent value investors do; identifying companies that are relatively cheap compared to:
- their long-term earnings records,
- other companies in the same business*1, and…
- companies of a similar size.
The difference is, by using econometric techniques to spit out one number, the adjusted ‘naked‘ PE ratio, he’s turning an imprecise art into a science. Here are the six companies listed in the UK with the lowest naked PEs of all:
| name/data (9/2/7) | price | mkt cap £m | PE | 8y PE | naked 8y PE |
| Watermark WMK | 25.25 | 11 | 2.6 | 3.4 | 4.0 |
| MICE MEG | 24.25 | 43 | 4.9 | 4.2 | 3.7 |
| Invensys ISYS | 299.5 | 2,385 | 20.6 | 5.0 | 3.4 |
| Vodafone VOD | 149.25 | 78,822 | 13.6 | 20.7 | 3.3 |
| Autologic ALG | 106.25 | 47 | 13.1 | 3.4 | 2.9 |
| Ferraris FER | 19.75 | 10 | 7.3 | 2.1 | 2.1 |
In the table, the PE is the historical price earnings ratio calculated using earnings per share reported for the latest full year. To calculate the eight year PE, Dr Anderson averages the earnings per share reported for each of the last eight years. Adjusting the eight-year PE ratio for the size of the company and its sector produces the naked eight year PE
Take Invensys, for example. Earnings last year were 14.5p, giving a fairly high PE of 20.6. Average earnings per year over the last eight years were 59.5p, which gives a far lower eight-year PE of 5.0. Typically, a company the size of Invensys would have a PE of 15.6. The average PE for industrials like Invensys is 11.6. Invensys’ eight-year PE is very low in comparison, which is why it has an extremely low naked PE of 3.4.
It’s naked because Dr Anderson’s stripped the PE ratio down and reassembled it so it better describes the potential of the company it’s applied to. He says a six share portfolio of the lowest naked PE shares rebalanced every year would have delivered average annual returns of 39% from 1975 to 2003. The portfolio lost money in only two years.
Should you buy the cheapest six stocks in the market? That’s up to you; whether you are convinced by Dr Anderson’s analysis and you are sufficiently contrarian to buy shares that have, in many cases, recently provided nothing but misery for investors, (I’ve linked the company names in the table to their price charts so you can see how much misery). Dr Anderson’s website explains the calculations in more detail, including links to his academic papers.
Most investors use a combination of statistics and judgement to pick shares. Jim Slater popularised the PEG (price to earnings to growth ratio) in the 1990’s. Ken Fisher wrote a book about the price to sales ratio a decade earlier. The PE ratio is over a hundred years old. Fans say these ratios predicted higher average returns for years (they still do sometimes), but their efficacy declined as they were accepted by investors. Perhaps it’s time for a new ratio.
I’ll return to the naked PE in three months time to update the table and check on how these six companies have done.
Footnotes:
- According to Dr Anderson, and contrary to the normal rule, in sectors like information technology a high PE ratio predicts higher returns. To improve the predictive power of the PE ratio for companies in those sectors he adjusts the naked PE down, if their PEs are above the sector average. I must confess, I’ve never heard of a value investor doing that
- Dr Anderson owns shares in Invensys and Vodafone.
Comments
24 Responses to “The cheapest six stocks on the market”
Leave a Reply
An interesting approach and the underlying approach is indeed what value investors try to do, find cheap but sound shares. As a long term Vodaphone holder I am still waiting patiently!! At least it now pays a little bit of divi.
back in 1991-5 there was a campaign by IFA’s predecessors to use a very similar measure, and the Johnson Fry Hi Y five was touted as a best way to choose shares. There was also a book by a “respected american academic” It of course went pear shaped. Interestingly in 1992 BTR (aka Invensys) was in the list, so maybe history does tell us something.
I was one of the investors in the Johnson Fry Hi 5 PEP in the early 1990s. The promised performance never materialised, and the scheme finally folded up a year or two ago. Anyone still invested at that time had the option of transferring to a conventional unit trust.
I’d be interested to hear how shares were selected for the Johnson Fry ‘Hi 5′ I don’t suppose you have any more details? Even my old friend Google doesn’t seem able to turn anything up
The Johnson Fry HY5 selections followed the system in Jim Slater’s book’How to improve your PEP’ and this followed the ‘Zulu’ principle.
The HY1 share was the second lowest price of the ten highest yielding shares on the FT30 index. If I remember rightly the HY5 was the HY1 share plus the four below that.
My wife had an HY1 from March 1996 to April 2004 and in general the share did very well except for one disastrous year which wiped almost 40% of the capital gain.
For a bit of fun, since May 04, I have tried to replicate the system for an HY1 using the same principle but using the ten highest capitalised companies on the FTSE100. It has done very well(so far).
Thanks Mr Wilkinson, and to the people who emailed me with more details. The trouble is now I’ve got conflicting reports of how the Johnson Fry HY5 selected its shares! At some point I’ll try to resolve them but for now I’ll offer a couple of observations based on the assumption the HY5 was a mechanical system where stocks are bought and sold at regular intervals according to strict criteria.
1. Although Dr Anderson tested his P/E ratio mechanically that’s because there’s no other way to test it, and compare its performance to the standard P/E ratio. He’s not necessarily suggesting people invest that way. Alternatively the naked P/E could be used as one item in the investors’ toolbox: To identify stocks that will potentially do well, subject to further analysis.
2. There are conflicting views on whether mechanical systems beat the market in the long-term. We track the STAR system on the main Interactive Investor site, which have beaten the market handsomely over 20 years or so. I’ve had long conversations with John Mulligan, the former investment banker who runs it. It’s clear that mechanical portfolios can do worse than the stock market for years at a time. If you don’t go in with your eyes open it’s likely you’ll capitulate at the worst possible moment - after years of poor performance. Bearing in mind it would be difficult for a fund to sustain years of redemptions, that *might* (I don’t know the facts) explain the failure of the Johnson Fry fund. For more on mechanical systems and STAR see: A simple system that trounces the FTSE.
My new ways of calculating the P/E are better at predicting winners than the traditional P/E, but that doesn’t mean low P/E shares beat high P/E every year. For example, in my PhD thesis I tested the new rules over subsets of years, and over the five years 1995-2000 high P/E beat low P/E whichever way you calculated it. When researchers say that low P/E beats high P/E, we are only talking about averages over decades. Most investors’ experience doesn’t extend that long, and if a rule fails them over several years they will understandably abandon it, whatever researchers tell them.
Is this the same as the ‘Dogs of the Dow or Ftse?I’ve used this since 1996 and apart from a collapse of 40+ % in one half year it’s worked reasonably well.
[...] a comment on The cheapest six stocks in the market Dr Anderson, a lecturer in business at Durham University Business School, confirms that high PE [...]
Like to learn more.
In “the little book that beats the market” they suggest ranking share by PE (lowest=1) and return on capital (highest=1) then add them together and select the lowest. the website http://www.magicformulainvesting.com does this for US stock. The idead of buying those companies with a good return on capital at lowest price is logical. They also suggest only investing over time in equal money blocks (as harmonic mean cost is lower than mathematical mean) and it eliminates good versus poor timing (NB 1995-2000 was the dotcom bubble so Hi5 would be relatively poor in that short space of time)
PE is only one variable in judging if a business is going to do well in the next few months/years, and thus, if the share price may rise. Better to look at cash generation, profit and USP of business as well.
I am surprised that nobody refers to another ratio, which is often used particularly for technology stocks: PEG or the ratio of P/E to the growth rate. It translates the idea that people are willing to pay high P/E (or seemingly high P/E) for high growth companies. The rule is quite: PEG
It was interesting to see Invensys listed beside Vodaphone. BTR used to say their business was buying and selling businesses. Given that shares are always over priced on the expectation of growth to come and then add on the premium to gain control and they are always paying too much. Deal makers have always to be going for the next deal to keep ahead of the wolf. In the end BTR failed. Think of others - Enron? Vodaphone have just made another megadeal the size of which is beyond comprehension. Where is the button B for you to press to get your money back? There are a variety of techniques for pointing out WHAT to buy but you have to know WHY you are buying it. I am looking for growth so I expect them to rise. If they don’t I sell. I hope this is useful. Doug.
To Charles Goldfinder: I wrote a short section in my PhD thesis on PEG ratios. There is sadly very little academic evidence that they work. This is due to the fact that earnings grow as a random walk, rather than being sustained over several years. There are companies that show sustained growth, but no more than you would expect from chance alone. If the growth rate isn’t sustained then there is no point trying to price it using earnings multiples. This was first pointed out by Little (1962) and was recently confirmed by Chan, Karceski and Lakonishok (2003) using large amounts of US data. I can send you the section from my PhD thesis on this if you are interested.
The inclusion of Invensys suggests something wrong in using the past EIGHT years. It has had a massive change of fortune–crash, then minor recovery– during that time.
Personally, I have found plain old PE served me well. A few years ago, construction companies used to be on PEs of 6 or 7, yet showed consistently rising profits–for at least one very simple and easily spottable reason , no import competition.
In contrast potty media and technology companies were on 20-30. I also put money into these, being dumb enough to listen to ‘analysts’–translation, big US banks giving bogus recommendations, so as to transfer their own losses to their clients and others.
Guess where I lost money and where I gained it.
[...] The predictive power of the dividend yield is enhanced by accounting for share buybacks and issues. Obvious really, so why didn’t I think of that? Especially since I’ve been busy reporting on Dr Keith Anderson’s high performance PE ratio. [...]
Nice work! I am assuming such a scheme should work in other stock markets as well, since results rely on fundamental values. Any idea what these six stocks would be in the US or in the pan-European stock market? (or, put differently: can a simple individual investor easily compute it using free data available on the web?) Thanks!
[...] the predictive power of the price to earnings ratio, first publicised on Interactive Investor (see: The six cheapest stocks in the market). CXO surmises: Careful parsing of earnings history can enhance the use of historical PE as an [...]
To Cuistot: My PhD work concentrated on the UK market, not least because I didn’t then have access to the main US data sources, CRSP and Compustat. Durham Business School does have these, so I hope to produce some similar work using US data in the next few months. Proposals from those more familiar with European markets would be welcome…
[...] sent me new data, revealing how the six shares with the lowest Naked PE ratios, the “six cheapest shares in the market“, he selected in February have done, and which shares have the lowest Naked PE ratios [...]
[...] The cheapest six stocks on the market [...]
[...] The cheapest six stocks on the market (February 2006). An explanation of the Naked Price Earnings Ratio used to select the stocks, with links to Dr Anderson’s research. [...]
[...] See: The cheapest six stocks on the market [...]