PE ratio

[Snoopy]

I've taken historical ASX data from Dec 2007 and put it alongside current and deleted any that don't match (which means anything that has substantially changed it's name, gone broke or been taken over). Then I ran a straight price comparison and deleted any obvious recapitalisations. Data is not going to be perfect, but should be enough to be interesting.

Perhaps surprisingly, the most significant difference between the gainers and the market averages is in P/E - with the average much higher for the gainers at the beginning of the measured period! ...

Lizard, can you explain why your P/E statistics for December 2007 are so high? I would describe the December 2007 whole of market P/E of 30.2 as high (perhaps not surprising at the end of a bull market run), but the 'gainers' historic December 2007 P/E of over 50 is extraordinary. Is it possible that because the gainers sample was so small that you have identified some kind of statistical aberration? Just who were these gainers and can you identify any underlying theme that is driving this very high -in absolute terms- P/E result?

Also, how did you calculations deal with P/E ratios for those companies that reported negative earnings?

...ROE was much lower, as was yield.

That ties in with other research results which shows the best overall outperformance statistic is to look for companies with valuable underlying assets in relation to sharemarket price, regardless of P/E.

Pr/NTA was slightly better, but not considerably and has ended the year higher,

Interesting is the decline in market capitalisation for 'the market' of 40% in parallel with the decline in 'price to net asset value' of 44% (seems logical, assets are not worth what the market once thought they were and so the market price declines in tandem)...

...seen in comparison with the *increase* in market capitalisation for 'the gainers' of 35%, compared to a decline in 'price to net asset value' of 31% for 'the gainers'. Why is the market taking such an optimistic view of 'the gainers' (boosting market capitalisation) yet the market price payable for those gainers underlying net asset values appears to have shrunk? Am I reading your chart correctly Lizard?

while the ROE for the gainers is now consistent with the market.

Brought about by new accounting rules requiring assets values to be marked to a market valuation based on the return those assets give?

The most notable point from this is that the market cap has gone up for "gainers", while P/E has fallen.....

The gainers P/E has fallen from around '50' Lizard, which as I said before was an extraordinarily high absolute figure. Surely you could only expect the P/E to fall from those levels?

The change in both these numbers (P/E and Market Cap) can be used to calculate that the average change in earnings for the gainers was an increase of 104%! Earnings for the market as a whole must have fallen by an average of 6.4%.

You have recorded the market prices precisely but what about the Earnings? Is it possible that because your comparison covers only 11 months that some of those reported earnings have not been updated year on year?

SNOOPY

[Lizard]

Snoopy, as you can probably imagine, it is a pretty rough exercise and dependent on the accuracy of the database source.

The averages for P/E exclude negative data. Therefore, a few very high multiples in the "gainers" have skewed the data - the median turns out to be a P/E of 16 at start for both "market" and "gainers".

Looking over the data, it looks like the gainers include a few large caps with small gains (e.g. CSL), some takeover targets (ORG, RPM, SHG) with a larger proportion of minnows - perhaps businesses emerging from the "embryonic" stage such as QML and ACG or resource minnows who struck it lucky (MAK?).

The Pr/NTA data is absent in the database for many very small companies or where negative and therefore cannot be included. The anomaly in the data for "gainers" seems to come from data for some of these companies being added to the database as their market cap increased. So unfortunately, perhaps not too meaningful!

I have since tried analysing results for different groups of stocks based on starting parameters (e.g. P/E<10) and not alot stands out - they are all pretty grim and fairly similar averages of 50-60% fall! Larger stocks seemed to have generally fallen less.

Not alot of lessons for what to hold through a bear market - except "don't".

[Snoopy]

Snoopy, as you can probably imagine, it is a pretty rough exercise and dependent on the accuracy of the database source.

The averages for P/E exclude negative data. Therefore, a few very high multiples in the "gainers" have skewed the data - the median turns out to be a P/E of 16 at start for both "market" and "gainers".

Looking over the data, it looks like the gainers include a few large caps with small gains (e.g. CSL), some takeover targets (ORG, RPM, SHG) with a larger proportion of minnows - perhaps businesses emerging from the "embryonic" stage such as QML and ACG or resource minnows who struck it lucky (MAK?).

OK, so it looks like your first finding Lizard:

"The most significant difference between the gainers and the market averages is in P/E - with the average much higher for the gainers at the beginning of the measured period!"

is not so significant if you change from the 'arithmetic average' to the 'median average'.

I guess that was always on the cards with a small sample that can be easily skewed by one or two pieces of outlier data.

Snoopy wrote:

"Why is the market taking such an optimistic view of 'the gainers' (boosting market capitalisation) yet the market price payable for those gainers underlying net asset values appears to have shrunk?"

The anomaly in the data for "gainers" seems to come from data for some of these companies being added to the database as their market cap increased. So unfortunately, perhaps not too meaningful!

Perhaps to balance out this 'new company' data anomoly, you could remove the December 2007 market capitalisation of those 'old' companies that went bust over the year?

I have since tried analysing results for different groups of stocks based on starting parameters (e.g. P/E<10) and not alot stands out - they are all pretty grim and fairly similar averages of 50-60% fall! Larger stocks seemed to have generally fallen less.

Not a lot of lessons for what to hold through a bear market - except "don't".

I think it is more a matter of selecting stocks by just one simple statistic (P/E ratio) not being a credible investment strategy, no matter how good that statistic appears to be on its own. By all means use P/E, but use other statistics in tandem.

In a 'financial crisis', I would expect those companies that do 'well' (in relative terms) to have quite a high P/E ratio. Why? Because companies that have their earnings valued highly by the market are valued highly because of the robustness of their underlying earnings streams in *all* market conditions.

SNOOPY

[Snoopy]

"In a 'financial crisis', I would expect those companies that do 'well' (in relative terms) to have quite a high P/E ratio. Why? Because companies that have their earnings valued highly by the market are valued highly because of the robustness of their underlying earnings streams in *all* market conditions."

Well Snoopy, Id love for you to find one of these stocks?

Who has a high PE in this market that has maintained a high PE because of robust underlying earnings

underlying earnings are so unpredicable now, I beg to disagree

to have both now is nearly impossible?? no

Out of my own portfolio I will pick three shares with

"a high PE in this market that has maintained a high PE because of robust underlying earnings"

for you Underdog.

1/ Lyttelton Port of Christchurch: PE of 22.8 before yesterday's bid, PE now is 26.6
2/ Contact Energy: PE of 15.9, even after yesterday's share price plunge. It was over 17 with the share price over $7.
3/ Lion Nathan: PE of 17.8 for FY2007. Haven't recalculated for this year yet. But earnings are up modestly 5%? while the share price has fallen by about 5%. That leaves the PE at around 16 which is being depressed by the LNN's hostile bid for Coca Cola Amatil which is on the table.

None of these companies is forecasting an earnings decrease. LNN is quite the opposite, predicting a noticable step up in earnings for FY2009.

SNOOPY

discl: hold LPC, CEN, LNN

[OldRider]

Google "Z score", to find a formula for checking a companies health.

I have a disproportionate investment in Qmastor for the size of the company,
Z score for QML is 10.71 which leaves me happy with the investment.

[Lizard]

Came across Piotroski the other day and thought he might belong on this thread as another possible fundamental stock-picking method. Suspect the theory works best during a market beat-up though, as imagine the stocks get hard to find when the market is going okay.

Basic theory is to select from stocks with a low Price/NAV (lowest 20%) and then screen them for 9 success factors:

• positive ROA,
• positive Op Cashflow/Assets,
• annual increase in ROA,
• Op Cashflow > Net Profit (before abnormals),
• decreasing "leverage" long-term debt/assets,
• increasing current ratio (current assets/current liabilities),
• no new equity issued in past year,
• increasing annual % gross margin,
• increasing asset turnover (sales/assets).

The paper discussing the test of this theory is found here.

According to Forbes, it seemed the theory worked pretty well in 2008. Not particularly surprising - Pr/NTA based selection usually works best in a market rout. The screens would definitely help - and makes me wonder if just using the screens and ignoring the Pr/NTA might be more successful in other years.

Unfortunately, it's all very well having a nice screen tool like this, but having the database that will allow it to be used is probably not realistic. And, if doing the calcs manually, then might as well be working to look for stocks that seem likely to pass those screens on the next result, rather than on the historical one.

[OldRider]

It would look to me after a quick read, a fairly reliable set of criteria for sorting out growth companies.
I use a scan which is simpler, though similar in that the result is strengthened by a high return on assets.
Its advantage is its simplicity as the data required can be got without much difficulty.

The results need to be sifted because there is always a predominance of IT and similar companies, which
have as much potential to go broke as they do to grow, however
over the years it has produced for me a number of gems, Oamps, Data #3 and Worley come to mind.

Interestingly, when I ran the project a month or two ago, on combined ASX and NZX data, not a single
NZX company made the list for further study. Over the years I have mentioned the formula I use without
attracting much interest, perhaps for new posters you could google CGVI if interested, the idea is not new, it had
its origin with Buffet and Munger many year ago.

The system though has little if any use for short term traders but for longer term investors I think it has it merits.

[BIRMANBOY]

Hi lizard, I was nodding off by page 3!! I suppose everyone is constantly looking for the perfect investment strategy and it seems that a fair proportion of traders have gone into the mode of "the more complicated the system the more probable the chance of beating the averages". What happened to the "buy low sell high" system. Surely this has the perfect formula...buy when a share is in a trough (not in a death spiral) and sell when it reaches a point of resistance. The great thing is you dont need any formulas just the historical charts to provide you with the necessary data. Too simple??? I dont know. What do the readers think?

Came across Piotroski the other day and thought he might belong on this thread as another possible fundamental stock-picking method. Suspect the theory works best during a market beat-up though, as imagine the stocks get hard to find when the market is going okay.


Basic theory is to select from stocks with a low Price/NAV (lowest 20%) and then screen them for 9 success factors:

• positive ROA,
• positive Op Cashflow/Assets,
• annual increase in ROA,
• Op Cashflow > Net Profit (before abnormals),
• decreasing "leverage" long-term debt/assets,
• increasing current ratio (current assets/current liabilities),
• no new equity issued in past year,
• increasing annual % gross margin,
• increasing asset turnover (sales/assets).

The paper discussing the test of this theory is found here.

According to Forbes, it seemed the theory worked pretty well in 2008. Not particularly surprising - Pr/NTA based selection usually works best in a market rout. The screens would definitely help - and makes me wonder if just using the screens and ignoring the Pr/NTA might be more successful in other years.

Unfortunately, it's all very well having a nice screen tool like this, but having the database that will allow it to be used is probably not realistic. And, if doing the calcs manually, then might as well be working to look for stocks that seem likely to pass those screens on the next result, rather than on the historical one.

[JemT]

Simple, at least to start with, is always the best way of doing this sort of thing. I am only just getting into company valuations (that is studying them, sadly I am not sure I can find a job actually doing them), so the only advice I can give at this stage is to start out simple then refine your techniques by adding layers. Make sure you always take a step back and look at your results from a common sense point of view. One of the worst valuations I have seen involved some guys doing a highly technical analysis, involving about 8 years' of past data, which came out with a valuation that could not be corroborated with the company itself.

Regarding specific ratios, from what I have learned:

Price/Sales can be useful in identifying overvalued companies (P/S > 2-3), but it falls down due to the fact that it does notreally take into account the operations of the company. Some companies might be overly indebted which means a large interest expense taking a chunk out of their bottom line, likewise some might have managers pissing away cash on various things, or large capital maintence requirements.

P/E is so popular that it has been suggested that (some) companies are (may be) valued on this basis. However you need to be careful in the areas of leverage, nonsensical profit figures (i.e. the companies who took a big deferred tax hit in 2010, or who have restructuring charges), and high growth expectations (internet companies/companies going into China etc).

EV/EBITDA, if you can get reliable comparative multiples then this is a good measure to use. Just remember you must always subtract debt, add surplus assets, and subtract minority interests to get the equity figure. It would be a big mistake if you forgot to subtract debt and perceived the company to be undervalued.

Price/Free Cash flow has issues around the consistent calculation of freee cash flow. There are so many ways to do this, so to get a good comparative figure you might need to compute the comparable companies' FCF figures yourself.

The best ratio often depends on the nature of the company. A cash flow based ratio is more suitable to a company that generates a lot of cash flow (Skycity, Vector, etc). An earnings growth based ratio will be better suited to a company with reasonable growth prospects.

This is just how I see things after about 10 weeks' of studying this area. The advice on here is fantastic because it is all so practical and based on real experience; textbooks only get you so far, and I am quite convinced that my university lecture does not actually know the first thing about buying/selling shares.

[buns]

I’m an Multiples (some could say lazy) guy

I will use the EBITDA Multiple if I don’t understand their capex very well, know they don’t spend much on it, or simply have no chance of finding the investment and maintenance CAPEX side of it. Because like the PE ratio ignores the balance sheet, I think the EBITDA multiple ignores a lot of the cash flows.

An EBIT/EV multiple reigns supreme, as this takes into account depreciation which is a result of your stay in businesses (maintenance I call them) assets, hence any increase and decrease in those should align to the same trend in CAPEX spend.

The downfall with this is that depreciation relates to prior periods, and current capex spend won’t be reflected. So if the business has embarked on any kind of growth or a major change in its asset base you would think the current deprecation numbers might not reflect the future ones from your current capex spend.

Taking this into account, one of my new favourites is the FCF/EV ratio – essentially it gives a time value, or pay back on your investment. For FCF I just use operational cash flow which is EBITDA less maintenance capex.

As a whole, using these is definitely an art. Depending on a companies size/maturity, or stage in its lifecycle the combinations of comparatives you use should be very different. I meantioned it on the RYM thread, that all of my comparables (valuation skill set) didn’t help to much for a company like that which demanded a higher valuation because of other things.

You need to be carefull.. Telecom for the last 5/6 years has had an EV/EBITDA around 5 or 6 and a PE below 10… Where would that have got you

The Devil is in the detail.