Arbitrary value investing help

[sinner]

Hi guys,

I am curious for the input of fundamental traders about a simple question:

Assume you have an investing mandate to invest in ASX20 securities using P/E to define "value".

Is there a P/E ratio, from a fundamental perspective which defines "cheap" in these really large stocks, a level at which you start getting interested in the price compared to the earnings? If so, what is your level? P/E<=10?

 

[prawn_86]

Is there a P/E ratio, from a fundamental perspective which defines "cheap" in these really large stocks, a level at which you start getting interested in the price compared to the earnings? If so, what is your level? P/E<=10?

I think average PE for top stocks is about 14 so anything sub 10 where you think the earnings are stable/growing would be worth a look imo

Really PE shouldnt be used alone and note that the data can also depend where you are getting the PE figure from. If it is historic and based on numbers before a profit down/upgrade then the quoted stats will be wrong.

 

[sinner]

I think average PE for top stocks is about 14 so anything sub 10 where you think the earnings are stable/growing would be worth a look imo

Really PE shouldnt be used alone and note that the data can also depend where you are getting the PE figure from. If it is historic and based on numbers before a profit down/upgrade then the quoted stats will be wrong.

Hey prawn,

thanks for the speedy reply, your first sentence was exactly the sort of answer I was looking for. I understand all the issues around data and fundamental numbers.

Thanks also for the idea of "average P/E" I hadn't considered that as a probably useful metric for this task.

 

[skc]

ASX20 is a bit tricky because there are really a few different sectors.

Banks - Big4
Financials - MQG, QBE, AMP, SUN
Big miner - BHP, RIO, NCM
Telco
Retail - WOW/WES
Energy - WPL, ORG, STO
Health - CSL
Property - WDC
Industrials - BXB

BXB and CSL trades in the PE 20x
Banks probably low teen.
Big miners now high single digit.
WOW used to command 16x now probably low 13-14.
Energy are trading at pretty low PE because of large capex plans

Be mindful of this variability amongst the ASX20 - the usefulness of the "average" figure really dpeends on how you plan to use it.

If you are after a simple "indicator" whether the market presents value, then a historical PE of the stock in question may be a better metric...

 

[craft]

Hi guys,

I am curious for the input of fundamental traders about a simple question:

Assume you have an investing mandate to invest in ASX20 securities using P/E to define "value".

Is there a P/E ratio, from a fundamental perspective which defines "cheap" in these really large stocks, a level at which you start getting interested in the price compared to the earnings? If so, what is your level? P/E<=10?

Short answer – NO.

Slightly longer answer.
Growth and profitability drive cash flow and in turn valuation. Price/current earnings tell you nothing about these aspects going forward. P/E tells you nothing about the financial structure either.

 

[sinner]

If you are after a simple "indicator" whether the market presents value, then a historical PE of the stock in question may be a better metric...

Hey skc, thanks for the response as well.

So there is no "overall market" P/E at which you would be interested in any large stock from any sector if the price was right?

If you are after a simple "indicator" whether the market presents value, then a historical PE of the stock in question may be a better metric...

Yerp I have been looking into CAPEs as one idea, but curious enough about an "arbitrary" number to ask about it as well.

 

[skc]

Hey skc, thanks for the response as well.

So there is no "overall market" P/E at which you would be interested in any large stock from any sector if the price was right?

Sure. I would say 8x or lower would make big cap shares very attractive from a historical perspective, irrespective of the sector.

But if I can buy CSL @ 12x I certainly won't wait for 8x.

 

[sinner]

Short answer – NO.

Slightly longer answer.
Growth and profitability drive cash flow and in turn valuation. Price/current earnings tell you nothing about these aspects going forward. P/E tells you nothing about the financial structure either.

Hey craft, another thanks for the response.

I understand and appreciate your input here. I hope you understood my question correctly, I am not asking whether or not you would/do use P/E to value stocks, I'm simply asking if you had to use P/E to value stocks in the ASX20, is there any number at which you would deem stocks to be "value" above which is "expensive"? If your answer to that question is no, that's ok, just want to clarify we are talking about the same "no". I am not trying to build a valuation model or anything, just had a few thoughts on the tram this morning that I thought you guys might be able to help answer.

Sure. I would say 8x or lower would make big cap shares very attractive from a historical perspective, irrespective of the sector.

Yerp, that was the sort of answer I was interested in. Thanks skc!

 

[craft]

Hey craft, another thanks for the response.

I understand and appreciate your input here. I hope you understood my question correctly, I am not asking whether or not you would/do use P/E to value stocks, I'm simply asking if you had to use P/E to value stocks in the ASX20, is there any number at which you would deem stocks to be "value" above which is "expensive"? If your answer to that question is no, that's ok, just want to clarify we are talking about the same "no". I am not trying to build a valuation model or anything, just had a few thoughts on the tram this morning that I thought you guys might be able to help answer.

Yep it’s a NO for me.

Wouldn't even bother picking an arbitrary number to do a Scan. The amount of junk and good companies returned would be little better then random and I certainly wouldn't trade based on it as too many positions needed to get any historical theoretical advantage based on low P/E studies.

If you forced me to pick a P/E it would be about 8. Based on a nominal GDP since 1974 of 8.1% and an Equity Risk Premium of 5%. [1/13%=7.7]

But to reiterate a stock with a P/E of 8 could still be overvalued depending on its growth and profitability outlook and another stock on a P/E of 16 may be undervalued, so the P/E is meaningless. And then there's the accounting vagrancies in the earnings measure as well.

 

[sinner]

Thanks for clarifying your no and giving in-depth justification for your answer craft, definitely appreciate it.

Just to make clear, this isn't for a scan and the assumption is that I have perfect earnings data.

 

[craft]

Thanks for clarifying your no and giving in-depth justification for your answer craft, definitely appreciate it.

Just to make clear, this isn't for a scan and the assumption is that I have perfect earnings data.

No worries Sinner and hope you turn up something useable. I’m sure there are anomalies based on ratio’s that might be worth exploiting on a statistical basis (the academic paper seems to suggest so). Limiting the initial universe limits the number of positions needed to exploit if you do turn something up, thats a good start.

Top decile of ROIC/net debt would be my choice to investigate. Obscure enough to stay robust, picks up the drivers of value and financial structure.

 

[odds-on]

Hi guys,

I am curious for the input of fundamental traders about a simple question:

Assume you have an investing mandate to invest in ASX20 securities using P/E to define "value".

Is there a P/E ratio, from a fundamental perspective which defines "cheap" in these really large stocks, a level at which you start getting interested in the price compared to the earnings? If so, what is your level? P/E<=10?

Sinner,

In response to your question, i would start to get interested in a single digit PE ratio.

Have you thought about using dividend yield? I would be getting interested in any ASX20 stock that has a double digit grossed up dividend yield especially when comparing the difference to 1 year term deposits against long term average difference. See below.

http://www.macquarie.com.au/mgl/au/.../november-2011/investment-strategy-sleep-well

Cheers

Oddson

 

[sinner]

. Limiting the initial universe limits the number of positions needed to exploit if you do turn something up, thats a good start.

Top decile of ROIC/net debt would be my choice to investigate. Obscure enough to stay robust, picks up the drivers of value and financial structure.

Purely out of curiosity, if I took the top decile of ASX200 based on ROIC/net debt and asked you to invest monthly in that universe using only P/E, would you have a number or would it still be no?

Sinner,

In response to your question, i would start to get interested in a single digit PE ratio.

Have you thought about using dividend yield? I would be getting interested in any ASX20 stock that has a double digit grossed up dividend yield especially when comparing the difference to 1 year term deposits against long term average difference. See below.

http://www.macquarie.com.au/mgl/au/.../november-2011/investment-strategy-sleep-well

Cheers

Oddson

Thanks for you response odds-on, so anything <10 is your number.

I have done a bunch of portfolio research on yield, including double-sort by payout ratio which is an idea I read in a meta-paper that has a lot of merit for yield based strategies.

The purpose of this exercise is different, it's not that I want to use P/E in the conventional way (because otherwise my experience is that no single fundamental metric performs as good as an aggregate/adaptive metric) but since it's not my area of expertise I thought it best to ask for some help. Got exactly what I was after from the exact people I wanted to ask Thanks guys.

 

[tech/a]

Question

Looking back say 5 yrs.
Which has had the best growth and dividend return
Low P/E
High P/E
Or is it not discernable?

Is there perhaps a range?

How much does a companies P/E vary year upon year?

Are P/E ratios ( low ones ) a reliable gauge to share price growth.
Surely someone would have a paper on this!

 

[odds-on]

I remember reading an interview with James Montier and he used a simple Benjamin Graham stock screen to look for stock ideas. He also used it as a useful indicator of the market. If i ever had the data (and time) i would enjoy doing some backtesting, pretty sure when the simple Benjamin Graham stock screen has plenty of candidates then it is the time to buy a dozen microcap spec stocks with cash in the bank!

 

[sinner]

Question

Looking back say 5 yrs.
Which has had the best growth and dividend return
Low P/E
High P/E
Or is it not discernable?

Is there perhaps a range?

How much does a companies P/E vary year upon year?

Are P/E ratios ( low ones ) a reliable gauge to share price growth.
Surely someone would have a paper on this!

Took a bit of digging but turns out there is

http://papers.ssrn.com/sol3/papers.cfm?abstract_id=1254643

The Low P/E Effect and Abnormal Returns for Australian Industrial Firms
Simone Kelly
Bond University - Finance

Jenna McClean
UBS

Ray McNamara
Bond University - School of Business

August 25, 2008
21st Australasian Finance and Banking Conference 2008 Paper

Abstract:
While the low P/E effect has been examined rather extensively in international markets particularly in the US, there is a notable absence of Australian market-based P/E studies. This research examines the relationship between the investment performance of Australian Industrial common stock and their P/E ratios in an attempt to uncover potential for a P/E based trading strategy. The excess and differential returns of P/E ranked portfolios containing 1310 Industrial firms over a 9 year period (January 1998 to December 2006) are examined. The results show the existence of a low P/E effect in the Australian capital market. Furthermore, the superior returns of low P/E stocks increase when a consensus of two business failure prediction models is applied to the portfolio of low P/E stocks. The statistically significant risk-adjusted returns afforded to hypothetical investors over the sample period (up to 12 ¿% per annum), not only provide support for a P/E based trading strategy, but also suggest a violation of the semi-strong form of the Efficient Market Hypothesis.

Please don't say "anything works in a bull market", I am only fulfilling your request for a paper, not stating my opinion on any topic and do wish to actually continue this conversation if possible.

EDIT: Adding in, that the above paper has nothing to really do with my questions, only techs. As I stated to odds-on, my experience of using fundamental factors (and momentum) in the conventional portfolio way provides the best results in aggregate or adaptive form, single metrics do not provide robust results compared to market cap weighting.

 

[craft]

Purely out of curiosity, if I took the top decile of ASX200 based on ROIC/net debt and asked you to invest monthly in that universe using only P/E, would you have a number or would it still be no?

Short answer - Maybe

Considering that the universe would interest me, seeing what value a P/E variable could add would be worth investigating. As for what number to choose for this universe – that’s a parameter I would have to test for, but higher than for the general population at a guess maybe 10-12.

It's a maybe for me mainly because I'm not the best mechanical system tester.

 

[Ves]

Sinner - I ripped the text out of SKC's post because I am lazy.

These are all different in terms of quality so they all have different requirements.

Banks - Big4

CBA - I actually bought this at $44 last year (P/E of around 10, I think - but wouldn't add unless it was closer to $35, small exposure is enough at this stage)
WBC / ANZ - under $18 each looks good (I have small positions in the lower $20s, but wouldn't add to them unless really tempted)
NAB - I don't like this one as much, so P/E would have to be about 7 or 8 to get me thinking

The main caveat for these fundamentally is what skeletons are hiding in the closet if the housing bubble pops? More worrying than cyclical credit growth.

Financials - MQG, QBE, AMP, SUN

Mixed bag here. QBE looked fairly cheap closer to $10. Not sure what that P/E is or was. AMP between 4 and 6, SUN around 8. MQG - I don't like the business model - maybe 4?

Big miner - BHP, RIO, NCM

I think these companies have two things against them: very cyclical industry (must be peaking soon) and the need to be increasingly wasteful in their acquisitions to prop up production numbers as profitability is falling. Don't laugh but I wouldn't touch these unless the P/E was 3 or 4. Even then I would be cautious - I am not a fan of mining companies, I find them hard to value.

Telco - TLS - Probably about 8 to 10, mainly because of the free cash flow. I missed out last year.

Retail - WOW/WES - both about 8 to 10, mainly because they are pretty reliable when it comes to earnings

Energy - WPL, ORG, STO - 3 or 4. Again heavy capex required to grow.

Health - CSL - similar to SKC, between 12 and 14 would be a good start.

Property - WDC - Between 8 and 10

Industrials - BXB - 3 or 4... but even that is stretching it. Horrible company. Didn't they once manage to lose 14 million shipping pellets???

Obviously I agree with craft. Wouldn't seriously look at buying any of them until I looked at the capital structure & profitability metrics. No point buying at any P/E if the E is going backwards or is not likely to go any where for years.

 

[odds-on]

Obviously I agree with craft. Wouldn't seriously look at buying any of them until I looked at the capital structure & profitability metrics. No point buying at any P/E if the E is going backwards or is not likely to go any where for years.

Ves,

Do you think it is actually possible to "value" companies in the ASX20? IMO a company becomes so big it is actually impossible to value them due to the size and complexity of their business operations. Everytime I try and read a Big 4 annual report I get a headache. The larger a company becomes the more it becomes just a "stock" and bar a few cursory balance sheet checks (current ratio, debt levels) it really is a case of understanding how they fit into the ASX "system" and using the correct valuation ratio.

Personally I think dividend yield is very important in the ASX20, especially with the stable earners. Outside the ASX20 I reckon the Price to Sales Ratio used in conjunction with the Current Ratio is a very useful screen.

Cheers

Oddson

 

[sinner]

Sinner - I ripped the text out of SKC's post because I am lazy.

Very good reply Ves, you have given me a few things to think of which I hadn't considered:

It seems like in your case, there is "a number", around 8 at which you would start getting interested, but then another multiplier is applied to the number to get a real opinion rating.

For example default value interest for you might be 8, and then for a company like MQG where one could easily think of a few obvious fundamental caveats we subtract one P multiple for each caveat (or divide by the total number of caveats) to get your "real value" opinion of P/E around 3-4.

That is a pretty interesting concept on its own.