Selecting Good Companies (ROE, PB and PER)

[thinknotleft]

Hi Musicwhiz

So let's say you can find a business which does return >20% ROE for 10 years consistently, does this mean you would pay a premium over book value for it? And how high is your threshold before you will think it's not worth paying for?

Your question depends on how high I wish to set my discount rate to be and my margin of safety (MOF) to be. Taking a discount rate of 10% and a high MOF, I think paying not more than a 2x book will be good for a 20% ROE business. (Don't ask how I derive at this result -- it's just a very rough guesstimate. )

Higher ROEs (if sustained) businesses may warrant an increasing higher P/B ratio (it's a non-linear process. Don't ask how I get this thought -- Derive yourself.)

historical financials do give a good peek at the Company's business model and hence track record which one can use to reliably project into the future. Barring exceptional circumstances or Acts of God or political upheavals, I should say a Company which generates very good FCF and high ROE should continue to do so.

Insofar, I have not encountered research supporting your hypothesis. Will be glad if you can point me to such research supporting
a) High ROEs or FCFs now imply high ROEs or FCFs 10 years later.
b) High ROEs or FCFs, for past 10 years, imply high sustained ROEs or FCFs for next 10 years.
( (a) and (b) are quite testable if one has the data.)

Instead, I have encountered article(s) supporting my hypothesis -- not many high ROEs now will maintain its ROEs years later. For example, Michael Maboussin's article titled 'ROIC Patterns and Shareholder Returns'.

For investing, my personal belief is one should not be a pessimist; neither should one be an optimist. Instead, one should strive hard to be a realist, and always a business analyst. A realist should look at the world through an objective and rational lens, and be as neutral as possible in his analysis (devoid of emotional factors).

Interestingly, a person who is devoid of emotions will not be able to make decisions (which includes investing decisions).

Personally, I do not strive to be neutral in my analysis. Rather, I try to know what's the weakness and assumptions in my analysis.

[cyclone]

Attached is Michael Maboussin's article titled 'ROIC Patterns and Shareholder Returns' in pdf format

[bluechipstamp]

I think paying not more than a 2x book will be good for a 20% ROE business. (Don't ask how I derive at this result -- it's just a very rough guesstimate. )

Instead of thinking in terms of P/B vs ROE, one can just use the P/E ratio. They're the same.

Using your above example, you will pay (P/B)/ROE < 2/0.2.

Since ROE is E/B, you get (P/B)/(E/B) < 2/0.2, or simply, P/E < 10.

Teh Hooi Ling had a article on BT some time ago analyzing coys based on P/B vs ROE. This turns out to be nothing more than a roundabout way of buying coys with low P/E.

[donmihaihai]

A normal assumption goes like this.

If a company generated returns i.e ROE of >20% for last 10 years consistently and if all factors stay the same, then there is a high chance that it will do the same for the next 10 years.

It sound great and look great on paper. But can anyone start throwing out names that
1) has ROE of >20% CONSISTENTLY for 10 years
2) has ROE of >20% CONSISTENTLY for 20 years.

I have none.

[d.o.g.]

It sound great and look great on paper. But can anyone start throwing out names that
1) has ROE of >20% CONSISTENTLY for 10 years
2) has ROE of >20% CONSISTENTLY for 20 years.

I have none.

I can name more than a few, some of which I currently own, used to own, or want to own. But that knowledge is part of my competitive edge, so I won't say more, except that these companies are out there. You just have to look hard enough.

[maniac]

A normal assumption goes like this.

If a company generated returns i.e ROE of >20% for last 10 years consistently and if all factors stay the same, then there is a high chance that it will do the same for the next 10 years.

It sound great and look great on paper. But can anyone start throwing out names that
1) has ROE of >20% CONSISTENTLY for 10 years
2) has ROE of >20% CONSISTENTLY for 20 years.

I have none.

How about

1) CSE-Global
2) Coca-Cola

?

[Musicwhiz]

A normal assumption goes like this.

If a company generated returns i.e ROE of >20% for last 10 years consistently and if all factors stay the same, then there is a high chance that it will do the same for the next 10 years.

It sound great and look great on paper. But can anyone start throwing out names that
1) has ROE of >20% CONSISTENTLY for 10 years
2) has ROE of >20% CONSISTENTLY for 20 years.

I have none.

I'd like to quote some examples of consistently high ROE without debt. It may not conform to 10 years worth of >20% ROE every year, but I am using averages here.

1) Kingsmen Creatives - Average ROE of 21.7% from FY 2003 to FY 2009 (7-years). ROE for 1H FY 2010 (annualized) is thus far 24.8%, on track for >20% again to add to the 8-year average of >20%.

2) SIA Engineering Co. Ltd - Average ROE of 22.5% from FY 2001 to FY 2010 (10 years). ROE for 1Q FY 2011 (annualized) was 21.1%, on track for >20% ROE to add to the 11-year average of >20%.

Of course, these are only two of the businesses I own which exhibit such characteristics; there may be many more others which I have yet to uncover, or perhaps it is fuelled by leverage? The above two examples are companies which have low debt (Kingsmen) or no debt (SIAEC).

[cif5000]

Just some familiar names...SGX and SPH came very close with a 10 year 20% ROE record. One is leveraged while the other is not.

One that really has a 10-year record and I foresee another 10 years is STE. Have a look at the Singapore budget to see where the money is going to.

All the 3 household names are trading nowhere near their book values.

[thinknotleft]

Instead of thinking in terms of P/B vs ROE, one can just use the P/E ratio. They're the same.

Using your above example, you will pay (P/B)/ROE < 2/0.2.

Since ROE is E/B, you get (P/B)/(E/B) < 2/0.2, or simply, P/E < 10.

Teh Hooi Ling had a article on BT some time ago analyzing coys based on P/B vs ROE. This turns out to be nothing more than a roundabout way of buying coys with low P/E.

Hi bluechipstamp

Your logic is not correct. The premise here is that the business must have a 20% ROE first. A P/E of 10 and below does not implies 20% ROE.

Teh Hooi Ling's article is not a roundabout way of buying coys with low P/Es, because you can have low P/Es and low ROEs.

[bluechipstamp]

A P/E of 10 and below does not implies 20% ROE.

Definitely. Your original comment to MW was about deriving the P/B ratio that you're willing to pay for a 20% ROE coy. I'm trying to show that this is simply the same as paying a low P/E for this 20% coy.