Forum Archive Index - February 2003

[Travis Morien <travismorien@yahoo.com>]

--- "tennyson@caverock.net.nz"
<tennyson@caverock.net.nz> wrote:
> At last, another person on this forum that can think
> for themselves!
> 
> Excellent response Morgy.  You just nailed Travis. 
> Or did you?  I 
> hope Travis will excuse me for butting in at this
> point, but I think 
> Morgy, you haven't quite got the point that Travis
> was trying to make 
> (probably due to Travis not presenting his argument
> clearly enough 
> for you). 

I'll try to keep it *really* simple in future, so
there are no more understandings from the technical
analysts of the group.
> 
> First up, I have no argument with any of your
> quotes.  Trying to 
> predict earnings one year out is a minefield, two
> years out is 
> incredibly difficult and ten years out is
> practically insane.

Yep, no disagreement there.

See for example:
http://www.afajof.org/Pdf/forthcoming/chan.pdf

But, I never said I was going to nail the earnings,
just test reasonableness.

There are a number of ways of estimating future
earnings, all involve simple arithmetic (no calculus)
and none involve precise forecasting.

Method one: EPS growth by a fixed %pa.

You can take today's EPS and grow it year after year
by an appropriate % growth rate.  You then estimate
dividends by multiplying this EPS by a div payout
ratio in line with historical averages, or your own
estimate.  You also assume a PER that you'll sell the
stock at in ten years time.

ie.. stock today trades at 25x.  After a detailed
analysis of the business, especially comparing this
with historical growth rates for that company and
similar companies, not to mention a healthy dose of
common sense, you think that the company shuld be able
to grow its earnings at between 4%pa and 8%pa.  You
are highly sceptical that the company will grow its
earnings at a greater rate than that because the best
the company ever did over the long term in the past
was 8% and you feel the future will be worse than the
past.  For the purposes of the model you assume that
in 10 years you'll be able to sell this company at 15x
earnings.

*very* simple stuff, I'll leave it as an exercise to
the reader to spend the required 2 minutes making an
Excel spreadsheet that will do this.

After plugging in these assumptions you notice that
the price the stock is trading at in the market is too
high because in order to pay a reasonable set of
dividends plus 15x earnings capital in 10 years the
company must grow at a minimum of 15%pa.  You think
this is highly unlikely, so you pass it by.

Or, you determine that the company only has to grow at
2%pa to justify the present price.  It appears to be a
bargain because this is less than your conservative
estimate.

Method two: Comparing sustainable return on equity
with growth in book value.

Company has $1 per share in book value.  After a look
at the business's past returns, and the trend in
profitability, and the growth in book value compared
with retained earnings, and of course a certain amount
of conservative judgement that incorporates regression
to the mean for extreme ROEs and the same scepticism
of high past performance you estimate the following:

What ROE can the company sustain in the future?

Will each dollar of retained earnings find its way to
increasing the book value of the company, or has BV
grown at a faster or slower rate than retained
earnings?

You also use the assumptions from the first model,
what div payout ratio and what final PE to sell the
stock.

So, our company has $1 per share in equity, based on
our judgement we assume that the company can earn 10%
on this and that historically the company has grown BV
by $1 for every dollar of earnings that they've not
paid as a dividend.

Assuming a 50% div payout and final sell price of 15x
earnings...

Year  BV    EPS   Div   Retained earnings
2003  $1    10c    5c       5c
2004  $1.05  10.5c  5.25c   5.25c
2005  $1.1025   etc

Again, a very simple thing to set up in a spreadsheet.

Real world example, in mid 2002 I noticed that
Brambles (BIL) then selling at 2.54x its book value,
could achieve an estimated return of about 12%ROE.  I
doubted they could do better than that though.  Using
an appropriate set of assumptions about future PE and
div payout and the efficiency of converting EPS -
dividends into an increase in book value I concluded
that BIL was overpriced, in fact I concluded it was
overpriced by about double.  No feasable set of
assumptions that I could make about any of these
parameters showed the possibility of BIL giving a good
return.  When you find that in order to support the
current price you have to come up with unreasonably
optimistic assumptions you know something is wrong.

I made similar observations of a number of high PE
growth stocks, none of them seemed fairly priced based
on any assumption I could make about ROE or growth in
book value.  With forecast returns over 10 years being
low or negative I concluded that these stocks could
not possibly be fairly priced.  My only regret is that
I told all my clients to sell, but didn't tell anyone
to short sell.

(Forecast return = internal rate of return of the
final column which = dividends +- sell price(buy
price)) 

Method three: Estimating sales and net profit margin.

Third model, you notice that a company has
consistently grown its sales per share by about 5%pa
for the last 10 years, and that they've achieved an
average net profit margin (EPS/Sales) of 2%.  Another
simple spreadsheet, take 2002's sales per share, grow
them by 5% and multiply sales by 2%, voila you've
estimated EPS, multiply that by div payout ratio,
multiply the EPS in 10 years by your final PER and
you've got your future cash flows.

So you apply this model to the stock you are thinking
of buying and find that the present price of the stock
is simple too high.  This would call for a large
increase in the rate of sales growth or the profit
margin.  You perform a few experiments to work out
combinations and fail to find a combination of sales &
margin that appears remotely plausible.

Of course, you can combine all of these models
together as I do.  What rate of sales growth is
required to maintain that ROE?  Too high?  Use a lower
ROE then.

What rate of ROE is required to maintain the EPS
growth assuming every dollar of retained earnings adds
to book value?  What rate of sales growth will be
required to maintain that ROE?  What margin?  Etc, and
in reverse.

Setting up such a spreadsheet takes just a few minutes
if you do a simple version (I've got a complex version
that produces charts, calculates after tax returns,
produces sensitivity analyses quite quickly etc). 
Using this spreadsheet takes about 10 seconds.  That
people rarely if ever do this (and that people laugh
at the concept) is disgraceful, but it does explain
why I've been able to identify so many mispriced
stocks and why I've made a profit compared to a market
that has been just plain awful.
> 
> However, there is one special class of company
> where, as strange as 
> it may seem, insanity can prevail.   This is the so
> called 'consumer 
> monopoly' as described in 'the Buffettology
> Workbook', and all those 
> other Buffettology books.   A consumer monopoly is a
> company that is 
> a well established leader in its chosen market, with
> a history of 
> profit growth.  It will generally also show a
> healthy return on 
> shareholders funds and be almost immune from
> competition being able 
> to devastate its profit margins.   There aren't many
> companies that 
> can satisfy these very strict criteria.   A failure
> at any of the 
> qualifying hurdles means any ten year earnings
> prediction is 
> nonsense.  But find the select few companies that
> clear those 
> hurdles, and it can work.  This is what Buffett
> does, but how to find 
> these companies?   It takes hard work, but this is
> the mission of the 
> 'Focus Investment Group' (one of the 'sharechat
> message boards').    
> Welcome aboard, if you would like to help us.

Yes, while it is true that Warren Buffett style
companies are generally the most predictable in terms
of being able to maintain high ROE and good margins,
this method works for any stock.

It doesn't need to be a superior business franchise,
if you do the math and find that the company needs to
grow profits at 20%pa and achieve a 40% ROE and a 20%
sales growth with a large increase in profit margins,
despite never having achieved such results in the past
and in your opinion being unlikely to do so in the
future, then it is almost certainly overpriced. 
Stocks with the opposite criteria, expectations built
in to them so bad that the company almost could not
possibly do that poorly, are probably undervalued.

BIL, CSL and several other stocks were overpriced by
the former criteria according to my models, CLI was
underpriced by the latter criteria.

Without modeling of this sort it simply isn't possible
to say based on any mathematical logic what the future
return of the stock will be - even ballpark figures. 
If you have no idea whether the stock you are buying
is going to average somewhere in the vicinity of 30%pa
or 10%pa or -20%pa because your only buying tool is a
chart then you just aren't an investor.

> There is one more point about predicting future
> earnings that needs 
> to be made clear.  When Travis says he is predicting
> future earnings 
> he does *not* mean he is trying to pin down earnings
> figures five, 
> six or seven years hence.  What Travis is plotting
> here are *future 
> earnings trends*.  An actual result for a particular
> year *may* be 
> right on the trend line.  But the result may also be
> significantly 
> above or significantly below it.   This sort of
> thing does not 
> invalidate his analysis.  Why not?

Yes Snoopy, pretty much.  Trends are important, and my
first analysis usually uses trends to estimate the
future return of the stock.  But it would be more
accurate to say that I use a test of believability. 
If I can believe the numbers that my model tells me i
have more confidence buying that stock than if it
feeds me unbelievable high results.

For an idea of what I mean, I'll leave it as an
exercise to the reader to try to work out what kind of
insanely high return on equity and big PER end
multiple is required to justify owning BIL at $9.50. 
When you are done with that, work out CSL at $50, $40,
$30 and even $20.  You may be shocked by what you see
when you apply this to many "growth" stocks.
> 
> There exists in business a phenomenon known as
> 'regression to the 
> mean'.  Put simply this means businesses have good
> and bad years and 
> although some of these good/bad fluctuations are
> quite unpredictable, 
> they tend to average out in the end.  

A company with 30% ROE is likely to earn a more normal
ROE eventually.  I'll plug an ROE of 10% into the
model to see what drops out.

Same for any number, what does the model say if you
change it from below (above) average to above (below)
average.

> The secret to running a successful business lies in
> hiring the right 
> people, getting the management structure right and
> understanding your 
> market.   Get these basics right and the single year
> earnings 
> forecasting blips pale into insignificance.   Bad
> years are balanced 
> out by good years and looking back over several
> years of trading 
> history the upward trend in earnings will still be
> apparent.  

I agree with all this, but it wasn't really what I was
trying to say.  What I wish to say is that if you
can't run some kind of simple valuation model to work
out what you are buying and what return you might
expect then you clearly don't deserve the title of
investor, you are a speculator.

Speculation is ok, but from what I've seen very few
people are even good speculators.  Most are third rate
speculators whose idea of a viable financial strategy
is to chase the performance of any asset class that
seems to have done well or believe any quack on the
net who says he can tell you what, for example, gold
is going to do, especially if that person says so with
a great deal of force and conviction and can provide
mediocre links on the net to prove the point that
another wacko things the same thing.
 
> Reduced to this level you can see that what Travis
> is saying is that: 
> if you have 
> 
> i/the right people 
> ii/doing the right things 
> iii/ in the right market, 
> 
> and 
> 
> iv/ this combination has been proven to work in the
> past, 
> 
> THEN
> 
> it is reasonable to conclude that the business will
> continue to grow 
> in the future.
> 
> To anyone who has run a business this is just common
> sense.  Travis 
> is *not* predicting that such a business will grow a
> predictable 
> amount every year without fail.  Travis is
> predicting the longer the 
> ingredients of a successful business are there, the
> better the chances 
> are it will be recognized by the market -
> eventually.  And as a long 
> term investor Travis is patient, able to bide his
> time waiting for 
> that 'eventually' to really happen.
> 
> SNOOPY

I'm predicting that CSL can't grow their profits at
50%pa, and anyone that thinks CSL can is a twit.  NB:
I first became bearish on CSL in mid 2001 after
reading a JB Were buy recommendation on the stock
where they did in fact assume this, and stated as much
in black and white.

I'm predicting that BIL's return on equity isn't
suddenly going to jump from 12% to 30% for no reason
other than the market priced that kind of
profitability in, so $9.50 was too much.

I'm predicting that Telstra's sales, which have
averaged 4-8% growth aren't suddenly going to double
that, so I'm not forecasting great things for this
stock.

I'm predicting that CLI ought to be able to grow its
profits at at least -10%pa or better.  Which is why I
told my clients to buy it in December when it was
trading at about $1.60.

I'm predicting that AMP, which trades at a similar PE
multiple on strongly (and temporarily) depressed
earnings and trades at 0.57x book value is more likely
to enjoy the kind of modest profit performance needed
to give a good return than the record high ROEs built
into WOWs price.

A commercial software program that does this sort of
analysis is http://www.stockval.com.au/

There are a number of very nifty articles at that site
as well.

I don't use Stockval myself, having implemented a
similar thing in Excel (before I'd even heard of this
program), but the author of this software also wrote
"A Wonderful Company at a Fair Price" which is a very
good read as well, one of these days I might buy the
software but for now I'm doing fine without it.


> 
> 
> 
> PS  You should also know that the mathematics needed
> to carry out 
> this analysis *can* be carried out on a simple four
> function 
> calculator.   OK, I'll admit that a business or
> scientific calculator 
> might have you pushing less buttons.  But nothing
> anywhere as complex 
> as calculus or heavy algebra is needed to solve
> these 'valuing as a 
> bond' problems.

A spreadsheet is best unless you have a lot of time on
your hands.  You can do all this stuff in your head if
you want, but after you've done it a few times
"because you can" you'll want to get that spreadsheet
done.  The maths only involves +-*/ there aren't even
square roots, let alone calculus or stats.  I'm
familiar with the original Ben Graham quote and
familiar with the context in which it was originally
written.  He was bashing the work of finance academics
whose occupation at that time was complex calculus
relating to the Capital Asset Pricing Model and Modern
Portfolio Theory and the arcane world of derivative
pricing.

Travis
www.travismorien.com

__________________________________________________
Do you Yahoo!?
Yahoo! Shopping - Send Flowers for Valentine's Day
http://shopping.yahoo.com

----------------------------------------------------------------------------
To remove yourself from this list, please use the form at
http://www.sharechat.co.nz/chat/forum/