Foolish economic girly-man

[wabmester]

Charlie, I re-read that section of Bernstein, and I'm convinced that the Gordon Equation is useless, that Bernstein is confused, and that we're both confused

Berstein states in several places that Gordon is only useful for predicting long-term returns. He gives an example based on long-term dividend and growth averages.

He then gives examples using current dividend rates and recent growth to predict future returns, but he insists that you can't use Gordon to predict short-term (e.g., annual) returns, so I have no idea why he's feeding short-term numbers into the equation.

Where you and I differ is that you use short-term values for both terms, and I use a long-term growth term to extrapolate current yields to predicted long-term, but frankly none of them seem very useful or accurate.

We all concede that Gordon doesn't consider P/E growth and that it can't predict changes in dividend (or earnings, or economic) growth, so ... what were we discussing?

 

[haha]

On the other hand, if the market had a 50% correction on Monday, then Gordon would predict a long term return of 16.7% going forward from Tuesday, assuming the yield  and earnings growth rates in $ do not change.  

I think you may be overlooking a few things. It might be best to use some sort of smoothed growth rate, so it reflects more than just the current time year.

But accepting the figures you put forth, prior to the correction you have 1.56% dividend yield and 6.8% "growth". What we are talking about is economic or corporate growth. So after the correction of 50%, ignoring whatever negative effects a 50% correction would like have on economic growth rates, we would have a yield of 1.56%/0.50 or 3.12%. We would then add the growth rate of 6.8%, and get a new Total Return expecation for a buy and hold investor of 9.92%. Not too bad, but remember that we are talking about nominal returns, and to get a growth rate of 6.8% over any long period, we would have to assume inflation of at least 3.8 to 4% pa. I don't have the latest figures, but I think long TIPs are around 2.25% real. Add back the 4% inflation, and we have TIPs at 6.25% nominal.

This is an equity premium considerably less than the historic average. This speaks loud and clear to me- the market could deline by 50% and still not be a bargain.

So invest if you want, but it is more speculating than investing. I must admit that this speculation has worked extremely well for a long time. Probably so well for so long, that deep in the bones many people may not be able to accept what logic tells them.

But The Shadow Knows.

Mikey

 

[wabmester]

... we have TIPs at 6.25% nominal.

This is an equity premium considerably less than the historic average. This speaks loud and clear to me- the market could deline by 50% and still not be a bargain.

Shh, Mikey, you just gave away the fact that TIPS are a bargain

I whole-heartedly agree with you on the inadequate equity risk premium, but to be fair, with 10-year treasuries at 4.3%, any expected growth in stocks north of 8% would justify valuations.

Which is why I think stocks are pretty fairly valued if you only consider today's environment: low interest rates and high earnings growth. Unfortunately, I don't think either one of them will be sustained very long, and stocks will tank when either leg gets kicked out.

 

[charlie]

Mikey, I think you are probably right about using a
"smoothed" earnings growth rate as a proxy for
dividend growth rate.

Let's consider an example:

Suppose a stock was priced at $100 today, and the
dividend is $1.50 and the earnings growth is predicted
to be $6.80 per year. This would translate to 8.3%
annual return per Gordon (accepting earnings growth
is a proxy for dividend growth),

Now if the stock price dropped to $50 tomorrow but
the dividend remained at $1.50 and the earnings
growth remained at $6.80, I think the predicted return
per Gordon would be 16.6%. Of course it is not likely
that the stock would drop 50% unless something was happening to change the earnings growth.

Wab, personally I think Gordon is better at predicting
the total market in the aggregate than an individual
stock. Bogle in his book "Bogle on Mutual Funds"
on page 249 shows a chart of actual vs. forecast returns
for 10 year periods from 1948 to 1992. The Gordon
prediction was not perfect but not bad either. Bogle
claims the correlation between actual vs fcst was .77.
The std deviation between actual and fcst was 3% on
a mean return of 11% per year.

I may be making a mistake in using the current TSM
dividend yield and earnings growth rate to predict
the long term return. However, Bernstein uses
earnings growth rate instead of dividend growth
rate in his example on page 52. I think using the
current dividend yield is accurate, but Mikey is
probably right that we should use a "smoothed"
earnings growth rate or dividend growth rate. I
don't know where to get that number so I just
use the current number published by Vanguard.

You may be right that Gordon is full of s---t
but it is fun talking about it.

Thanks to both of you.

Cheers,

Charlie

 

[cute fuzzy bunny]

Chuck - where ya getting those earnings growth numbers? Those are WAY higher than anything I've seen. I think the morningstar example was using a 1.25% annual earnings growth percentage as the 'expected' number going forward.

 

[wabmester]

Charlie, I did some random reading, and I still think you're abusing Gordon twice:

1) The growth rate is assumed to be constant, so a snapshot of growth is pretty useless.

2) Earnings growth is not a good proxy for dividend growth.   You're basically double-counting since part of that earnings will be used to pay dividends and part will be reinvested for growth.

So, you can use one form or the other.   For dividends, it's:

returns = dividend/price + dividend_growth

and for earnings, it's simply:

returns = earnings/price

Pretty good reference here:

http://www.investopedia.com/articles/04/012104.asp

 

[wabmester]

I think the morningstar example was using a 1.25% annual earnings growth percentage as the 'expected' number going forward.

Morningstar was using "real" growth. Chuck was using nominal.

 

[charlie]

TH, check the "holdings" tab on Vanguard's TSM.
It shows an earnings growth rate of 6.8%.

Wab, the slopes of the earnings growth and dividend
growth charts look about the same to me on P 60 of "4 Pillars...". Bernstein says that the compound annual dividend growth rate is about 4.5% while Bogle uses 6% for the 25 year period prior to 1992. Thanks for
your link. I will check it out. Maybe 6.8% is too high,
but it is not out of the ballpark of reasonableness.

BTW, the quants at Vanguard's Asset Allocation
Fund project a long term annual return for Index
500 to be 9.9%, 5.8% for long bonds and 2.4%
for cash. Check out the "news" tab on AA if you
have time.

Cheers,

Charlie

 

[wabmester]

Charlie, let me know when this horse is dead  

Vanguard is giving you the 5-year average for nominal earnings growth of  TSM.    Would you believe me if I told you that the 5-year average for *dividend* growth of the same was close to zero?

I could only find dividends for the S&P 500 from 1998-2002 to compare, here are the nominal dividend growth numbers:

1998: 4.51%
1999: 3.02%
2000: -2.58%
2001: -3.37%
2002: 0.4%

5-year average: 0.4%

Granted, this is not exactly the same period as Vanguard is looking at, but it's got to be pretty close.  Dividends have simply not grown as fast as earnings.

 

[johnblake]

You people are clearly not 'foolish economic girly-men'.

 

[charlie]

OK Wab, I give up on using earnings growth as a
proxy for dividend growth. Thanks for your research.

I checked the link you provided above. It states that
real dividend growth is roughly 2-3% per year. If
we use 2.5% real then the Gordon model projects
1.56% + 2.5% = 4.06% real for TSM, or about 7%
if you add back in inflation at 3%. The earnings
model cited in your link projects 4.7% real or about
7.7% with inflation. Both numbers suck but we have
to live with them, I suppose.

The horse is dead now as far as I am concerned.

Thanks again!

Charlie

 

[Cut-Throat]

As far as mining, thats the beauty of the gordon equation.  The inputs are all real numbers.  Over decade long periods of time its ability to predict returns is quite solid.

Its simply earnings and dividends.  Theres no ability to mine anything, exclude anything, or enhance anything.  Unless all the companies in the index being analyzed are fudging their earnings.

TH,

I think I'm gonna take a few swings at this dead horse!

If you have been following this thread, it seems that there is agreement on the Equation but not the numbers that go into them

The Answers vary so wildly, I have no idea what to believe.

After reading this thread what would calculate now

 

[cute fuzzy bunny]

This is a worthy read. Guy should have pulled this off his web site.

Some economist in 1997 predicting the future returns of the stock market over a 5-30 year period, using "normal" p/e's and whatnot.

http://sd.znet.com/~schester/financial_advice/expected_return_sp500.html

He laid out several scenarios from "most optimistic" to "pessimistic".

REALLY interesting read...

 

[charlie]

For the record, the author of the link provided by TH
published the article on Feb 3, 1977. His expected
inflation adjusted average annual return was -8%
the optimistic was -5% to 1% and the pessimistic was
-18% for the next 5 years.

The actual average annual return for the 5 year period
was 12.4%. If you assume an average inflation of 3%
over the period, the inflation adjusted average return
was 9.4%.

The really interesting thing about the article was how
wrong he was!! I agree with TH that he should have pulled this off his website.

Cheers,

Charlie

 

[cute fuzzy bunny]

1997, not 1977!

What I found to be ironic is his methodology and measurements were all made using sound economic calculations and reasonable historic valuations and growth numbers.

Which means that either "its all different now" or that we ended up valued way higher than we should be.