Update to the update: apparently, rereading the paper, Dimson, Marsh
and Staunton refer to chap 16 of the Goetmann et al article (by Jorion
and Goetzmann) and claim the J&G article is biased too low since it
excludes foreign stock market dividends. If true, then I guess Dimson
et al are right after all.

RL

raylopez99 wrote:
- quote -

Update: This paper by Dimson, Marsh and Staunton is informative but a
bit too optimistic about worldwide equity returns over the last 100
years.

For example, Table I of Dimson is grossly in error, according to the
data in the Goetzmann and Ibbotson book, chapter 16, "Global Stock
Markets in the 20th Century". The real equity returns in the Dimson
paper are wildly biased upwards (every single country shows upwards
bias from what's reported in Goetzmann et al). For example, Belgium
had a -0.26%/yr return says Geotzmann from 1921 to 1996 while Dimson
reports 2.4%/yr. The years are a bit off (1921 to 1996 for the former
and '1900' to 2005 for the latter--and every single country has a
'1900' value, when Goetzmann specifically says finding good data that
far back in every country is a problem) but if anything World War I
should have hurt Belgium even more, and I can't imagine the years
1995-2005 were boom years to the point that it would affect the
averages this much. Similarly South Africa had a -1.8%/yr return
(real) says Goetzmann et al while Dimson et al report 7.25%/yr.

RL


David Moore wrote:
- quote -

On Tue, 5 Dec 2006 08:23:54 -0600, "raylopez99" <raylopez99[at]yahoo.comwrote:


- quote -

Other resources can be found on this newsgroup's FAQ located at
http://financial-planning.algebra.co...of_Information

To suggest other resources for the FAQ or to help with submissions for
the FAQ, submit comments/material directly to the FAQ Editor (Tad
Borek) at his address mifp_faq at pacbell dot net


-HW "Skip" Weldon
Columbia, SC

David Moore wrote:
- quote -

Thanks.

- quote -

(according to Amazon), but I think part of it is covered in the book I
have (regarding world markets and survivorship bias)

RL

Another fact is that the median investor makes less than 7 percent
long-term real rate of return (see
http://news.morningstar.com/article/....asp?id=178504), which
is the market average. If you add trading costs -which lowers expected
returns by around 1 to 2 percent of total investment in the stock
market-, and progressive income taxes, the puzzle is smaller. I think
that a well-measured model which includes all these facts and
short-term liquidity constraints may well explain most of the puzzle,
even in the context of a standard neoclassical model.
David Moore wrote:
- quote -

A free, recent, and very informative article on the equity premium is
Dimson, Marsh, and Staunton, "The worldwide equity premium: a smaller
puzzle" at http://papers.ssrn.com/sol3/papers.c...ract_id=891620.

You did read the definitive book by the same authors, Triumph of the
Optimists: 101 Years of Global Investment Returns" Princeton U Press,
2002, right?

David

Yeah I'm now reading the part of the book that gets into modeling. The
first, and most widely quoted model, assumes the different risk premia
(not the asset classes, but the risk premia) are statistically
independent of each other. This "building blocks" model adds the risk
premia for inflation, default, horizon, equity together, linearly.

The book mentions the problem you refer to, as the "equity premium
puzzle", first pointed out in 1985 by Mehra and Prescott.

RL

jose.bailen[at]gmail.com wrote:
- quote -

Interesting. When I was at Chicago I remember several seminars on the
"equity premium puzzle": no model seems to explain in a satisfactory
way the huge difference between the long-term rate of return of stocks
(about 7 percent for big caps, in real terms), and the rate of return
of a safe asset (T-bills had an average real rate of return of less
than 0.7 percent between 1926 and 2002). Some models play with
individual preferences -the utility function- although their results
are not validated by empirical evidence, which shows a much lower risk
aversion; some other models play with a loss aversion (asymmetric
preferences, the pain of losing a given amount is greater than the
satisfaction of winning) other models play with liquidity constraints
and incomplete markets (this is the most plausible explanation in my
opinion); some other models play with trading costs, and so on...


raylopez99 wrote:
- quote -

I shelled out quite a bit of money (for my standards) and bought a new
book Equity Risk Premium: Goetzmann and Ibbotson (2006), which is a
collection of historical essays and journal articles with updated
commentary.

If I have time I'll post some highlights.

I'm not a financial planner but the data was surprising.

For example: going back to 1825 (yes!) the stock market was found to
be quite profitable (though in the 19th century the market favored
dividends more than capital gains, and stock prices stayed roughly
level but gave out profits in dividends).

Hence, for total return (large company stocks):

- quote -

Cross-correlation of assets are given, including real estate, corporate
bonds, metals (Au, Ag), etc.

Of interest regarding residential real estate is the low volatility
(STD):

from 1947 to 1978: residential housing: 6.88%/yr (geometric), STD =
3.28%! Compare with US Treasury notes: 3.7%/yr, STD = 3.71%. So real
estate was less volatile than Treasury notes. Amazing.

This review does not do justice to the book--which also gets into the
issue of how to build models for the equity risk premium, survivorship
bias, the global stock market, etc.

Highly recommended.

RL