OT: historical data for total return where? MSCI Barra is not free anymore

[raylopez99]

Does anybody know where I can get free data going back 20 to 25 years
for stock data (US) that includes dividends reinvested? Here's my
dilemma: I can go to Yahoo/Google and get price data for the S&P500,
Dow, etc, but this data does not include the dividend. Usually the
last 10 years the dividend has been rather small, but it skews my
numbers when the dividend is not included in the number. I would like
to get either month-by-month data or year end data for any 'broad' US
stock index. At one point, I was getting this for free at MSCI Barra
but last I logged on they've changed their policy and this data is no
longer available.

So, in summary, I need free data to do the following:

get the total return between two points in time, starting in 1980.
Say I want to know the return of the stock market (including dividends
reinvested) from March 12, 1986 to June 1, 2005. I would get two
numbers for these two dates (or closest date to these two dates), plug
them into my financial calculator, then get the geometric return per
year.

Thank you

RL

[Bernarda Zamora]

I am not sure. José is the expert: he has obtained historical returns
for the XX century from Shiller data
(http://www.econ.yale.edu/~shiller/data.htm)

[raylopez99]

Thanks, very helpful. I need confirmation that I'm doing this
calculation right--that you can add dividend data with the price level
data from Schiller's data.

***Update: see below ***

Example:

I want to calculate the total market return from Schiller's data
between December 2002 and December 2007 (i.e., last five years).
First, I get the data: go to
http://www.econ.yale.edu/~shiller/data.htm then click on "http://
www.econ.yale.edu/~shiller/data/ie_data.xls" (Stock market data used
in my book, Irrational Exuberance [Princeton University Press 2000,
Broadway Books 2001] are available here)

Next, I notice that the data from the spreadsheet for December 2002
is: S&P component: 899.18, while for Dividend: 16.07. Since it
appears this data is summed, that is, dividend is constantly
increasing meaning it's being added to last year's number, I believe I
can add these two numbers together, to get: 899.18 + 16.07 = 915.25.
I plug this into "PV" in my financial calculator.

Next, I take the data from the .xls for December 2007: S&P component:
1479.22, while for Dividend: 27.73. Adding these two numbers gives
1506.95, which I plug into my financial calculator as "FV". I enter
"0" for PMT, I enter 5*12 = 60 for N (months), and then I compute "i
%" (interest, compounded) as: 0.83*12 = 10.01% a year.

So the geometric return is 10.01%/yr

BTW, using Yahoo data for S&P500 with NO dividend, found here:
http://finance.yahoo.com/q/hp?s=%5EGSPC I get roughly the same
number, or using from 12/31/2002 to 12/28/2007 or 879.82 to 1478.49, i
% = .87 *12 = 10.43%/yr, which is greater than 10.01%. There may be
two reasons for this: the price of the dividend is not material (1-2%
a year the last couple of decades), and, the price of the dividend is
"captured"--when people anticipate a year end dividend, the price of
the stock goes up, and, once the
dividend is paid, the price of the stock goes down, so the dividend
distribution is "captured" in the raw non-dividend adjusted data from
Yahoo. If so, this explains why the dividend doesn't matter.
However, I wonder if I'm doing something else wrong, since I notice
that data for S&P500 returns with dividends reinvested for the 3, 5
and 10 year period ending 2/29/2008 in Morningstar is greater than
what I computed using the Yahoo (raw) S&P500 data, specifically,
Morningstar says the returns are (pre-tax): 5.25%, 11.49% and 3.97%
(respectively), while from Yahoo I computed: 4.65%, 10.4% and 2.8%
respectively, always slightly less, which I figured was the dividend
component (hence my original post).

***UPDATE: I found that in fact from Yahoo at http://finance.yahoo.com/q/hp?s=%5EGSPC
the last column *is* dividend adjusted, so you can use this column--
much ado about nothing. However, I did learn something from this
thread: I also found out that during volatile times such as at
present, the compounded long term return can vary dramatically. For
instance for the five years ending in 2/29/2008, the return is 9.21%/
yr, but, if you choose just two days earlier, 2/27/2008, the five year
return is 9.94%/yr, quite a difference for just two days.

Thank you

RL