Any sources for daily stock market index valuations for the long term?

February 15, 2009 1:14 PM

Is there a good source for daily stock market information for the last 100 years or so?

I am unsatisfied with the modeling tools available online and would like to do some more sophisticated models on my own. For example, I am interested in modeling the effect of aftershock behaviors (withdrawl, cutting investment, etc.) and whatnot using historical data that isn't _just_ annual returns?

Is there a source of data for the stock market indices available online? Open and close prices would be handy as well, preferably in some easily parsed format (but I will settle for anything that is in machine readable form)?

posted by rr to work & money (9 comments total)

btw, http://schwert.ssb.rochester.edu/dstock.htm is one option but I'd like something covering up to the recent decline. I would also like data that I could use to post an online application for others to use.
posted by rr at 1:17 PM on February 15

Which markets?

biz.yahoo.com has the NYSE daily open/close prices back to 1965. I maintain a small stock index at work and have found their data fairly accurate. There is a "Download to spreadsheet" link at the bottom of the table on the first page that'll download the entire dataset. The problem is in the fine print at the bottom: "By accessing the Yahoo! site, you agree not to redistribute the information found therein. "
posted by jwells at 2:02 PM on February 15

Thanks for the question. NYSE would be fine, but it would be very useful do have data all the way back to pre-depression era.
posted by rr at 2:05 PM on February 15

Robert Shiller has monthly data going back to 1898.
posted by procrastination at 2:15 PM on February 15

thanks for the shiller pointer -- doh, I should have looked there first.

I _really_ want daily data though. The reason is, I'm basically trying to put together a site where people can define behaviors (panic after a big drop and sell, buy what was successful last week, buy on buzz [hence volume] and so on) and then execute a simulation for how those behaviors affected their returns over time. The goal is not to be a quantitative tool but rather to interactively show how investor behavior can cause returns that are the polar opposite of investment returns..
posted by rr at 2:56 PM on February 15

The DJI used to publish all their open and close price data back to like 1896 in .xls, but I can't find it any more. This may be what I am thinking of, but I don't want to register to find out.
posted by 517 at 3:25 PM on February 15

Also this.
posted by 517 at 3:26 PM on February 15

Do you have access to a University library?

DataStream would definitely have this information, and in the level of granularity you're after.

Re: looking back over a century and trying to draw any meaningful conclusions - I can see a few problems with such a long horizon (just off the top of my head late at night mind you)

• Not all indices are created equal; The oh-so-quoted Dow Jones 30 Industrials, for example, is a price weighted index while the S&P 500 is market capitalisation weighted. Direct comparisons of the two series to your suggested impulse function(s) would be problematic.
• On September 16th 2005 the S&P 500 and a few other indices from Standard and Poors were reweighted from market based to free float; while I haven't worked with data from these indices across this event myself, this would almost certainly introduce a bias into results looking at these time series. Not sure how you'd isolate this.
• Many researchers have looked at the phenomenon of Downward Sloping Demand Curves for US equities; over a sufficiently long horizon changes in these demand curves would also introduce a bias (if, of course, one subscribes to this theory - sidenote - we've seen evidence of such curves in other markets e.g., The Nikkei, see Liu, S., 2003, 'Changes in the Nikkei 500: New Evidence for Downward Sloping Demand Curves for Stocks', International Review of Finance ).
• Cyclical demand for shares (correlation with the business cycle?)
• The rise of institutional investors; the last twenty years has seen the emergence of private market entities capable of operating on and influencing the markets on a scale that previously was only attributable to central banks. These entities are capable of moving the markets and its not always clear (like now) when they're actively kicking indices around like a football (like now).
• There are probably some macro factors that will cause problems with direct comparisons e.g., interest rates, disposable income, business cycle, that you'd have to think about isolating from your underlying time series.

All in all an interesting piece of research. Definitely get access to DataStream.
posted by Mutant at 3:42 PM on February 15 [1 favorite has favorites]

Cool, thank you.
posted by rr at 3:44 PM on February 15