Economic indicators predicting stock market?

D

DVB

After running correlations and building regressing models, I couldn't come up with many significant relationships between economic data and the stock market returns; therefore, I ask the for opinions of anyone who cares to comment. Did you find any statistically significant relationships between economic data: GDP, corp. profits, change in interest rates, etc and the future (or even current) returns of the stock market?
The obvious answer to this question is that stock prices tend to discount the economic data, and by the time the data is out, it is already priced in the market. Therefore, any kind of forward looking relationships will not exist: great GDP growth in the last quarter will not predict stock market returns next quarter. Then following this rationale there has to be a significant relationship b/n current quarter corporate profits and SP500 returns, because as the information about earnings, sales etc. leaks out to the market as the quarter progresses the stock prices will appreciate or depreciate to reflect the fundaments. The core assumption of financial theory is that the stock prices reflect growth in corporate earnings in a long run. Well, the statistics shows otherwise: I ran returns of SP500 vs. Corporate profits by quarter (1970-2003)and the R^2 came out to be 1.5% (!!!). Statistically it means that only 1.5% of SP500 change in any quarter is explained by a change in corporate profits. Makes one question a value of fundamental analysis and attempt to predict corporate earnings, does’t it? Even if one is a forecasting genius, and predicts corporate profits growth with 100% accuracy, it will help to explain 1.5% move in SP500! I may be better of flipping a coin to decide either to go long or short, the odds will be much better.

Any comments?

DVB
 
Before making any "models" one must understand how things really work. It seems that most people don't even understand the basics, so go here
http://www.inflationdeflation.com/
It's an addon to Monopoly game that would make you understand more than 1000 pages of official financial theory based on fake efficiency theory.
 
i have almost 15years of weekly S&P 500 earnings data, most of the fed data, and some other data from the economic consulting company that i have worked for in the past, and have found definate correlations to the market. . .

however, there are problems with the data, such as the structure of the SP500 has changed throughout those 15 years. . .

however, if you want to really discuss this topic, lets,
email me at qqqdaytrader@hotmail.com

and we can discuss your inputs, and expected outputs. . .

sg
 
"Makes one question a value of fundamental analysis and attempt to predict corporate earnings, does’t it? Even if one is a forecasting genius, and predicts corporate profits growth with 100% accuracy, it will help to explain 1.5% move in SP500! I may be better of flipping a coin to decide either to go long or short, the odds will be much better. "

its mostly a waste of time to try and time the market with fundamentals. the lag is too great and if you discover a problem the rest of the market may not care. in the market it doesn't matter until it matters. trying to figure out what will get the market to recognize a problem is not gameable.
a perfect example is right now. it is widely agreed the the us cant continue to run deficits at the rate it is. at this time the market doesn't care about that. at some point it will and react but it will take an event to set it off. when and what that event will be no one can know at this time.
 
I haven't tried an intellegent model of the stock market, though I've played with some things like interest rates and so forth and didn't find anything to pique my interest.

However, I've done a fair amount of modeling/research on exchange rates.

I'll tell you this; conventional theory of rates (PPP and the Quantity Theory of Money) doesn't closely track rates and none of the "definitive" models/research/papers on the subject that produce high r^2 last for long out of sample. My guess it that there are similar problems in the equity markets.

This doesn't mean that it can't be predicted; it means that times change and models must change with the times. With markets built so much around expectations and shock events, short or intermediate term predictions don't hold well.

In the case of PPP, it is a good indication of where things should go, but real prices tend to stray from and come back to it (like prices around a moving average).

You will find ways to "predict" market levels, but unless you are able to build it to react dynamically to the changing marketplace, it won't do you any good if you are looking for a specific price to target/trade/enter/exit.

Good luck and good trading to you,
 
You must log in or register to reply here.
Back
Top