Burnt

So I'm demonstrating a lack of faith feeling that more is required of me than I can physically accomplish. I know through the enabling power of atonement, everything is possible, but I'm just overwhelmed, grumpy, and have high blood pressure. BUT! I decided I'm not going to vent. This blog entry is solely for the purpose of doing a blog entry (points).
Anyways, there's something I found interesting in my Econometrics textbook. I'll translate to normal human words w/o the details of regression analysis:
There's this thing called the efficient markets hypothesis (EMH). We'll set up a model to characterize this concept. Our dependent variable (y)is the weekly percentage return (from Wednesdady close to Wednesday close)on the New York Stock Exchange composite index. A summary of the efficient markets hypothesis states that information observable to the market prior to a certain week, t-1, should not help to predict the return during week t. In other words, if we are using only one independent variable, y(t-1), the expected value of the return on a certain week is...hm, this is hard to word: the expected value of that week? This seems redundant, but what it is saying is that we expect the returns of one week to not depend whatsoever on the week before. Using data from January 1976 to March 1989 (the year I was born!) we get the model:

return = .180 + .059 return (t-1)

What you want to pull from this function is the .059. This (in a 'dummified' way) implies that 6% of returns for a given week can be explained by the returns of the previous week. Aka, investment opportunities will be noticed and will disappear almost instantaneously. Put even more generalized and stupidly, the New York Stock Exchane is almost all luck! (sort of)

However, when you get a t statistic for this regression, you get a value of 1.55, which means that this is statistically insignificant, which means that this isn't very accurate, or telling. Haha, what Econometrics can tell us! Basically, the NY Stock Exchange is a big '?'. You could assume that it's a big '?'. or you could be an 'econometrics-ist' and realize that 2 and 3 dimensional regressions aren't that helpful. If you want accurate data, you need to consider more explanatory variables, which deal w/ 4+ dimensions, and hire economists like Shaunee. :) And make that big money while you're at it.