Quantitative Modeling: Point Me In The Right Direction
Hey monkeys,
I'm interested in learning more about and developing my own quantitative model to help predict future returns on indexes, currencies, derivatives, or individual stocks. I have a firm grasp on multivariate regressions and advanced hypothesis testing (Significant difference in mean returns, variances etc etc) but I am still confused on how to proceed after regressing data. I can look at an output and understand the numbers (R^2, T-stat etc), but am lost when it comes to taking those factors and further applying it to predicting returns.
Example:
Say I was using returns on the S&P 500 as the dependent variable and was regressing currencies, treasury and corporate yields, and macro data like CPI or retail sales. Let's just say that there is a strong inverse relationship between the index and the difference between 5yr-2yr treasuries. Now what? Do I go back to the data, look for when a steepening yield curve has caused a drop in the S&P, and then hypothesis test those returns against times when the yield curve wasn't steepening, and if its significant, that's a trading strategy??
Forgive me if that is completely wrong, I'm just trying to wrap my head around this. If any of you guys know of books, videos, websites or pretty much ANYTHING that does a solid job of explaining this sort of modeling using real world concepts, please let me know. Also, if you have built a model yourself and wouldn't mind letting me take a peak at it just to get a general sense of what I need to do, I will forever be internet indebted to you. Don't worry, I have no intentions of taking your work. 1) Its ethically wrong (Thank you CFA Institute) and 2) I want to build one from scratch and actually learn the ropes.
Thanks!