i want to work with dynamic factor models to model the stock prices and develop a statistical arbitrage strategy based on these models. i looking for resources to understand these models better and maybe some examples.
My favorite paper on this topic is Ng, Victor, Engle, Robert F., Rothschild, Michael (1992)."A multi-dynamic-factor model for stock returns." Journal of Econometrics 52(1-2): 245-266.
I personally already ran some tests, but I did not find any evidence that a dynamic factor model is a good predictor of stock prices (out-of-sample). I also found the same results for Machine Learning Models, such Neural Network. Although in the sample the results are good, out-of-the-sample they are really poor.
If you want a predictor of stock market at the aggregate level, a good paper is:
"Li, Yan and Ng, David T. and Swaminathan, Bhaskaran, Predicting Market Returns Using Aggregate Implied Cost of Capital (March 18, 2013). Journal of Financial Economics (JFE), Forthcoming. Available at SSRN: http://ssrn.com/abstract=1787285 or http://dx.doi.org/10.2139/ssrn.1787285"
I have a research that aims to predict the market at the firm level. If you want to take a look:
"Azevedo, Vitor G., The Implied Cost of Capital and the Time-Series of Expected Returns (June 1, 2016). Available at SSRN: http://ssrn.com/abstract=2787972"
Dear Kamyar,
Have you tried the Fam-French 3 factor model or Carhart 4 factor model?
Recently, the trend is going towards factors such as consumption, habit formation, momentum and liquidity factors.
See page 57 Eq 20.
Hope this helps
Dear Kamyar,
I have not seen any empirical study that shows the comparison of various models in your research topic. But, I want to suggest you test the Neural Network Model as a dynamic and intersting method.
Good luck!
Hello,
Do you mean risk premia factors (size, momentum, vol..) There is a large literature concerning this. with empirical results as well.
Good luck!
- Badges
- Science topic
- Similar topics
- Mathematics
- Statistics