Introduction
• Standard approaches in asset pricing and market
efficiency assume rational expectation:
• No learning problem: investors know (·) for ()
• IS return predictability = risk premium/mispricing
• As technology has improved, the set of available
and potentially valuation-relevant predictor
variables expanded enormously.
• →High dimensional prediction problems in finance
• When investors learn about (·) in big data setting,
in-sample predictability takes place in equilibrium.
• → necessity for OOS testing
2022/1/13 Long Zhen 2
评论0