Luc

Questions#

How am I going to go about measuring the effects of trading volume and investor sentiment individually, by just doing a panel regression with firm and time fixed effects for them on the relevant stocks? I am unsure about this; it is regarding hypothesis 2 and 3 I believe.

  • Hypotheses 2&3:
    • Investor sentiment $\rightarrow$ increased mispricing
    • Business press sentiment $\rightarrow$ increased mispricing

-Both time- and firm-fixed effects are possible, conduct analyses with and without time fixed effects.

  • Regression on the alpha’s (each firm $i$ at time $t$ has a unique observation, ‘abnormal return’)

Is it ok to just have the one hypothesis regarding hype? And then just taking individual conclusions from trading volume and investor sentiment from the OLS?? Maybe I should rephrase the hypotheses 1 and 2 (not mention mispricing, but just mention that they will have a significant effect on something, like put them in an OLS regression with returns, and measure whether or not those two factors have a significant effect on excess returns??)

  • It is not clear to me what you mean exactly. OLS should not yield reliable conclusions in a panel setting.
  • Fixed effects estimates: here

Can I use the Nasdaq 100 instead of the S&P 500? Has to do with Data Availability and information regarding constituents!

  • Via e-mail: it is possible, but it is better to use a somewhat bigger $N$.

Is it ok to do rebalancing and file preparation in Excel, and then export this data to Stata where I will do all the analyses (regressions, fixed effects, etc.?)

  • Yes! If you need any help with this, we can discuss it during the meeting.

Mikael Paaso mentioned that there might be bias, because Trading Volume will be bigger if the firm is bigger, so maybe I have to somehow control for Firm Size, what do you think? How could I go about this?

  • Yes, just include firm size in the regression for the $\alpha$’s. There should be a discussion leading to this fact, and a similar discussion could also lead to the inclusion of other control variables.