Jonathan

Questions#

I adjusted the tables as you advised in order to have the results per economic region in one concise presentation. These have been split into panels per table which are explained in footnotes below each table.

I have added a section regarding the societal effect of QE

I have done a comparison between participation rates in the US economy and the total QE deployed by the federal reserve

I have added an explanation at the start of the empirical analysis in regards to the selection of the optimal number of lags for the VAR model

In terms of the multicollinearity issue I found multiple papers explain that it will increase the sensitivity of the model, in such a way that a change to the dataset could create large deviations in the results of the model.

Remarks#

  • Fig. 1 and fig. 2 are very nice, and clearly illustrate the relevance of QE.

Based on the results of the analysis we find large differences between the coefficients estimators when the economic variable is the independent variable and when the financial market variable is the independent variable.

  • It would be nice to briefly recap the logic of your methodology first. You use VAR models to estimate the interrelationships. Why?

  • Two aspects:

    • Simultaneity
    • Stationary (cointegrated) time series

  • Be concise and don’t engage in too much theorizing in the introduction (level of detail is sometiems too high in the introduction)

Hypotheses#

Given this interpretation if the coefficients have a difference between the financial and economic variable to be greater than or equal to 0.5.

Why 0.5? I think you just want the coefficients to be statistically significantly different from each other, right?

In order to evaluate the coefficients of this hypothesis and determine whether to reject the null hypothesis, the equation measuring financial market performance would need to have a relatively low lag, most likely to be equal to 1 and have large coefficients of the QE variable which is greater than or equal 1.0.

Similarly, here I don’t understand why you would want to have the coefficient > 1?

H3, H4: What coefficient signs do you expect here? What if you have models with multiple lags?

  • Descriptive stats.: why do you have the skewness and kurtosis reported?

  • You can also determine the long-run equilibrium values of the system, for each regression separately (see any time series econometrics book for the details, relatively easy calculation coming from the coefficients).

  • If some of the models' stationarity assumptions are violated, it might be better to relegate it to the appendix.

  • Is there a reason why you omitted stars indicating statistical significance?