Q2. How do the authors calculate the correlation coe cients?
after calculating the Fisher transformation, the authors use the delta method to obtain the standard errors of the correlation coe cients.
Q3. Why do the authors think that the null of no correlation test is biased?
the authors think that this test may be biased to reject the null of no correlation simply because there are more zeroes than ones in the countries’ reference cycles18since expansions are typically longer than recessions.
Q4. What is the definition of the -period ahead forecast error?
Out of this specification, the -period ahead forecast error is+ + = 1X=0+ (6)5Den Haan presents a more general model by allowing for both linear and quadratic deterministic trends.
Q5. What is the way to classify data into groups?
They classify the data into groups ( is given by the user) satisfying the requirements that each group must contains at least one object, and that each object must belong to exactly one group.
Q6. What is the percentage of the group of countries experiencing the first recession of the new century?
In this respect, while more than 80% of the European countries experienced the first recession of the new century, this percentage is less than 40% for the group of accessing countries.
Q7. How does Den Haan propose to solve these problems?
In order to solve these problems he proposes to use the correlations of the VAR forecast errors at di erent horizons as a measure of comovements of the series.