Discussion of Cyclical Behavior of Unemployment and Job Vacancies Evidence by carlmartin


									         Discussion of

     Cyclical Behavior of
Unemployment and Job Vacancies:
    Evidence from Canada
              Overview of Paper
Motivation: To see how well the Mortensen-Pissarides can
      match business cycle movements in Canada

   Does the “Shimer Puzzle” hold for Canada as well?
                Overview of Paper
Documents cyclical movements in:
   • Unemployment (countercyclical)
   • Vacancies (procyclical)
   • The vacancy to unemployment ratio (procyclical)
   • The job-finding rate (procyclical)
   • The separation rate (countercyclical)

Shows the Beveridge curve exists for Canada
                 Overview of Paper
Presents a simply search model with shocks to labor
productivity, simulates it and compares results to data
     Shows Shimer’s puzzle holds for Canada

Next, try adding Non-Market Activity (a la Hagedorn and
Manovskii (2005)) to fix the problem
      Concludes this doesn’t work
                 Place in Literature:
              Hall (2005) & Shimer (2005) Work on MP model & B.C.

Hagedorn & Manovskii (2005)                MP models with alternate wage determination

      Current paper

   Claim: Canada is basically the same as the US
   (so parameters and model structure virtually identical)

   Therefore can use Canada vs US study as a way of
evaluating Hagedorn and Manovskii (2005) argument that
more complete measures of the value of leisure and home
production can help improve the fit of MP over the business
cycle since countries have different UI systems
   Obvious Question: How Similar are the
     results for Canada and the US?
   Average Separation rate: 0.03 vs. 0.038

   Average Job Duration: 2.8 years vs. 2.5 years

   Magnitude of Unemployment fluctation: 32% vs 38%
   above or below trend

   *Is the job duration implied in Canada consistent with
tenure data?
            Some Business Cycle Statistics
                           U      V      V/U   F       S      P
Standard deviation         0.162   0.237 0.367   0.105  0.096 0.021
                             0.19  0.202 0.382   0.118  0.075   0.02

Quarterlyautocorrelation   0.956    0.956 0.959     0.791    0.795    0.876
                           0.936     0.94 0.941     0.908    0.733    0.878
Correlation matrix U           1   -0.689 -0.851    -0.66    0.682   -0.322
                               1   -0.894 -0.971   -0.949    0.709   -0.408
                     V                  1 0.958     0.712   -0.475    0.568
                                        1 0.975     0.897   -0.684    0.364
                     V/U                       1    0.753   -0.595     0.52
                                               1    0.948   -0.715    0.394
                     F                                  1   -0.155    0.232
                                                        1   -0.574    0.396
                     S                                           1   -0.396
                                                                 1   -0.524
                     P                                                    1
Canadian Case
American Case
American Case
Canadian Case
            My conclusion
 Canada and US not similar on all dimensions
**large difference in role played by separations
                 My conclusion
     Canada and US not similar on all dimensions
    **large difference in role played by separations

This is interesting because it is not what Hall (2005) and
             Shimer (2005) argue for the US

 *Note it is more consistent with evidence presented by
                      Davis (2005)
                  Shimer’s Puzzle
*Although the Shimer Puzzle exists for Canada, the magnitude
                  of the puzzle is different

For the US: Standard MP explains less than 10% of volatility
           of the vacancy to unemployment ratio

   For Canada: Standard MP explains between 15%-30%
        (without adding shocks to separations rate)
                 Other comments:
• Given separations are more important for Canada, should
  consider putting in separations shocks linked to
  productivity shocks

• Note that Shimer sets s=0.1 in his AER article, not 0.09 as
  you claim in the paper (& use saying it is the same value as
  in his calibration)

• Why do you want to use his value for s instead of
  calibrating it directly to the Canadian data given you
  change the other parameters (except r)

• Why calibrate the UI rate in Canada to the Midpoint of the
  replacement range instead of the average or median rate?
               Big Picture Comment
Many papers are focusing on trying to make the MP model
consistent with the business cycle facts, but there is still the
possibility that the MP model isn’t the right one.

E.g., Malcomson & Mavroeidis (2006) suggest that less than
30% of unemployment in the US is due to Matching frictions
(high wage frictions account for slightly more than 10% and
Efficiency wage frictions about 60% according to their

To top