Comments on News and Noise in the Post-Great Recession Recovery by Renato Faccini and Leonardo Melosi Federico Ravenna Danmarks Nationalbank and University of Copenhagen Konstanz, May 2018
The views expressed in this presentation are those of the author and do not necessarily reflect the position of Danmarks Nationalbank
This paper Builds a DSGE model with search frictions in the labor market and nominal rigidities, where TFP news shocks are given a fair chance to explain the business cycle Measures the role of changes in beliefs - noise - on the business cycle Estimates the share of the employment dynamicsduring the post-great recession recovery that can be ascribed to changes in beliefs
Context News of future TFP increases result in a fall in labor supply because of a wealth effect. To obtain a countervailing increase in labor demand, introduce search frictions and a quadratic post-matching hiring cost. This ensures that TFP news shock result in buildup of employment ahead of TFP realization.
Comments
Comments 1 The choice of hiring costs: implications 2 What should be part of the Great Recession history in a model with search unemployment 3 Estimation: a success and two puzzles 4 What should the benchmark be for a model trying to explain the post-great recession recovery?
The choice of hiring costs: implications
The choice of hiring costs: implications Wholesale firms must post vacancies to obtain new employees. If a job produces output Z t and W t is the wage paid to the worker, than the value of a filled job in terms of final goods is Q N t = Z t P w t P t W t P t + (1 δ N )E t Λ t,t+1 Q N t+1, Assume: cost of posting vacancies κv t From the job posting condition, q t Q N t = κ, q t is the probability of filling a vacancy, so Z t P w t P t = W t + κ (1 δ N )E t Λ t,t+1 P t q t κ q t+1 If κ = 0, this yields the standard result that P w t /P t = W t/p t Z t = MC t.
The choice of hiring costs: implication 1 Q N t = Z t P w t P t W t P t + (1 δ N )E t Λ t,t+1 Q N t+1, Assume post-match hiring cost: G [q t V t (i)/n t (i)] Obtain Q N t P w t = Z t W t + G +(1 δ N )E t Λ t,t+1 Q N t+1 P t P t N, }{{} t Hiring cost not associated with number of vacancy posted, only with number of new hires, so not directly affected by aggregate labor market conditions.
The choice of hiring costs: implication 2 Q N t = Z t P w t P t W t P t + (1 δ N )E t Λ t,t+1 Q N t+1, Assume post-match hiring cost: G [q t V t (i)/n t (i)] Obtain Q N t P w t = Z t W t + G +(1 δ N )E t Λ t,t+1 Q N t+1 P t P t N, }{{} t Shocks to matching efficiency do not affect equilibrium since choice variable is number of hires, and vacancies adjust residually.
The choice of hiring costs: implication 3 Q N t = Z t P w t P t W t P t + (1 δ N )E t Λ t,t+1 Q N t+1, Assume post-match hiring cost: G [q t V t (i)/n t (i)] Obtain Q N t P w t = Z t W t + G +(1 δ N )E t Λ t,t+1 Q N t+1 P t P t N, }{{} t There is no congestion externality: equivalent to model with quadratic cost of adjusting number of workers.
The choice of hiring costs: implication 4 Q N t = Z t P w t P t W t P t + (1 δ N )E t Λ t,t+1 Q N t+1, Assume post-match hiring cost: G [q t V t (i)/n t (i)] Obtain Q N t P w t = Z t W t + G P w t +(1 δ N )E t Λ t,t+1 Q N t+1 P t P t N t P, }{{ t } Assumption: costs are in terms of wholesale good, so changes in markup directly affect costs of hiring
The choice of hiring costs: implication 5 Q N t = Z t P w t P t W t P t + (1 δ N )E t Λ t,t+1 Q N t+1, Assume post-match hiring cost: G [q t V t (i)/n t (i)] Obtain Q N t P w t = Z t W t + G P w t +(1 δ N )E t Λ t,t+1 Q N t+1 P t P t N t P, }{{ t } It matters if costs are sunk or not: non-sunk costs can be passed on to new employees as lower wages, since they are not paid until the bargain is concluded! Lower volatility of labor market tightness
The choice of hiring costs Hiring costs help the model because they imply smooth adjustment of labor. Positive news brings about a fall in hiring costs. Estimation favors post-match hiring costs. Pre-match hiring costs negligible? Recovery: evidence of hard-to-fill positions Low hiring rate. June 2014: 24 days to fill a vacancy. Longest delay since 2001. At beginning of recovery: 16 days
The choice of hiring costs
What should be part of the Great Recession history in a model with search unemployment
[A] Volatility in separations
[B] Behaviour of hours per worker Nonfarm Business Sector
[C] Composition of unemployment TFP of employed workers may be very different from TFP of unemployed workers
6.00 5.00 4.00 3.00 2.00 1.00 College 0.00 1.00 2.00 3.00 1992 1994 1996 1998 2000 2002 2004 2006 2008 2010 Unemployment rate by educational attainment (demeaned) >24
6.00 5.00 4.00 3.00 2.00 1.00 0.00 1.00 College High School 2.00 3.00 1992 1994 1996 1998 2000 2002 2004 2006 2008 2010 Unemployment rate by educational attainment (demeaned) >24
Estimation: a success and two puzzles
The role of news shocks in the estimated model TFP shocks and changes in beliefs can explain an important share of unemployment over the business cycle
Behavior of consumption Only 1% of consumption growth explained by TFP and TFP news-shocks. But these shocks have a large and persistent impact on consumption
Behavior of consumption
Consumer sentiment and beliefs University of Michigan s consumer sentiment index seems highly correlated with TFP news shocks (and noise)
115.0 0.8 105.0 0.3 95.0 85.0 0.2 75.0 Sentiment and TFP news shock 0.7 65.0 55.0 1.2
6.0 115.0 5.0 105.0 4.0 95.0 3.0 85.0 2.0 75.0 Sentiment and SPF Unemployment fcst 1.0 65.0 0.0 55.0 1.0
Consumer sentiment and beliefs
What should the benchmark be for a model trying to explain the post-great recession recovery?
What is the benchmark? The model uses noise to explain divergence between TFP and unemployment in recovery. What other shocks could play this role? Christiano, Eichenbaum, Trabandt (2015) TFP can explain missing deflation during great recession Consumption wedge Financial wedge How can we use the data to identify uniquely beliefs shocks?