Discussion of Regime Switches, Agents™Beliefs, and Post …

Francesco�s paper Comments Wrapping up

Discussion of "Regime Switches, Agents�Beliefs,and Post-WW II U.S. Macro Dynamics"

by Francesco Bianchi

Discussant: Efrem Castelnuovo (University of Padua)

2nd International Conference in Memory of Carlo GianniniBank of Italy, January 19/20, 2010

Efrem Castelnuovo University of Padua

Discussion Francesco Bianchi�s


Francesco�s paper

What Francesco does

I AD/AS model à la LS (2004)I Regime-dependence of the policy parameters &heteroskedasticity (two independent chains)

I Agents endowed with probability distributions over futureregimes - role for credibility, communication

I Estimates it for the post-WWII U.S. economy - Bayesianalgorithm to estimate MS-DSGE via Gibbs sampling




Francesco�s paper

Main �ndings

I Recurrent policy switches during the post-WWII U.S. period -�pre- vs. post-Volcker�split unwarranted

I Policy switches and heteroskedasticity: Not synchronizedI Beliefs over future regimes: Relevant drivers




Francesco�s paper

Main �ndings


I Policy switches and heteroskedasticity: Not synchronized

I Beliefs over future regimes: Relevant drivers




Francesco�s paper

Main �ndings


I Policy switches and heteroskedasticity: Not synchronizedI Beliefs over future regimes: Relevant drivers




Comments

My reaction

I Very rich paper, worth reading

I Formal evidence in favor of recurrent switching (ML favorsit), beliefs counterfactuals: Novelty!

I Comments on

1. Trend in�ation

2. Indeterminacy and the price puzzle

3. Role of beliefs

4. �Scale e¤ects�

5. Business cycle measurement




Comments

My reaction

I Very rich paper, worth readingI Formal evidence in favor of recurrent switching (ML favorsit), beliefs counterfactuals: Novelty!

I Comments on

1. Trend in�ation


3. Role of beliefs






Comments

My reaction


I Comments on

1. Trend in�ation


3. Role of beliefs






Comments

My reaction


I Comments on

1. Trend in�ation


3. Role of beliefs






Comments

My reaction


I Comments on

1. Trend in�ation


3. Role of beliefs






Comments

My reaction


I Comments on

1. Trend in�ation


3. Role of beliefs






Comments

My reaction


I Comments on

1. Trend in�ation


3. Role of beliefs






Comments

My reaction


I Comments on

1. Trend in�ation


3. Role of beliefs






Comments

Time-varying trend in�ation

"Our results suggest that a reduced innovation variancefor the in�ation target was the single most importantimprovement in monetary policy during theVolcker-Greenspan years." [CPS (2009)]




Comments

Francesco�s assumption

I Constant TI over the entire post-WWII U.S. period:Questionable

I Switching TI: Schorfheide (2005) vs. LWZ (2007): UnclearI Drifting TI: Kozicki and Tinsley (2005), Stock and Watson(2007), Ireland (2007), CSbordone (2008), CPS (2009), ...seems to be there

I Omitted TVTI in a RS world: Consequences?




Comments



I Switching TI: Schorfheide (2005) vs. LWZ (2007): Unclear

I Drifting TI: Kozicki and Tinsley (2005), Stock and Watson(2007), Ireland (2007), CSbordone (2008), CPS (2009), ...seems to be there





Comments








Comments








Comments

TV vs. �xed TI: Consequences

1955 1959 1963 1967 1971 1975 1979 1983 1987 1991 1995 1999 2003 20070

0.2

0.4

0.6

0.8

1Strong Systematic Policy

Martin

Burns

Miller

Volcker

Greenspan

Bernanke

Fixed TargetTrend Inflation

1955 1959 1963 1967 1971 1975 1979 1983 1987 1991 1995 1999 2003 20070

0.2

0.4

0.6

0.8

1High Volatilities

I Castelnuovo, Greco, and Raggi (2010): E¤ects of TVTIEfrem Castelnuovo University of Padua



Comments

Indeterminacy and the VAR facts

I Fact #1. LS (2004): Indeterminacy in the pre-Volcker era,sunspot shocks, in�ation exp. not anchored

I Fact #2. Trivariate VARs: Price puzzle, but just pre-�79(Hanson (2004), Boivin and Giannoni (2006))

I Passive monetary policy and the price puzzle: Any link?




Comments








Comments








Comments

Role of indeterminacy

I Castelnuovo and Surico (2009): Indeterminacy and VARs




Comments

Beliefs, mpolicy, and heteroskedasticity: Facts

I Role of beliefs: How important (vs. policy shits, shocks�svolatility)? On which dimension(s)?

I Counterfactuals to assess their contribution in terms of

1. In�ation persistence

2. Great Moderation




Comments





2. Great Moderation




Comments





2. Great Moderation




Comments





2. Great Moderation




Comments

Small vs. Medium-Scale Models

I Switches: Relevance of the model scale?

I Schorfheide (2005): Evidence in favor of RS TI vs. LWZ(2007): Reject it

I Your paper: Policy switches vs. SZ (2006), JP (2008): Rejectthem

I Policy switches in your fmkt: Omitted information?




Comments


I Switches: Relevance of the model scale?I Schorfheide (2005): Evidence in favor of RS TI vs. LWZ(2007): Reject it






Comments








Comments








Comments

Business cycle as an observable

I HP �ltering: Robust to model misspeci�cation, but ...

1. Distortion of the relevant frequencies (Canova (1998))

2. Perron and Wada (2009): Piecewise-linear trend, worthchecking

I �Multiple �ltering�approach (Canova and Ferroni (2009))




Comments









Comments









Comments









Wrapping up

Wrapping up

I Nice paper! I would like to understand better ...

1. Role of TVTIn�ation

2. Ability of the model to match some GModeration facts

3. Contribution of beliefs

4. How many results survive the �scale e¤ect�

5. Robustness to alternative business cycle representations

I Look forward to reading other papers in Francesco�s agenda



Documents

Discussion of Regime Switches, Agents™Beliefs, and Post …