Regional President Speech

Mainideas Whattheauthorsdo Interpretations Unstableisnotoptimal Otheranalyses Whyworry Discussion of Evans and Honkapohja, “Robust Learning Stability.” JamesBullard PresidentandCEO FederalReserveBankofSt. Louis1 28May2008 FrontiersinMonetaryTheoryandPolicy—IMES,BOJ 1 ViewsexpressedarethoseoftheauthoranddonotnecessarilyreflectofficialpositionsoftheFOMCorthe FederalReserveSystem.

... solongassmallexpectationaldeviationsfromRE (cid:15) dissipate,insteadofaccumulating. Rationalexpectationsequilibriumisattheheartofthe (cid:15) learningliterature. Butrationalexpectationsisnotapracticalassumption. (cid:15) Actualprivatesectorexpectationsareunlikelytobe (cid:15) preciselyconsistentwithREE. Thatisstillallright... (cid:15) Thatconditionisknownasexpectationalstability. (cid:15) Mainideas Whattheauthorsdo Interpretations Unstableisnotoptimal Otheranalyses Whyworry Learning and monetary policy Rationalexpectationsisacornerstoneofmodern (cid:15) macroeconomictheory.

... solongassmallexpectationaldeviationsfromRE (cid:15) dissipate,insteadofaccumulating. Butrationalexpectationsisnotapracticalassumption. (cid:15) Actualprivatesectorexpectationsareunlikelytobe (cid:15) preciselyconsistentwithREE. Thatisstillallright... (cid:15) Thatconditionisknownasexpectationalstability. (cid:15) Mainideas Whattheauthorsdo Interpretations Unstableisnotoptimal Otheranalyses Whyworry Learning and monetary policy Rationalexpectationsisacornerstoneofmodern (cid:15) macroeconomictheory. Rationalexpectationsequilibriumisattheheartofthe (cid:15) learningliterature.

... solongassmallexpectationaldeviationsfromRE (cid:15) dissipate,insteadofaccumulating. Actualprivatesectorexpectationsareunlikelytobe (cid:15) preciselyconsistentwithREE. Thatisstillallright... (cid:15) Thatconditionisknownasexpectationalstability. (cid:15) Mainideas Whattheauthorsdo Interpretations Unstableisnotoptimal Otheranalyses Whyworry Learning and monetary policy Rationalexpectationsisacornerstoneofmodern (cid:15) macroeconomictheory. Rationalexpectationsequilibriumisattheheartofthe (cid:15) learningliterature. Butrationalexpectationsisnotapracticalassumption. (cid:15)

... solongassmallexpectationaldeviationsfromRE (cid:15) dissipate,insteadofaccumulating. Thatisstillallright... (cid:15) Thatconditionisknownasexpectationalstability. (cid:15) Mainideas Whattheauthorsdo Interpretations Unstableisnotoptimal Otheranalyses Whyworry Learning and monetary policy Rationalexpectationsisacornerstoneofmodern (cid:15) macroeconomictheory. Rationalexpectationsequilibriumisattheheartofthe (cid:15) learningliterature. Butrationalexpectationsisnotapracticalassumption. (cid:15) Actualprivatesectorexpectationsareunlikelytobe (cid:15) preciselyconsistentwithREE.

... solongassmallexpectationaldeviationsfromRE (cid:15) dissipate,insteadofaccumulating. Thatconditionisknownasexpectationalstability. (cid:15) Mainideas Whattheauthorsdo Interpretations Unstableisnotoptimal Otheranalyses Whyworry Learning and monetary policy Rationalexpectationsisacornerstoneofmodern (cid:15) macroeconomictheory. Rationalexpectationsequilibriumisattheheartofthe (cid:15) learningliterature. Butrationalexpectationsisnotapracticalassumption. (cid:15) Actualprivatesectorexpectationsareunlikelytobe (cid:15) preciselyconsistentwithREE. Thatisstillallright... (cid:15)

Thatconditionisknownasexpectationalstability. (cid:15) Mainideas Whattheauthorsdo Interpretations Unstableisnotoptimal Otheranalyses Whyworry Learning and monetary policy Rationalexpectationsisacornerstoneofmodern (cid:15) macroeconomictheory. Rationalexpectationsequilibriumisattheheartofthe (cid:15) learningliterature. Butrationalexpectationsisnotapracticalassumption. (cid:15) Actualprivatesectorexpectationsareunlikelytobe (cid:15) preciselyconsistentwithREE. Thatisstillallright... (cid:15) ... solongassmallexpectationaldeviationsfromRE (cid:15) dissipate,insteadofaccumulating.

Mainideas Whattheauthorsdo Interpretations Unstableisnotoptimal Otheranalyses Whyworry Learning and monetary policy Rationalexpectationsisacornerstoneofmodern (cid:15) macroeconomictheory. Rationalexpectationsequilibriumisattheheartofthe (cid:15) learningliterature. Butrationalexpectationsisnotapracticalassumption. (cid:15) Actualprivatesectorexpectationsareunlikelytobe (cid:15) preciselyconsistentwithREE. Thatisstillallright... (cid:15) ... solongassmallexpectationaldeviationsfromRE (cid:15) dissipate,insteadofaccumulating. Thatconditionisknownasexpectationalstability. (cid:15)

FailureoftheTaylorPrincipleisassociatedwith (cid:15) expectationalinstability. contemporaneousdataspecificationofthepolicyrule. (cid:15) decreasinggaininthelearningrule. (cid:15) InthesimplestNewKeynesianmodel,theTaylorPrinciple (cid:15) isassociatedwithE-stability. Thatresultassumes (cid:15) Comforting. (cid:15) Mainideas Whattheauthorsdo Interpretations Unstableisnotoptimal Otheranalyses Whyworry What the literature has shown REEmayormaynotbeexpectationallystableinNew (cid:15) Keynesianmodels.

contemporaneousdataspecificationofthepolicyrule. (cid:15) decreasinggaininthelearningrule. (cid:15) FailureoftheTaylorPrincipleisassociatedwith (cid:15) expectationalinstability. Thatresultassumes (cid:15) Comforting. (cid:15) Mainideas Whattheauthorsdo Interpretations Unstableisnotoptimal Otheranalyses Whyworry What the literature has shown REEmayormaynotbeexpectationallystableinNew (cid:15) Keynesianmodels. InthesimplestNewKeynesianmodel,theTaylorPrinciple (cid:15) isassociatedwithE-stability.

contemporaneousdataspecificationofthepolicyrule. (cid:15) decreasinggaininthelearningrule. (cid:15) Thatresultassumes (cid:15) Comforting. (cid:15) Mainideas Whattheauthorsdo Interpretations Unstableisnotoptimal Otheranalyses Whyworry What the literature has shown REEmayormaynotbeexpectationallystableinNew (cid:15) Keynesianmodels. InthesimplestNewKeynesianmodel,theTaylorPrinciple (cid:15) isassociatedwithE-stability. FailureoftheTaylorPrincipleisassociatedwith (cid:15) expectationalinstability.

contemporaneousdataspecificationofthepolicyrule. (cid:15) decreasinggaininthelearningrule. (cid:15) Comforting. (cid:15) Mainideas Whattheauthorsdo Interpretations Unstableisnotoptimal Otheranalyses Whyworry What the literature has shown REEmayormaynotbeexpectationallystableinNew (cid:15) Keynesianmodels. InthesimplestNewKeynesianmodel,theTaylorPrinciple (cid:15) isassociatedwithE-stability. FailureoftheTaylorPrincipleisassociatedwith (cid:15) expectationalinstability. Thatresultassumes (cid:15)

decreasinggaininthelearningrule. (cid:15) Comforting. (cid:15) Mainideas Whattheauthorsdo Interpretations Unstableisnotoptimal Otheranalyses Whyworry What the literature has shown REEmayormaynotbeexpectationallystableinNew (cid:15) Keynesianmodels. InthesimplestNewKeynesianmodel,theTaylorPrinciple (cid:15) isassociatedwithE-stability. FailureoftheTaylorPrincipleisassociatedwith (cid:15) expectationalinstability. Thatresultassumes (cid:15) contemporaneousdataspecificationofthepolicyrule. (cid:15)

Comforting. (cid:15) Mainideas Whattheauthorsdo Interpretations Unstableisnotoptimal Otheranalyses Whyworry What the literature has shown REEmayormaynotbeexpectationallystableinNew (cid:15) Keynesianmodels. InthesimplestNewKeynesianmodel,theTaylorPrinciple (cid:15) isassociatedwithE-stability. FailureoftheTaylorPrincipleisassociatedwith (cid:15) expectationalinstability. Thatresultassumes (cid:15) contemporaneousdataspecificationofthepolicyrule. (cid:15) decreasinggaininthelearningrule. (cid:15)

Mainideas Whattheauthorsdo Interpretations Unstableisnotoptimal Otheranalyses Whyworry What the literature has shown REEmayormaynotbeexpectationallystableinNew (cid:15) Keynesianmodels. InthesimplestNewKeynesianmodel,theTaylorPrinciple (cid:15) isassociatedwithE-stability. FailureoftheTaylorPrincipleisassociatedwith (cid:15) expectationalinstability. Thatresultassumes (cid:15) contemporaneousdataspecificationofthepolicyrule. (cid:15) decreasinggaininthelearningrule. (cid:15) Comforting. (cid:15)

Agentsaretryingtorobustlytrackthechanging (cid:15) environmentinwhichtheyoperate. Keeppolicymakerinformationinlinewithreality. (cid:15) Replacedecreasinggainwithconstantgain. (cid:15) ConcentrateonoperationalrulesinthesenseofMcCallum. (cid:15) Inparticular,contemporaneousvaluesofoutputand (cid:15) inflationarenotknownwhenpolicydecisionsaremade. Mainideas Whattheauthorsdo Interpretations Unstableisnotoptimal Otheranalyses Whyworry What the authors do Instrumentrulesplusvariousformsof“optimal”policy. (cid:15)

Keeppolicymakerinformationinlinewithreality. (cid:15) Agentsaretryingtorobustlytrackthechanging (cid:15) environmentinwhichtheyoperate. ConcentrateonoperationalrulesinthesenseofMcCallum. (cid:15) Inparticular,contemporaneousvaluesofoutputand (cid:15) inflationarenotknownwhenpolicydecisionsaremade. Mainideas Whattheauthorsdo Interpretations Unstableisnotoptimal Otheranalyses Whyworry What the authors do Instrumentrulesplusvariousformsof“optimal”policy. (cid:15) Replacedecreasinggainwithconstantgain. (cid:15)

Keeppolicymakerinformationinlinewithreality. (cid:15) ConcentrateonoperationalrulesinthesenseofMcCallum. (cid:15) Inparticular,contemporaneousvaluesofoutputand (cid:15) inflationarenotknownwhenpolicydecisionsaremade. Mainideas Whattheauthorsdo Interpretations Unstableisnotoptimal Otheranalyses Whyworry What the authors do Instrumentrulesplusvariousformsof“optimal”policy. (cid:15) Replacedecreasinggainwithconstantgain. (cid:15) Agentsaretryingtorobustlytrackthechanging (cid:15) environmentinwhichtheyoperate.

Keeppolicymakerinformationinlinewithreality. (cid:15) Inparticular,contemporaneousvaluesofoutputand (cid:15) inflationarenotknownwhenpolicydecisionsaremade. Mainideas Whattheauthorsdo Interpretations Unstableisnotoptimal Otheranalyses Whyworry What the authors do Instrumentrulesplusvariousformsof“optimal”policy. (cid:15) Replacedecreasinggainwithconstantgain. (cid:15) Agentsaretryingtorobustlytrackthechanging (cid:15) environmentinwhichtheyoperate. ConcentrateonoperationalrulesinthesenseofMcCallum. (cid:15)

Inparticular,contemporaneousvaluesofoutputand (cid:15) inflationarenotknownwhenpolicydecisionsaremade. Mainideas Whattheauthorsdo Interpretations Unstableisnotoptimal Otheranalyses Whyworry What the authors do Instrumentrulesplusvariousformsof“optimal”policy. (cid:15) Replacedecreasinggainwithconstantgain. (cid:15) Agentsaretryingtorobustlytrackthechanging (cid:15) environmentinwhichtheyoperate. ConcentrateonoperationalrulesinthesenseofMcCallum. (cid:15) Keeppolicymakerinformationinlinewithreality. (cid:15)

Mainideas Whattheauthorsdo Interpretations Unstableisnotoptimal Otheranalyses Whyworry What the authors do Instrumentrulesplusvariousformsof“optimal”policy. (cid:15) Replacedecreasinggainwithconstantgain. (cid:15) Agentsaretryingtorobustlytrackthechanging (cid:15) environmentinwhichtheyoperate. ConcentrateonoperationalrulesinthesenseofMcCallum. (cid:15) Keeppolicymakerinformationinlinewithreality. (cid:15) Inparticular,contemporaneousvaluesofoutputand (cid:15) inflationarenotknownwhenpolicydecisionsaremade.

... butmostproduceexpectationalinstabilityinthissetting. (cid:15) Policymakersfollowingtheserecommendedapproachesin (cid:15) thisenvironmentwouldbesurprisedtofindthatthe economydoesnotcoordinateontheintendedequilibrium. Toobtainexpectationalstability,usethe (cid:15) expectations-basedrulesofEvans-Honkapohja(2003, 2006). Mainideas Whattheauthorsdo Interpretations Unstableisnotoptimal Otheranalyses Whyworry Main findings Therearemanyrecommendedgoodor“optimal”policies (cid:15) forthismodel...

Policymakersfollowingtheserecommendedapproachesin (cid:15) thisenvironmentwouldbesurprisedtofindthatthe economydoesnotcoordinateontheintendedequilibrium. Toobtainexpectationalstability,usethe (cid:15) expectations-basedrulesofEvans-Honkapohja(2003, 2006). Mainideas Whattheauthorsdo Interpretations Unstableisnotoptimal Otheranalyses Whyworry Main findings Therearemanyrecommendedgoodor“optimal”policies (cid:15) forthismodel... ... butmostproduceexpectationalinstabilityinthissetting. (cid:15)

Toobtainexpectationalstability,usethe (cid:15) expectations-basedrulesofEvans-Honkapohja(2003, 2006). Mainideas Whattheauthorsdo Interpretations Unstableisnotoptimal Otheranalyses Whyworry Main findings Therearemanyrecommendedgoodor“optimal”policies (cid:15) forthismodel... ... butmostproduceexpectationalinstabilityinthissetting. (cid:15) Policymakersfollowingtheserecommendedapproachesin (cid:15) thisenvironmentwouldbesurprisedtofindthatthe economydoesnotcoordinateontheintendedequilibrium.

Mainideas Whattheauthorsdo Interpretations Unstableisnotoptimal Otheranalyses Whyworry Main findings Therearemanyrecommendedgoodor“optimal”policies (cid:15) forthismodel... ... butmostproduceexpectationalinstabilityinthissetting. (cid:15) Policymakersfollowingtheserecommendedapproachesin (cid:15) thisenvironmentwouldbesurprisedtofindthatthe economydoesnotcoordinateontheintendedequilibrium. Toobtainexpectationalstability,usethe (cid:15) expectations-basedrulesofEvans-Honkapohja(2003, 2006).

Otherwiseconvergenceoccurs. (cid:15) WearetalkingaboutlocallytotheREE. (cid:15) Ifpolicyisexecutedcorrectly... (cid:15) ... observerswouldseeonlytheREEvaluesofkeymacro (cid:15) variables. Learningwouldappeartobeunimportant,eventhoughit (cid:15) isenforcingtheREE. Theactuallearningdynamicswouldonlybeobservedin (cid:15) anunstablesitutation. Oneanalogy: off-equilibriumplayingames. (cid:15) Anotheranalogy: defaultpunishmentinmodelswith (cid:15) endogenousdebtconstraints. Mainideas Whattheauthorsdo Interpretations Unstableisnotoptimal Otheranalyses Whyworry How to think about E-stability Stabilityisnotacommonmodeofanalysisin (cid:15) macroeconomicsandmonetarypolicy.

Otherwiseconvergenceoccurs. (cid:15) WearetalkingaboutlocallytotheREE. (cid:15) Ifpolicyisexecutedcorrectly... (cid:15) ... observerswouldseeonlytheREEvaluesofkeymacro (cid:15) variables. Learningwouldappeartobeunimportant,eventhoughit (cid:15) isenforcingtheREE. Oneanalogy: off-equilibriumplayingames. (cid:15) Anotheranalogy: defaultpunishmentinmodelswith (cid:15) endogenousdebtconstraints. Mainideas Whattheauthorsdo Interpretations Unstableisnotoptimal Otheranalyses Whyworry How to think about E-stability Stabilityisnotacommonmodeofanalysisin (cid:15) macroeconomicsandmonetarypolicy. Theactuallearningdynamicswouldonlybeobservedin (cid:15) anunstablesitutation.

WearetalkingaboutlocallytotheREE. (cid:15) Ifpolicyisexecutedcorrectly... (cid:15) ... observerswouldseeonlytheREEvaluesofkeymacro (cid:15) variables. Learningwouldappeartobeunimportant,eventhoughit (cid:15) isenforcingtheREE. Oneanalogy: off-equilibriumplayingames. (cid:15) Anotheranalogy: defaultpunishmentinmodelswith (cid:15) endogenousdebtconstraints. Mainideas Whattheauthorsdo Interpretations Unstableisnotoptimal Otheranalyses Whyworry How to think about E-stability Stabilityisnotacommonmodeofanalysisin (cid:15) macroeconomicsandmonetarypolicy. Theactuallearningdynamicswouldonlybeobservedin (cid:15) anunstablesitutation. Otherwiseconvergenceoccurs. (cid:15)

Ifpolicyisexecutedcorrectly... (cid:15) ... observerswouldseeonlytheREEvaluesofkeymacro (cid:15) variables. Learningwouldappeartobeunimportant,eventhoughit (cid:15) isenforcingtheREE. Oneanalogy: off-equilibriumplayingames. (cid:15) Anotheranalogy: defaultpunishmentinmodelswith (cid:15) endogenousdebtconstraints. Mainideas Whattheauthorsdo Interpretations Unstableisnotoptimal Otheranalyses Whyworry How to think about E-stability Stabilityisnotacommonmodeofanalysisin (cid:15) macroeconomicsandmonetarypolicy. Theactuallearningdynamicswouldonlybeobservedin (cid:15) anunstablesitutation. Otherwiseconvergenceoccurs. (cid:15) WearetalkingaboutlocallytotheREE. (cid:15)

... observerswouldseeonlytheREEvaluesofkeymacro (cid:15) variables. Learningwouldappeartobeunimportant,eventhoughit (cid:15) isenforcingtheREE. Oneanalogy: off-equilibriumplayingames. (cid:15) Anotheranalogy: defaultpunishmentinmodelswith (cid:15) endogenousdebtconstraints. Mainideas Whattheauthorsdo Interpretations Unstableisnotoptimal Otheranalyses Whyworry How to think about E-stability Stabilityisnotacommonmodeofanalysisin (cid:15) macroeconomicsandmonetarypolicy. Theactuallearningdynamicswouldonlybeobservedin (cid:15) anunstablesitutation. Otherwiseconvergenceoccurs. (cid:15) WearetalkingaboutlocallytotheREE. (cid:15) Ifpolicyisexecutedcorrectly... (cid:15)

Learningwouldappeartobeunimportant,eventhoughit (cid:15) isenforcingtheREE. Oneanalogy: off-equilibriumplayingames. (cid:15) Anotheranalogy: defaultpunishmentinmodelswith (cid:15) endogenousdebtconstraints. Mainideas Whattheauthorsdo Interpretations Unstableisnotoptimal Otheranalyses Whyworry How to think about E-stability Stabilityisnotacommonmodeofanalysisin (cid:15) macroeconomicsandmonetarypolicy. Theactuallearningdynamicswouldonlybeobservedin (cid:15) anunstablesitutation. Otherwiseconvergenceoccurs. (cid:15) WearetalkingaboutlocallytotheREE. (cid:15) Ifpolicyisexecutedcorrectly... (cid:15) ... observerswouldseeonlytheREEvaluesofkeymacro (cid:15) variables.

Oneanalogy: off-equilibriumplayingames. (cid:15) Anotheranalogy: defaultpunishmentinmodelswith (cid:15) endogenousdebtconstraints. Mainideas Whattheauthorsdo Interpretations Unstableisnotoptimal Otheranalyses Whyworry How to think about E-stability Stabilityisnotacommonmodeofanalysisin (cid:15) macroeconomicsandmonetarypolicy. Theactuallearningdynamicswouldonlybeobservedin (cid:15) anunstablesitutation. Otherwiseconvergenceoccurs. (cid:15) WearetalkingaboutlocallytotheREE. (cid:15) Ifpolicyisexecutedcorrectly... (cid:15) ... observerswouldseeonlytheREEvaluesofkeymacro (cid:15) variables. Learningwouldappeartobeunimportant,eventhoughit (cid:15) isenforcingtheREE.

Anotheranalogy: defaultpunishmentinmodelswith (cid:15) endogenousdebtconstraints. Mainideas Whattheauthorsdo Interpretations Unstableisnotoptimal Otheranalyses Whyworry How to think about E-stability Stabilityisnotacommonmodeofanalysisin (cid:15) macroeconomicsandmonetarypolicy. Theactuallearningdynamicswouldonlybeobservedin (cid:15) anunstablesitutation. Otherwiseconvergenceoccurs. (cid:15) WearetalkingaboutlocallytotheREE. (cid:15) Ifpolicyisexecutedcorrectly... (cid:15) ... observerswouldseeonlytheREEvaluesofkeymacro (cid:15) variables. Learningwouldappeartobeunimportant,eventhoughit (cid:15) isenforcingtheREE. Oneanalogy: off-equilibriumplayingames. (cid:15)

Mainideas Whattheauthorsdo Interpretations Unstableisnotoptimal Otheranalyses Whyworry How to think about E-stability Stabilityisnotacommonmodeofanalysisin (cid:15) macroeconomicsandmonetarypolicy. Theactuallearningdynamicswouldonlybeobservedin (cid:15) anunstablesitutation. Otherwiseconvergenceoccurs. (cid:15) WearetalkingaboutlocallytotheREE. (cid:15) Ifpolicyisexecutedcorrectly... (cid:15) ... observerswouldseeonlytheREEvaluesofkeymacro (cid:15) variables. Learningwouldappeartobeunimportant,eventhoughit (cid:15) isenforcingtheREE. Oneanalogy: off-equilibriumplayingames. (cid:15) Anotheranalogy: defaultpunishmentinmodelswith (cid:15) endogenousdebtconstraints.

ConsiderthebreakdownofBrettonWoods. (cid:15) AndmaybeweshouldworryaboutSweden,aswewillsee (cid:15) below. Indeed,wewouldnotexpecttoseealotofthesecases... (cid:15) ... policymakerswouldhavetoabandontheirpolicies. (cid:15) ButconsiderFigure2inthepaper. (cid:15) InflationandtheoutputgapbehaveasiftheREEhasbeen (cid:15) attainedformanyperiods. Buteventually,expectationalerrorsaccumulateanddrive (cid:15) theeconomyawayfromthetargetedequilibrium. Doesithappen? (cid:15) Mainideas Whattheauthorsdo Interpretations Unstableisnotoptimal Otheranalyses Whyworry Expectational instability Somemightworrythattheunstablecaseis“notreally (cid:15) observed.”

ConsiderthebreakdownofBrettonWoods. (cid:15) AndmaybeweshouldworryaboutSweden,aswewillsee (cid:15) below. ... policymakerswouldhavetoabandontheirpolicies. (cid:15) ButconsiderFigure2inthepaper. (cid:15) InflationandtheoutputgapbehaveasiftheREEhasbeen (cid:15) attainedformanyperiods. Buteventually,expectationalerrorsaccumulateanddrive (cid:15) theeconomyawayfromthetargetedequilibrium. Doesithappen? (cid:15) Mainideas Whattheauthorsdo Interpretations Unstableisnotoptimal Otheranalyses Whyworry Expectational instability Somemightworrythattheunstablecaseis“notreally (cid:15) observed.” Indeed,wewouldnotexpecttoseealotofthesecases... (cid:15)

ConsiderthebreakdownofBrettonWoods. (cid:15) AndmaybeweshouldworryaboutSweden,aswewillsee (cid:15) below. ButconsiderFigure2inthepaper. (cid:15) InflationandtheoutputgapbehaveasiftheREEhasbeen (cid:15) attainedformanyperiods. Buteventually,expectationalerrorsaccumulateanddrive (cid:15) theeconomyawayfromthetargetedequilibrium. Doesithappen? (cid:15) Mainideas Whattheauthorsdo Interpretations Unstableisnotoptimal Otheranalyses Whyworry Expectational instability Somemightworrythattheunstablecaseis“notreally (cid:15) observed.” Indeed,wewouldnotexpecttoseealotofthesecases... (cid:15) ... policymakerswouldhavetoabandontheirpolicies. (cid:15)

ConsiderthebreakdownofBrettonWoods. (cid:15) AndmaybeweshouldworryaboutSweden,aswewillsee (cid:15) below. InflationandtheoutputgapbehaveasiftheREEhasbeen (cid:15) attainedformanyperiods. Buteventually,expectationalerrorsaccumulateanddrive (cid:15) theeconomyawayfromthetargetedequilibrium. Doesithappen? (cid:15) Mainideas Whattheauthorsdo Interpretations Unstableisnotoptimal Otheranalyses Whyworry Expectational instability Somemightworrythattheunstablecaseis“notreally (cid:15) observed.” Indeed,wewouldnotexpecttoseealotofthesecases... (cid:15) ... policymakerswouldhavetoabandontheirpolicies. (cid:15) ButconsiderFigure2inthepaper. (cid:15)

ConsiderthebreakdownofBrettonWoods. (cid:15) AndmaybeweshouldworryaboutSweden,aswewillsee (cid:15) below. Buteventually,expectationalerrorsaccumulateanddrive (cid:15) theeconomyawayfromthetargetedequilibrium. Doesithappen? (cid:15) Mainideas Whattheauthorsdo Interpretations Unstableisnotoptimal Otheranalyses Whyworry Expectational instability Somemightworrythattheunstablecaseis“notreally (cid:15) observed.” Indeed,wewouldnotexpecttoseealotofthesecases... (cid:15) ... policymakerswouldhavetoabandontheirpolicies. (cid:15) ButconsiderFigure2inthepaper. (cid:15) InflationandtheoutputgapbehaveasiftheREEhasbeen (cid:15) attainedformanyperiods.

ConsiderthebreakdownofBrettonWoods. (cid:15) AndmaybeweshouldworryaboutSweden,aswewillsee (cid:15) below. Doesithappen? (cid:15) Mainideas Whattheauthorsdo Interpretations Unstableisnotoptimal Otheranalyses Whyworry Expectational instability Somemightworrythattheunstablecaseis“notreally (cid:15) observed.” Indeed,wewouldnotexpecttoseealotofthesecases... (cid:15) ... policymakerswouldhavetoabandontheirpolicies. (cid:15) ButconsiderFigure2inthepaper. (cid:15) InflationandtheoutputgapbehaveasiftheREEhasbeen (cid:15) attainedformanyperiods. Buteventually,expectationalerrorsaccumulateanddrive (cid:15) theeconomyawayfromthetargetedequilibrium.

ConsiderthebreakdownofBrettonWoods. (cid:15) AndmaybeweshouldworryaboutSweden,aswewillsee (cid:15) below. Mainideas Whattheauthorsdo Interpretations Unstableisnotoptimal Otheranalyses Whyworry Expectational instability Somemightworrythattheunstablecaseis“notreally (cid:15) observed.” Indeed,wewouldnotexpecttoseealotofthesecases... (cid:15) ... policymakerswouldhavetoabandontheirpolicies. (cid:15) ButconsiderFigure2inthepaper. (cid:15) InflationandtheoutputgapbehaveasiftheREEhasbeen (cid:15) attainedformanyperiods. Buteventually,expectationalerrorsaccumulateanddrive (cid:15) theeconomyawayfromthetargetedequilibrium. Doesithappen? (cid:15)

AndmaybeweshouldworryaboutSweden,aswewillsee (cid:15) below. Mainideas Whattheauthorsdo Interpretations Unstableisnotoptimal Otheranalyses Whyworry Expectational instability Somemightworrythattheunstablecaseis“notreally (cid:15) observed.” Indeed,wewouldnotexpecttoseealotofthesecases... (cid:15) ... policymakerswouldhavetoabandontheirpolicies. (cid:15) ButconsiderFigure2inthepaper. (cid:15) InflationandtheoutputgapbehaveasiftheREEhasbeen (cid:15) attainedformanyperiods. Buteventually,expectationalerrorsaccumulateanddrive (cid:15) theeconomyawayfromthetargetedequilibrium. Doesithappen? (cid:15) ConsiderthebreakdownofBrettonWoods. (cid:15)

Mainideas Whattheauthorsdo Interpretations Unstableisnotoptimal Otheranalyses Whyworry Expectational instability Somemightworrythattheunstablecaseis“notreally (cid:15) observed.” Indeed,wewouldnotexpecttoseealotofthesecases... (cid:15) ... policymakerswouldhavetoabandontheirpolicies. (cid:15) ButconsiderFigure2inthepaper. (cid:15) InflationandtheoutputgapbehaveasiftheREEhasbeen (cid:15) attainedformanyperiods. Buteventually,expectationalerrorsaccumulateanddrive (cid:15) theeconomyawayfromthetargetedequilibrium. Doesithappen? (cid:15) ConsiderthebreakdownofBrettonWoods. (cid:15) AndmaybeweshouldworryaboutSweden,aswewillsee (cid:15) below.

Taylorrulefairsbetterthansomeotherrecommended (cid:15) rulesstudiedlaterinthepaper. Instabilityrequiresthecombinationofoperationalrules (cid:15) withconstantgainlearning. Mainideas Whattheauthorsdo Interpretations Unstableisnotoptimal Otheranalyses Whyworry Taylor-type rules Somepotentialforinstabilityinthecalibratedcase. (cid:15)

Instabilityrequiresthecombinationofoperationalrules (cid:15) withconstantgainlearning. Mainideas Whattheauthorsdo Interpretations Unstableisnotoptimal Otheranalyses Whyworry Taylor-type rules Somepotentialforinstabilityinthecalibratedcase. (cid:15) Taylorrulefairsbetterthansomeotherrecommended (cid:15) rulesstudiedlaterinthepaper.

Mainideas Whattheauthorsdo Interpretations Unstableisnotoptimal Otheranalyses Whyworry Taylor-type rules Somepotentialforinstabilityinthecalibratedcase. (cid:15) Taylorrulefairsbetterthansomeotherrecommended (cid:15) rulesstudiedlaterinthepaper. Instabilityrequiresthecombinationofoperationalrules (cid:15) withconstantgainlearning.

Resultsforcommitmentcasesimilar. (cid:15) Itrepresentsoptimalpolicyunderdiscretionifthe (cid:15) policymakerhasapreferenceforinterestratesmoothing. Theruleis (cid:15) ϕλ ϕα i = π + x x . (1) t t t α α i i Producesdeterminacyifα sufficientlylarge. i (cid:15) Producesexpectationalinstabilityifα sufficientlysmall. i (cid:15) SimilarresultsforDuffy/Xiaoundercommitment. (cid:15) Worrisome. (cid:15) Theexpectations-basedapproachofEvansand (cid:15) Honkapohjasolvesthisproblemandprovidesrobust expectationalstability. Mainideas Whattheauthorsdo Interpretations Unstableisnotoptimal Otheranalyses Whyworry Duffy and Xiao TheDuffy/XiaoruleisaTaylor-typerule. (cid:15)

Resultsforcommitmentcasesimilar. (cid:15) Theruleis (cid:15) ϕλ ϕα i = π + x x . (1) t t t α α i i Producesdeterminacyifα sufficientlylarge. i (cid:15) Producesexpectationalinstabilityifα sufficientlysmall. i (cid:15) SimilarresultsforDuffy/Xiaoundercommitment. (cid:15) Worrisome. (cid:15) Theexpectations-basedapproachofEvansand (cid:15) Honkapohjasolvesthisproblemandprovidesrobust expectationalstability. Mainideas Whattheauthorsdo Interpretations Unstableisnotoptimal Otheranalyses Whyworry Duffy and Xiao TheDuffy/XiaoruleisaTaylor-typerule. (cid:15) Itrepresentsoptimalpolicyunderdiscretionifthe (cid:15) policymakerhasapreferenceforinterestratesmoothing.

Theruleis (cid:15) ϕλ ϕα i = π + x x . (1) t t t α α i i Producesdeterminacyifα sufficientlylarge. i (cid:15) Producesexpectationalinstabilityifα sufficientlysmall. i (cid:15) SimilarresultsforDuffy/Xiaoundercommitment. (cid:15) Worrisome. (cid:15) Theexpectations-basedapproachofEvansand (cid:15) Honkapohjasolvesthisproblemandprovidesrobust expectationalstability. Mainideas Whattheauthorsdo Interpretations Unstableisnotoptimal Otheranalyses Whyworry Duffy and Xiao TheDuffy/XiaoruleisaTaylor-typerule. (cid:15) Itrepresentsoptimalpolicyunderdiscretionifthe (cid:15) policymakerhasapreferenceforinterestratesmoothing. Resultsforcommitmentcasesimilar. (cid:15)

Producesdeterminacyifα sufficientlylarge. i (cid:15) Producesexpectationalinstabilityifα sufficientlysmall. i (cid:15) SimilarresultsforDuffy/Xiaoundercommitment. (cid:15) Worrisome. (cid:15) Theexpectations-basedapproachofEvansand (cid:15) Honkapohjasolvesthisproblemandprovidesrobust expectationalstability. Mainideas Whattheauthorsdo Interpretations Unstableisnotoptimal Otheranalyses Whyworry Duffy and Xiao TheDuffy/XiaoruleisaTaylor-typerule. (cid:15) Itrepresentsoptimalpolicyunderdiscretionifthe (cid:15) policymakerhasapreferenceforinterestratesmoothing. Resultsforcommitmentcasesimilar. (cid:15) Theruleis (cid:15) ϕλ ϕα i = π + x x . (1) t t t α α i i

Producesexpectationalinstabilityifα sufficientlysmall. i (cid:15) SimilarresultsforDuffy/Xiaoundercommitment. (cid:15) Worrisome. (cid:15) Theexpectations-basedapproachofEvansand (cid:15) Honkapohjasolvesthisproblemandprovidesrobust expectationalstability. Mainideas Whattheauthorsdo Interpretations Unstableisnotoptimal Otheranalyses Whyworry Duffy and Xiao TheDuffy/XiaoruleisaTaylor-typerule. (cid:15) Itrepresentsoptimalpolicyunderdiscretionifthe (cid:15) policymakerhasapreferenceforinterestratesmoothing. Resultsforcommitmentcasesimilar. (cid:15) Theruleis (cid:15) ϕλ ϕα i = π + x x . (1) t t t α α i i Producesdeterminacyifα sufficientlylarge. i (cid:15)

SimilarresultsforDuffy/Xiaoundercommitment. (cid:15) Worrisome. (cid:15) Theexpectations-basedapproachofEvansand (cid:15) Honkapohjasolvesthisproblemandprovidesrobust expectationalstability. Mainideas Whattheauthorsdo Interpretations Unstableisnotoptimal Otheranalyses Whyworry Duffy and Xiao TheDuffy/XiaoruleisaTaylor-typerule. (cid:15) Itrepresentsoptimalpolicyunderdiscretionifthe (cid:15) policymakerhasapreferenceforinterestratesmoothing. Resultsforcommitmentcasesimilar. (cid:15) Theruleis (cid:15) ϕλ ϕα i = π + x x . (1) t t t α α i i Producesdeterminacyifα sufficientlylarge. i (cid:15) Producesexpectationalinstabilityifα sufficientlysmall. i (cid:15)

Worrisome. (cid:15) Theexpectations-basedapproachofEvansand (cid:15) Honkapohjasolvesthisproblemandprovidesrobust expectationalstability. Mainideas Whattheauthorsdo Interpretations Unstableisnotoptimal Otheranalyses Whyworry Duffy and Xiao TheDuffy/XiaoruleisaTaylor-typerule. (cid:15) Itrepresentsoptimalpolicyunderdiscretionifthe (cid:15) policymakerhasapreferenceforinterestratesmoothing. Resultsforcommitmentcasesimilar. (cid:15) Theruleis (cid:15) ϕλ ϕα i = π + x x . (1) t t t α α i i Producesdeterminacyifα sufficientlylarge. i (cid:15) Producesexpectationalinstabilityifα sufficientlysmall. i (cid:15) SimilarresultsforDuffy/Xiaoundercommitment. (cid:15)

Theexpectations-basedapproachofEvansand (cid:15) Honkapohjasolvesthisproblemandprovidesrobust expectationalstability. Mainideas Whattheauthorsdo Interpretations Unstableisnotoptimal Otheranalyses Whyworry Duffy and Xiao TheDuffy/XiaoruleisaTaylor-typerule. (cid:15) Itrepresentsoptimalpolicyunderdiscretionifthe (cid:15) policymakerhasapreferenceforinterestratesmoothing. Resultsforcommitmentcasesimilar. (cid:15) Theruleis (cid:15) ϕλ ϕα i = π + x x . (1) t t t α α i i Producesdeterminacyifα sufficientlylarge. i (cid:15) Producesexpectationalinstabilityifα sufficientlysmall. i (cid:15) SimilarresultsforDuffy/Xiaoundercommitment. (cid:15) Worrisome. (cid:15)

Mainideas Whattheauthorsdo Interpretations Unstableisnotoptimal Otheranalyses Whyworry Duffy and Xiao TheDuffy/XiaoruleisaTaylor-typerule. (cid:15) Itrepresentsoptimalpolicyunderdiscretionifthe (cid:15) policymakerhasapreferenceforinterestratesmoothing. Resultsforcommitmentcasesimilar. (cid:15) Theruleis (cid:15) ϕλ ϕα i = π + x x . (1) t t t α α i i Producesdeterminacyifα sufficientlylarge. i (cid:15) Producesexpectationalinstabilityifα sufficientlysmall. i (cid:15) SimilarresultsforDuffy/Xiaoundercommitment. (cid:15) Worrisome. (cid:15) Theexpectations-basedapproachofEvansand (cid:15) Honkapohjasolvesthisproblemandprovidesrobust expectationalstability.

SvenssonandWoodford(2005)recommendedrule (cid:15) equation(18)inthepaper. Operationalversionscanbeassociatedwithinstability (cid:15) underlearningforreasonablegainparameters. WorrisomeforSweden? Figure2? (cid:15) Thereisnothingoptimalaboutinstability. (cid:15) Mainideas Whattheauthorsdo Interpretations Unstableisnotoptimal Otheranalyses Whyworry Commitment Nointerestratesmoothing,timelessperspective,goalisto (cid:15) implementtheFOC.

Operationalversionscanbeassociatedwithinstability (cid:15) underlearningforreasonablegainparameters. WorrisomeforSweden? Figure2? (cid:15) Thereisnothingoptimalaboutinstability. (cid:15) Mainideas Whattheauthorsdo Interpretations Unstableisnotoptimal Otheranalyses Whyworry Commitment Nointerestratesmoothing,timelessperspective,goalisto (cid:15) implementtheFOC. SvenssonandWoodford(2005)recommendedrule (cid:15) equation(18)inthepaper.

WorrisomeforSweden? Figure2? (cid:15) Thereisnothingoptimalaboutinstability. (cid:15) Mainideas Whattheauthorsdo Interpretations Unstableisnotoptimal Otheranalyses Whyworry Commitment Nointerestratesmoothing,timelessperspective,goalisto (cid:15) implementtheFOC. SvenssonandWoodford(2005)recommendedrule (cid:15) equation(18)inthepaper. Operationalversionscanbeassociatedwithinstability (cid:15) underlearningforreasonablegainparameters.

Thereisnothingoptimalaboutinstability. (cid:15) Mainideas Whattheauthorsdo Interpretations Unstableisnotoptimal Otheranalyses Whyworry Commitment Nointerestratesmoothing,timelessperspective,goalisto (cid:15) implementtheFOC. SvenssonandWoodford(2005)recommendedrule (cid:15) equation(18)inthepaper. Operationalversionscanbeassociatedwithinstability (cid:15) underlearningforreasonablegainparameters. WorrisomeforSweden? Figure2? (cid:15)

Mainideas Whattheauthorsdo Interpretations Unstableisnotoptimal Otheranalyses Whyworry Commitment Nointerestratesmoothing,timelessperspective,goalisto (cid:15) implementtheFOC. SvenssonandWoodford(2005)recommendedrule (cid:15) equation(18)inthepaper. Operationalversionscanbeassociatedwithinstability (cid:15) underlearningforreasonablegainparameters. WorrisomeforSweden? Figure2? (cid:15) Thereisnothingoptimalaboutinstability. (cid:15)

Avoidscommittingtoaparticularrecursivealgorithmto (cid:15) describelearning. But,expectationalstabilitystillplaysaroleinthatanalysis. (cid:15) Mainideas Whattheauthorsdo Interpretations Unstableisnotoptimal Otheranalyses Whyworry Woodford NRE Woodford(2008)hasconsideredanalternativeapproachto (cid:15) checkingtherobustnessofpolicytothepossibilitythat expectationsmaynotbefullyrational.

But,expectationalstabilitystillplaysaroleinthatanalysis. (cid:15) Mainideas Whattheauthorsdo Interpretations Unstableisnotoptimal Otheranalyses Whyworry Woodford NRE Woodford(2008)hasconsideredanalternativeapproachto (cid:15) checkingtherobustnessofpolicytothepossibilitythat expectationsmaynotbefullyrational. Avoidscommittingtoaparticularrecursivealgorithmto (cid:15) describelearning.

Mainideas Whattheauthorsdo Interpretations Unstableisnotoptimal Otheranalyses Whyworry Woodford NRE Woodford(2008)hasconsideredanalternativeapproachto (cid:15) checkingtherobustnessofpolicytothepossibilitythat expectationsmaynotbefullyrational. Avoidscommittingtoaparticularrecursivealgorithmto (cid:15) describelearning. But,expectationalstabilitystillplaysaroleinthatanalysis. (cid:15)

“extra”termintheactuallawofmotion. (cid:15) expectationalstabilityconditionsunchanged. (cid:15) BullardandSuda(2008). (cid:15) Standardrecursivelearningexercise,butreplaceclassical (cid:15) econometricianswithBayesianeconometricians. Mainresultsstillhold: (cid:15) Stabilitystillanissue. (cid:15) Mainideas Whattheauthorsdo Interpretations Unstableisnotoptimal Otheranalyses Whyworry Bayesian approaches ManyfeelthatrecursivelearningshouldbeBayesian. (cid:15)

“extra”termintheactuallawofmotion. (cid:15) expectationalstabilityconditionsunchanged. (cid:15) Standardrecursivelearningexercise,butreplaceclassical (cid:15) econometricianswithBayesianeconometricians. Mainresultsstillhold: (cid:15) Stabilitystillanissue. (cid:15) Mainideas Whattheauthorsdo Interpretations Unstableisnotoptimal Otheranalyses Whyworry Bayesian approaches ManyfeelthatrecursivelearningshouldbeBayesian. (cid:15) BullardandSuda(2008). (cid:15)

“extra”termintheactuallawofmotion. (cid:15) expectationalstabilityconditionsunchanged. (cid:15) Mainresultsstillhold: (cid:15) Stabilitystillanissue. (cid:15) Mainideas Whattheauthorsdo Interpretations Unstableisnotoptimal Otheranalyses Whyworry Bayesian approaches ManyfeelthatrecursivelearningshouldbeBayesian. (cid:15) BullardandSuda(2008). (cid:15) Standardrecursivelearningexercise,butreplaceclassical (cid:15) econometricianswithBayesianeconometricians.

“extra”termintheactuallawofmotion. (cid:15) expectationalstabilityconditionsunchanged. (cid:15) Stabilitystillanissue. (cid:15) Mainideas Whattheauthorsdo Interpretations Unstableisnotoptimal Otheranalyses Whyworry Bayesian approaches ManyfeelthatrecursivelearningshouldbeBayesian. (cid:15) BullardandSuda(2008). (cid:15) Standardrecursivelearningexercise,butreplaceclassical (cid:15) econometricianswithBayesianeconometricians. Mainresultsstillhold: (cid:15)

expectationalstabilityconditionsunchanged. (cid:15) Stabilitystillanissue. (cid:15) Mainideas Whattheauthorsdo Interpretations Unstableisnotoptimal Otheranalyses Whyworry Bayesian approaches ManyfeelthatrecursivelearningshouldbeBayesian. (cid:15) BullardandSuda(2008). (cid:15) Standardrecursivelearningexercise,butreplaceclassical (cid:15) econometricianswithBayesianeconometricians. Mainresultsstillhold: (cid:15) “extra”termintheactuallawofmotion. (cid:15)

Stabilitystillanissue. (cid:15) Mainideas Whattheauthorsdo Interpretations Unstableisnotoptimal Otheranalyses Whyworry Bayesian approaches ManyfeelthatrecursivelearningshouldbeBayesian. (cid:15) BullardandSuda(2008). (cid:15) Standardrecursivelearningexercise,butreplaceclassical (cid:15) econometricianswithBayesianeconometricians. Mainresultsstillhold: (cid:15) “extra”termintheactuallawofmotion. (cid:15) expectationalstabilityconditionsunchanged. (cid:15)

Mainideas Whattheauthorsdo Interpretations Unstableisnotoptimal Otheranalyses Whyworry Bayesian approaches ManyfeelthatrecursivelearningshouldbeBayesian. (cid:15) BullardandSuda(2008). (cid:15) Standardrecursivelearningexercise,butreplaceclassical (cid:15) econometricianswithBayesianeconometricians. Mainresultsstillhold: (cid:15) “extra”termintheactuallawofmotion. (cid:15) expectationalstabilityconditionsunchanged. (cid:15) Stabilitystillanissue. (cid:15)

butitisnotthatclearhowtomapthisideaintoreality. (cid:15) Measurementissues. (cid:15) Potentialgames. (cid:15) Mainideas Whattheauthorsdo Interpretations Unstableisnotoptimal Otheranalyses Whyworry Responding to expectations Policymakersseemtodothisattimes... (cid:15)

Measurementissues. (cid:15) Potentialgames. (cid:15) Mainideas Whattheauthorsdo Interpretations Unstableisnotoptimal Otheranalyses Whyworry Responding to expectations Policymakersseemtodothisattimes... (cid:15) butitisnotthatclearhowtomapthisideaintoreality. (cid:15)

Potentialgames. (cid:15) Mainideas Whattheauthorsdo Interpretations Unstableisnotoptimal Otheranalyses Whyworry Responding to expectations Policymakersseemtodothisattimes... (cid:15) butitisnotthatclearhowtomapthisideaintoreality. (cid:15) Measurementissues. (cid:15)

Mainideas Whattheauthorsdo Interpretations Unstableisnotoptimal Otheranalyses Whyworry Responding to expectations Policymakersseemtodothisattimes... (cid:15) butitisnotthatclearhowtomapthisideaintoreality. (cid:15) Measurementissues. (cid:15) Potentialgames. (cid:15)

ThankstotheBankofJapanforinvitingmetoparticipate. (cid:15) Iappreciatetheattentiontostabilityissues,whichIthink (cid:15) areinsufficientlyanalyzedinmacroeconomics. Instabilitycanproducethe“bigticketlosses”that (cid:15) policymakersreallyworryabout. Mainideas Whattheauthorsdo Interpretations Unstableisnotoptimal Otheranalyses Whyworry Final thoughts Verynicepaperinaveryniceconference. (cid:15)

Iappreciatetheattentiontostabilityissues,whichIthink (cid:15) areinsufficientlyanalyzedinmacroeconomics. Instabilitycanproducethe“bigticketlosses”that (cid:15) policymakersreallyworryabout. Mainideas Whattheauthorsdo Interpretations Unstableisnotoptimal Otheranalyses Whyworry Final thoughts Verynicepaperinaveryniceconference. (cid:15) ThankstotheBankofJapanforinvitingmetoparticipate. (cid:15)

Instabilitycanproducethe“bigticketlosses”that (cid:15) policymakersreallyworryabout. Mainideas Whattheauthorsdo Interpretations Unstableisnotoptimal Otheranalyses Whyworry Final thoughts Verynicepaperinaveryniceconference. (cid:15) ThankstotheBankofJapanforinvitingmetoparticipate. (cid:15) Iappreciatetheattentiontostabilityissues,whichIthink (cid:15) areinsufficientlyanalyzedinmacroeconomics.

Mainideas Whattheauthorsdo Interpretations Unstableisnotoptimal Otheranalyses Whyworry Final thoughts Verynicepaperinaveryniceconference. (cid:15) ThankstotheBankofJapanforinvitingmetoparticipate. (cid:15) Iappreciatetheattentiontostabilityissues,whichIthink (cid:15) areinsufficientlyanalyzedinmacroeconomics. Instabilitycanproducethe“bigticketlosses”that (cid:15) policymakersreallyworryabout.