1 Benefits from U.S. Monetary Policy Experimentation in the Days of Samuelson and Solow and Lucas Timothy Cogley Riccardo Colacito Thomas J. Sargent Ramon.

1

“Benefits from U.S. Monetary Policy Experimentation in the Days of

Samuelson and Solow and Lucas”

Timothy CogleyRiccardo Colacito

Thomas J. Sargent

Ramon Marimon

Universitat Pompeu Fabra & CREi

Barcelona, April 1, 2005

Discussion of

2

25+ years later

3

Sims dixit (p.24):

“…if there is uncertainty about the structure of the economy, then even with a fixed policy objective function, widely understood, the form of the dependence of policy on observed history will shift over time as more is learned about (or opinions shift about) the structure of the economy. One could continually re-estimate the structure and, each period, re-announce an explicit relation of policy variables to history.”

4

Tom Sargent’s work on

“The Conquest of

the U.S. Inflation”• Princeton, 1999

• & with Tim Cogley Review of

Economic Dynamics (forth.)follows this dictum

5

e.g., in Cogley and Sargent3 competing models are‘continuously reestimated’ andrelative weights assigned for a(Bayesian) composite model

to ‘explain’ data and Fed choices...

6

but Sims adds:

“However it is simpler to announce the stable objective function once and then each period solve only for this period’s policy variables values instead of computing a complete policy reaction function. This is done by making conditional projections from the best existing reduced form model, and picking the best looking projected future time path.” (WP p.24)

7

Cogley, Colacito & Sargent (2005) follows this second dictum

With two models that ‘fit historical data equally well’ (1948-63):

8

recognizing what Sims said:

“Empirical macro-economists sometimes express frustration at the limited amount of information in economic time series, and it does not infrequently turn out that models reflecting rather different behavioral hypothesis fit the data equally well.” (p.26)

9

In a world that differs from ‘Conquest’ and others

• There is no uncertainty over the specific models, only about (the weight on SS)

• Therefore, models are not updated, only and the corresponding decision

• Private agents and monetary authority share same beliefs

• There is no Kydland & Prescott’ time-inconsistency• Expected inflation in estimations: regressing inflation on ‘a

constant along with one lag of inflation and unemployment’ (so much for Lucas model!)

...in the ‘Days of Samuelson and Solow and Lucas’ (1963?)

10

...in the old ‘Days of Samuelson and Solow and Lucas’

11

and the ‘old fashioned’, pre VARs OLS, provides clues,

12

More than the Lucas model being in doubt, the ‘old fashioned’ eclectic-nested model

already ‘hints’ that

“a little of doubt about the Lucas model is enough to make the central bank acquiesce to Keynesian prescriptions”

CC&S ‘show’ this beautifully using Bayesian experimenters and ‘anticipated utility’ policy makers

13

“a little of doubt...”

14

but the paper is, mostly, about ‘the value of experimentation’, in a world where...

• The CB must ‘learn the true model’• Bayesian updating converges to the true model,

whatever this is• Experimentation is ‘one sided’: either ‘do nothing’

(Lucas) or move along the ‘Phillips curve trade-off’ (Samuelson-Solow)

• CB cares a lot about unemployment• CB experimentation does not “enhance the

difficulty of [private agents] learning by experience” (M. Friedman, A Program for Monetary Stability, 1960)

15

as a result,• Experimentation has little value• Particularly if beliefs about SS are high (natural

experimentation)• While, if Lucas is right, Keynesian policy makers

discover their mistake quickly Did they? Was Lucas wrong? Where they too cautious?

SS is right

L is right

16

and the paper is also about ‘misspecification and caution’

• Distrusting the ‘prior over models’ makes a cautious CB give more weight to the ‘do nothing guy’: Lucas

meet the T2 map!(CC&S +Hansen)

17

Distrusting the

‘models specifications’

increases variances

and makes a cautious

CB give more weight

to ‘active policies,’

closer to the SS model

meet the T1 map!

Alas! Fear + Fear No Fear

18

Are they falling into Sims’ criticism?“While it ought to appear unreasonable, this pattern of

giving careful attention to minor sources of uncertainty, while ignoring major sources by making dogmatic assumptions, is not uncommon in economics” (‘Pitfalls...’ AER Papers& Proc. 2001)

I do not think so,I think it is more that, after having gone through the interesting exercise, one can see that perhaps ‘the Days of Samuelson and Solow and Lucas’ may not be the most interesting models to study ‘Experimentation and Caution in Economics’

And this shows that we have gone a long way...

19

20

In 25 Years, THE VAR has had major impact in new generations

as in

and thousands moreBut...

21

as in 25 Years, THE WALL has had major impact in new generations too

22

23

end of my discussion!