What effects is EMU having on the euro area and its member countries?

16 17 June 2005, Frankfurt am Main

Session V: Inflation persistence and inflation differentials

Discussant: William Dickens Brookings Institution, Washington D.C.

Comments on Price Setting and Inflation Persistence: did EMU Matter?

European Central Bank 17 June, 2005 Frankfurter Hof Hotel

What I Want to Do
Heap praise on the authors and the
organizers of the IPN. Go over results of Angeloni et. al. Express some concerns about the results and their interpretation. Present some preliminary results from the International Wage Flexibility Project (a sister project of IPN which has also been generously supported by the ECB).

This is Hard Work

Perhaps no one in this room knows better than I
do how much work goes into a project like the IPN. Until you have tried to coordinate 13 teams of researchers in 12 countries dealing with more than a dozen unique sets of data and tried to administer a common protocol and make sense of the results So my hat is off to the authors. Been there, done that, its not easy!

Angeloni et. al. Present Two Types of Evidence

Micro data on frequency and magnitude of
price changes from surveys of consumer and producer prices Regression estimates of inflation persistence (sum of coefficients in a regression of inflation on many lags) using product level inflation data.

Did EMU Matter? Major Findings

PRICE CHANGES: No. Other than an abrupt
and limited increase in the frequency of price changes around the time of cash conversion there is no evidence of any change in the frequency or magnitude of price changes. INFLATION PERSISTENCE: Some evidence that inflation persistence declined in mid 90s, but same thing happened in US and likely reflects change in nature of inflationary expectations (which is likely due in some sense to EMU).

Why Do We Care About Frequency of Price Changes?

What I would most like to see is reference to a
model of firm behavior that would allow me to judge whether the estimated frequency is high or low relative to some policy relevant benchmark. Ideally what would be reported and tested for change would be parameters such as implied cost of adjusting and/or not adjusting prices. But, at the very least it would be good if measures of frequency of change were adjusted for inflation rates.

Adjust Change Frequencies for Inflation?

We might expect the frequency of price changes
to decline considerably with declining inflation. What I think we care about (and without an explicit model Im not sure) is whether EMU (and the single market more broadly) affected the costs of changing, or not changing, prices. The fact that there has been no change in the frequency of price reductions in countries that have had significant drops in their rates of inflation may indicated an increase in price flexibility.

Why Do We Care About Persistence?

Persistence measures are tied to new
Keynesian Phillips curve and as such are directly connected with policy concerns (how quickly will shocks abate and how aggressive must policy be to stabilize inflation and output). But link is tenuous due to nature of estimates (reduced form auto-regressions rather than estimates of models). Thus we dont know how persistent persistence is (could be structural or ephemeral ).

We May Not Believe Model Estimates

but at least they give us some breakdown
of the source of persistence. (Is it due to past inflation affecting current inflation, the way expectations are formed, or auto-correlation in real or nominal innovations?) I would be remiss as a discussant if I didnt take this opportunity to mention my favorite model of this phenomena Akerlof, Dickens and Perry (2000).

ADP 2000 and Persistence

Our model predicts the decline in persistence
as inflation falls as the result of people increasingly ignoring expectations in price setting with lower inflation. We find that once we take account of this the parameters of our Phillips Curve for US data are quite stable. Our model suggests that an inflation target of about 3% minimizes unemployment for the US (model doesnt fit well for European data).

The International Wage Flexibility Project

What is the IWFP?

13 Country study of wage inflation
supported by the ECB and directed by Erica Groshen (NY Fed.) and me using micro data on individual and occupational wages analyzed by teams in each country familiar with the data broader than analyzing wage rigidity (sand and grease), but that is the part that Im going to talk about today

Country Teams

Austria Belgium Denmark Finland France Germany Italy

Norway Portugal Sweden Switzerland United Kingdom United States

Disclaimers
Preliminary results! Mainly based on
European Community Household Panel data which is quite noisy. My views only (not the views of the ECB, Brookings, or any other members of the IWFP project team).

Questions
Could DNWR be an explanation for Europes
persistent unemployment problems or While ADP model fits very well for US and Canadian data it fits very poorly and gives nonsensical results when applied to European data. Could presence of real rigidity caused by centralized wage setting institutions be the reason for this? Some studies done in Europe (notably UK) show much less Downward Nominal Wage Rigidity than US and Canadian data, so is DNWR a problem in Europe? Might problem be downward real rigidity instead?

How Could We Answer Questions?

Initially we were unsure about how to get
at presence of different types of rigidity. At first meeting some very interesting results emerged examining wage change histograms for job stayers (all results are for job stayers).

1986, Inflation: 0% .2 .15 .1 .05 0 1990, Inflation: 5.3% .2 .15 .1

1987, Inflation: 1.9%

1988, Inflation: 1.9%

1989, Inflation: 5%

1991, Inflation: 5.2%

1992, Inflation: 3.4%

1993, Inflation: 2.5%

Fraction

.05 0 1994, Inflation: .4% .2 .15 .1 .05 0 1998, Inflation: -0.2% .2 .15 .1 .05 0 -.25 -.1 0 .1 .25 -.25 -.1 0 .1 .25 1999, Inflation: 1.7% -.25 -.1 0 .1 .25 -.25 -.1 0 .1 1995, Inflation: 1.9% 1996, Inflation: .8% 1997, Inflation: .4%

The Distribution of Nominal Wage Changes, Switzerland 1986 - 199

Does Low Inflation Eliminate Downward Wage Rigi

Belgium
400,000 350,000 300,000 N and % 250,000 200,000 150,000 100,000 50,000 0

Growth Rate 1978-1979

0.16 0.14 0.12 0.1 0.08 0.06 0.04 0.02 0

[ -2

[ -1

[ -1

[ -1

-1 0[ [ -8 , -7 [ [ -5 , -4 [ [ -2 , -1 [ [1 ,2 [ [4 ,5 [ [7 ,8 [1 [ 0, 11 [1 [ 3, 14 [1 [ 6, 17 [1 [ 9, 20 [

9[

6[ -1

-1

0,

-1 4,

7,

1,

3[

Growth Rate

Fraction

N and %
.15 .1

100,000

150,000

200,000

250,000

300,000

50,000

.05

-15

-10

-5

Growth Rate 1981-1982

1984

Growth Rate

UK
0

10

15

20

25

30

[ -2 0 , -1 9 [ [ -1 8 , -1 7 [ [ -1 6 , -1 5 [ [ -1 4 , -1 3 [ [ -1 2 , -1 1 [ [ -1 0 , -9 [ [ -8 , -7 [ [ -6 , -5 [ [ -4 , -3 [ [ -2 , -1 [ [0,1 [ [2,3 [ [4,5 [ [6,7 [ [8,9 [ [10 ,11 [ [12 ,13 [ [14 ,15 [ [16 ,17 [ [18 ,19 [ 0.1 0.02 0.04 0.06 0.08 0.12

35

10

Can we use wage change histograms to diagnose nature and extent of wage rigidity?

Wage Change Distribution

(For workers with same wage facing same minimum wage)
Observatuions Swept No Nominal Wage Change Observations Swept Up to Zero By Downward Nominal Wage Rigidity and Menu Costs Wage Reduced to Minimum Down to Zero by Menu Costs Observations Swept Up to Rate of Price Inflation By Downward Real Wage Rigidity No Real Wage Change

Change in Log Wage

Frequency 0.26

-0.22

-0.19

-0.16

-0.13

-0.10

-0.07

-0.04

-0.01

0.02

0.05

0.08

0.11

0.14

0.17

0.20

0.23

11

 If people make mistakes reporting their wages then

we see wage changes where there are none. If we compute wage=income/hours we see wage changes due to overtime, bonuses, or mistakes in reporting hours. If we use social security data we have similar problems since most countries dont have data that allow us to accurately identify base wage. All evidence suggests that for most data sets frequency and extent of errors make this a very serious problem (evidence suggests that in many data sets most reported wage cuts are actually errors of these sorts).

Big Problem is Measurement Error

New Approach:
Use information in correlation of changes between years
Abowd and Card (1989) suggest that wages have
two components: permanent changes transient (one period) changes (which result in negative serial correlation of wage changes) New method identifies transient changes as errors and uses auto-covariance and frequency of sign switching in changes to identify error rate and error variance. This information is used to identify semi-nonparametric mixed method of moments estimate of true wage distribution.

12

Rigidity Measures
Estimated true wage change distributions are
Downward nominal wage rigidity Symmetric nominal wage rigidity (menu costs) Downward real wage rigidity

then used as data for GMM estimate of a simple model of three types of rigidity

Measure is fraction of notional wage cuts (or

small changes) that dont happen due to rigidity (so could be independent of rate of inflation). No measures of insensitivity to fundamentals (another concept of rigidity).

Validating New Error Correction Method

Results applying new method to US largely fit with
those of other studies (a very high degree or downward nominal rigidity). Gottschalk method identifies true wages and measurement error in quarterly data. True wage changes have autoauto-correlation of zero (our key assumption). Finnish data have very few errors and procedure makes very few corrections Portuguese have good and bad data. Procedure doesnt correct good data and makes distribution of bad data resemble the good data.

13

US PSID Error Correction Model

1971
0.05.1.15.2 0 .05 .1 .15

1976

-4

-3

-2

-1

10

11

12

13

14

15

16

17

-4 18

-3 19

-2

-1

10

11

12

13

14

15

16

17

18

19

observed distribution constrained distribution

estimated true wage distribu tion observed distribution constrained distribution

estimated true wage distribution

1981
0 .05 .1 .15 0 .1 .2 .3

1992

-4

-3

-2

-1

10

11

12

13

14

15

16

17

18 -4 19 -3

-2

-1

10

11

12

13

14

15

16

17

18

19

observed distribution constrained distribution

estimated true wage distribution

observed distribution constrained distribution

estimated true wage distribution

Finland Service Sector

1991
0 .05 .1.15 .2 0 .2 .4 .6 .8

1993

-4

-3

-2

-1

10

11

12

13

14

15

16

17

18 -4

19 -3

-2

-1

10

11

12

13

14

15

16

17

18

19

observed distribution constrained distribution

estimated true wage distribution

observed distribution constrained distribution

estimated true wage distribution

1998
0 .1 .2 .3 .4
.4 .6

2001

-4

-3

-2

-1

10

11

12

13

14

15

16

0
17 18

.2

-4

19

-3

-2

-1

10

11

12

13

14

15

16

17

18

19

observed distribution constrained distribution

estimated true wage distribution

observed distribution constrained distribution

estimated true wage distribution

14

Portugese Data
Base wage+other labor earnings -1994
0.12 0.12

Base wage - 1994

0.1

0.1

0.08

0.08

0.06

0.06

0.04

0.04

0.02

0.02

0 -25 -23 -21 -19 -17 -15 -13 -11 -9 -7 -5 -3 -1 1 3 5 7 9 11 13 15 17 19 21 23 25 27 29 31 33 35 37 39 41 43 45 47 49

0 -25 -23 -21 -19 -17 -15 -13 -11 -9 -7 -5 -3 -1 1 3 5 7 9 11 13 15 17 19 21 23 25 27 29 31 33 35 37 39 41 43 45 47 49

-0.02

-0.02

Base wage+other labor earnings-1992 Base wage-1992

0.15 0.15

0.13

0.13

0.11

0.11

0.09

0.09

0.07

0.07

0.05

0.05

0.03

0.03

0.01

0.01

-25 -0.01

-22

-19

-16

-13

-10

-7

-4

-1

11

14

17

20

23

26

29

32

35

38

41

44

47

50 -0.01

-25

-22

-19

-16

-13

-10

-7

-4

-1

11

14

17

20

23

26

29

32

35

38

41

44

47

50

Real and Nominal Rigidity in 12 Countries

From ECHP and PSID

1.1 0.9 Fraction Affected 0.7 0.5 0.3 0.1

Ne
Real

th Ita ly er la n A u ds st ria G UK er m an Ire y la F i nd nla n Fr d an Be ce lg iu m

Nominal

-0.1

nm De

re

k ec e US

ar

15

Real and Nominal Rigidity in 10 EMU Countries

0.7 0.6 0.5 R igidity 0.4 0.3 0.2 0.1 0 1995 1996 1997 1998 1999 2000 2001
Real Nominal

Dispersion of Wage Rigidity in the EMU

0.38 0.36 0.34 0.32 0.3 0.28 0.26 0.24 0.22 0.2 1995 1996 1997 1998 1999 2000 2001

Rigidity

SD Real Rigidity

SD Nominal Rigidity

16

Conclusions
Too early to conclude much, but Much of Europe is different from the US showing
few signs of DNWR and substantial evidence of DRWR. Real and nominal rigidity seem to be negatively related and there are both high and low real rigidity countries in EMU Preliminary evidence suggests that wage setting institutions (centralized collective bargaining) are important determinants of regime.

17