Too Many Cooks? Committees in Monetary Policy Helge Berger International Monetary Fund & Free University Berlin Volker Nitsch Darmstadt University & KOF, ETH Zurich
Dec 29, 2015
Too Many Cooks? Committees in Monetary Policy
Helge BergerInternational Monetary Fund &
Free University Berlin
Volker NitschDarmstadt University &
KOF, ETH Zurich
Motivation 1 — The Idea
• The quality of monetary policy decisions is, among other things, a function of the design of the decision-making body
• Design dimensions include size and (in federal set-ups) centralization, and representation
• Size of considerable recent theoretical & practical interest
• Example: debate on ECB reform (GC max = 21 > 5-10 avg)
• First attempt at providing simple theory & stylized facts:What arguments may matter for board/MPC size?What arguments do matter? Is board/MPC size part of a larger CB design problem?
Motivation 2 — MPC Size and Monetary Policy
• The number of board members is objective and measurable.
• Quality of decision-making process is also influenced by:Chairman:
• Determining the agenda• Meeting procedures
Internal or external members:• Internal members’ career depend on the chairman?• External members and conflict of interest?
Government representatives:• Voting / no voting rights?
Determinants of MPC Size
• MPC design in general, and decision on size in particular, involve trade-off between costs and benefits (Sibert 2006, Berger 2006, Gerling et al. 2005, Fujiki 2005)
• Benefits Information: larger MPCs better at collecting and
processing information under uncertainty(Gerlach-Kristen 2006, Blinder/Morgan 2005, Lombardelli et al. 2005, Berk/Bierut 2004)
Moderation: larger MPC moderate extreme views(Riboni/Ruge-Murcia 2006, Sibert 2006, Blinder 1998)
Political Economy: larger MPCs may help insulate individuals from political pressure(Blinder 2006)
Determinants of MPC Size 2
• CostsDecision-making: increasing in membership size (Berger 2006)
Example:
Information: decreasing returns & free-riding (Sibert 2006)
403530252015105
4000
3000
2000
1000
0
n
e=1
e=5
C = e [n (n -1)] / 2 + e n
30h27
14h18
4h9
C(n, e=5 min)n
30h27
14h18
4h9
C(n, e=5 min)n
Determinants of MPC Size 3
• Costs (continued)Example: Bundesbank reform
German Bundesbank asserts that its 1992 MPC reform triggered by German unification helped preventing an increase in MPC size which “would have greatly complicated that body’s decision-making processes” (Deutsche Bundesbank 1992, p.50)
Example: ECB reformECB admitted fear that anticipated increase in the number of national central bank governors attending euro area MPC meetings after EMU enlargement “will not necessarily make deliberations easier…”. (ECB 2003, p.83)
Determinants of MPC Size 4
• Country characteristics: Goodfriend (2005): “efficient size of a policy committee might vary across countries” Information-related benefits are likely to be increasing in
country size & heterogeneityDecision-making costs could be lower in environments
used to more democratic decision processesOther economic, political, and institutional factors,
e.g. income levels, fiscal behavior, monetary policy regime, other features of the central bank set-up may matter, too
• Use this variation to identify determinants of MPC size
How to think about the underlying design process?
Literature
• Costs and benefits of MPC size (applied theoretical, institutional or experimental perspectives)(Blinder 1998, Gerlach-Kristen 2006, Blinder and Morgan (2005), Lombardelli et al. 2005, Sibert 2006)
• Empirics of MPC design(Berger et al. 2008, Erhart and Vasquez-Paz 2007)
• Effects of institutional features of central banking on monetary policy(Cukierman et al. 1992, de Haan and Kooi 2000, Göhlmann and Vaubel 2007)
• Effects of board size (and other board characteristics) on corporate performance (Yermack 1996, Hermalin and Weisbach 2003, de Andres and Vallelado 2008)
Data
• New data set on MPCs around the world: 33 countries1950-2006annualde jure/de factovarious other features (e.g., composition)
• Compute other measures in addition to MPC size (e.g., governor/membership turnover rate)
Country Sample
Argentina Iceland Pakistan
Australia Ireland Portugal
Austria Israel Singapore
Belgium Italy Spain
Botswana Japan Sweden
Brazil Korea Switzerland
Canada Malaysia Thailand
Denmark Mauritius Trinidad & Tobago
Finland Netherlands Turkey
France New Zealand U.K.
Germany Norway U.S.A.
Data on Board Size
23
21
222930303130292827
22
251816
15
1310
11
7
0
2
4
6
8
10
1950 1955 1960 1965 1970 1975 1980 1985 1990 1995 2000 2005
Data on Board Size
0
2
4
6
8
10
12
1950 1955 1960 1965 1970 1975 1980 1985 1990 1995 2000 2005
5 countries
9 countries
11 countries 4 countries
Episodes of Major Changes in Board Size
78
91
01
1#
me
mb
er
1978 1979 1980 1981 1982 1983
Argentina
05
10
15
20
# m
em
ber
1992 1993 1994 1995 1996 1997
Brazil
68
10
12
14
# m
em
ber
1990 1991 1992 1993 1994 1995
France0
24
68
10
# m
em
ber
1986 1987 1988 1989 1990 1991
New Zealand
78
91
0#
me
mb
er
1967 1968 1969 1970 1971 1972
Turkey
45
67
89
# m
em
ber
1995 1996 1997 1998 1999 2000
Japan
56
78
9#
me
mb
er
1995 1996 1997 1998 1999 2000
Singapore
02
46
81
0#
me
mb
er1994 1995 1996 1997 1998 1999
United Kingdom
Methodology
• Aim: Measure the “success” of monetary policy
• Event-study approach
• Fatas, Mihov, and Rose (JMCB, 2007):
Controls: Openness, Budget deficit, BusCycle, Per capita income, GDP
Does board size matter?
Inflationit+1 = + 1 Targetit + 2 Successit + Yi + i
Event-Study Analysis: Inflation
010
020
030
040
0%
-3 -2 -1 0 1 2 3
Fewer Members
010
2030
40%
-3 -2 -1 0 1 2 3
More Members5
67
8%
-3 -2 -1 0 1 2 3
w/o Argentina & BrazilFewer Members
02
46
8%
-3 -2 -1 0 1 2 3
w/o Argentina & BrazilMore Members
Event-Study Analysis: Real GDP Growth
01
23
4%
-3 -2 -1 0 1 2 3
Fewer Members
01
23
45
%
-3 -2 -1 0 1 2 3
More Members0
12
34
5%
-3 -2 -1 0 1 2 3
w/o Argentina & BrazilFewer Members
01
23
45
%
-3 -2 -1 0 1 2 3
w/o Argentina & BrazilMore Members
Sample Baseline Full (with Argentina, Brazil)
De Facto Membership -0.28** -2.78** 5.27* -22.88*
De Facto Membership Squared 0.13** 1.42*
De Jure Quant. Monetary Target -12.93** -14.86** -117.45** -117.18**
Quant. Monetary Success -2.92** -1.85# 22.82# 33.59*
Turnover Rate Governor 8.58* 13.07** 278.30** 311.48**
Openness (% GDP) 0.001 0.003 -0.07 -0.09
Budget Balance (% GDP) -0.88* -0.83* -1.68 -0.77
Business Cycle (Growth –Avg Growth)
-0.11 -0.19 -9.22# -9.24#
Log Real GDP per capita -5.49** -6.40** -31.91** -31.44**
Log Real GDP -1.11# -1.98* 0.03 -9.83#
Observations 733 733 760 760
Adj. R-squared 0.22 0.25 0.18 0.22
Optimal Board Size
0
2
4
6
8
10
12
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20
0
20
40
60
80
100
120
140
160
180
200
Benchmark (LHS) W/ Argentina & Brazil (RHS)
Other Dependent Variables
Dependent variable Inflation variability Output growth
De Facto Membership -0.09** -0.53** -0.017* 0.08*
De Facto Membership Squared 0.02** -0.005**
…
normalized inflation: NormInflit = Inflationit / (1 + Inflationit)
Other MPC Size Measures
De Facto Size (Second Quartile)
-5.38*
De Facto Size (Third Quartile)
-8.25**
De Facto Size (Fourth Quartile)
-2.46**
De Jure Membership -0.44** -2.20*
De Jure Membership Squared
0.10*
De Facto Voting Membership
-0.29** -1.89*
De Facto Voting Membership Squared
0.10*
Turnover Rate Membership
-3.53**
Other MPC Characteristics
De Facto Membership -2.75*(1.09)
-3.06**(1.05)
-2.80**(1.03)
-3.15**(1.08)
De Facto Membership Squared 0.14*(0.06)
0.15**(0.05)
0.15**(0.05)
0.19**(0.05)
Government Representatives -0.38(0.54)
-0.78(0.67)
Regional Representatives -0.16(0.21)
-1.02**(0.38)
Industry Representatives -0.19#(1.06)
-0.75**(0.15)
MPC Size and Quantitative Targets
De Facto Membership -2.71**(0.95)
-2.69**(0.95)
-2.69**(0.94)
De Facto Membership Squared 0.12**(0.05)
0.12**(0.04)
0.12**(0.04)
Dummy for Large Committees (>10 Members) -9.16* (4.58)
2.11 (3.20)
-8.76# (4.63)
De Jure Quant. Monetary Target -15.49** (4.38)
-14.80** (4.26)
-15.71** (4.53)
De Jure Quant. Monetary Target × Large Committee
12.05* (4.99)
12.99* (5.46)
Quant. Monetary Success -2.00* (0.98)
-2.15# (1.11)
-1.74 (1.13)
Quant. Monetary Success × Large Committee 0.64(1.87)
-1.31 (1.68)
Conclusions
• Evidence that MPC size affects monetary policy outcomes (such as inflation) important feature of CB designU-shaped relationship; minimum level of inflation is reached
at about 8 to 10 members
• Other MPC characteristics also of importance
• A lot is left to be doneStrengthen and extend empirical work: panel estimation
techniques; extend sample Explore interaction with other CB characteristics
Appendix
Size Facts
Distribution (%) Obs.
3 4-6 7-9 10-12 13
Policy 4 28 47 11 10 50
Implementation 4 10 10 40 40 95
Number of Members in Governing Bodies 2003
Source: Lybek & Morris (2005)
Bank (Federal) Council Size
Bank (Central) Council Size
Bundesbank pre-1957 10 Australia 9
Bundesbank 1998 17 Canada 7
Fed 12 New Zealand 1
ECB 2001 18 Sweden 6
ECB (EMU27 w/o, w/ ref.) 33, 21 UK 9
Number of Members in Governing Bodies 2003