Explicit inflation targets and central bank independence: friends or foes?

Abstract

The paper studies the relationship between two institutional innovations in monetary policy of the past few decades: central bank independence (CBI) and explicit inflation targeting (EIT). The aim is to make inferences about the optimal institutional design of monetary policy, and the right sequencing of policy reform. Our reduced-form model unifies several approaches in the literature, and offers three novel institutional findings (that we square with existing empirical evidence). First, instrument-CBI is a complement to EIT, whereas goal-CBI acts as a strategic substitute for EIT in ensuring low inflation and policy credibility. Second, out of these two ‘commitment technologies,’ EIT is shown to be socially superior to goal-CBI. Third and controversially, countries that first implement goal-CBI are then less likely to adopt the desirable EIT regime. This is because independent central bankers may have less need to do so (their independence partly substitutes for EIT), as well as less willingness to do so (due to a higher degree of accountability associated with a transparently legislated target). Our analysis therefore implies that developing and emerging market countries should go down the New Zealand route—legislate EIT together with instrument-CBI, but stay clear of goal-CBI. Unfortunately, many transition countries have followed the opposite Fed/Bundesbank route, which we show may have adverse welfare consequences through several channels.

Appendices

Appendix 1: Proof of Proposition 1

As overriding is not observable the public forms an expectation of it, $R_{t}^{e}$, and sets inflation expectations rationally based on (9)

$$ (\pi_{t}^{e} )^{ * } = \bar{\pi } + \frac{{\lambda^{g} }}{2\omega }\left[ {1 - CBI(1 - R_{t}^{e} )} \right] $$

(19)

Moving forward let us examine the overriding choice. Under M _t  = 1 the public knows the true $g_{t}^{T}$, i.e. has full information about the opponents’ preferences and can thus perfectly anticipate overriding, $(R_{t}^{e} )^{ * } = R_{t}$. Using this constraint with (2), (9) and (19) yields

$$ \frac{{\partial u^{{g^{W} }} }}{\partial R} = - \frac{{\lambda^{g} CBI}}{2\omega }\left[ {1 - CBI(1 - R_{t} )} \right] < 0 $$

(20)

This implies that $R_{t}^{ * } = 0$, $\forall g^{T}$. The public therefore sets $(R_{t}^{e} )^{ * } = 0$ in (19), leading to $x_{t}^{ * } = 0$ from (1). $\square$

Appendix 2: CBI index (Sousa 2002)

 

Criteria	Points
Personal independence
1. Appointment of the central bank board members	1.00
2. Mandate duration of more than half of the central bank board members	1.00
3. Policymaker (or other fiscal branches representatives) participation at central bank meetings, where monetary decisions are taken	1.00
Political independence
4. Ultimate responsibility and authority on monetary policy decisions	1.00
5. Price stability	1.00
6. Banking supervision	1.00
7. Monetary policy instruments	1.00
Economic and financial independence
8. Policymaker financing	1.00
9. Ownership of the central bank’s (equity) capital	1.00

Appendix 3: AC index (Sousa 2002)

Criteria and methodology of this index is adopted from de Haan et al. (1999). We only use the ‘final responsibility’ component that we believe best proxies the degree of EIT.

 

Criteria	Points
Final responsibility
1. Is the central bank subject of monitoring by Parliament?	1.00
2. Has the policymaker (or Parliament) the right to give instruction?	1.00
3. Is there some kind of review in the procedure to apply the override mechanism?	1.00
4. Has central bank possibility for an appeal in case of an instruction?	1.00
5. Can the central bank law be changed by a simple majority in Parliament?	1.00
6. Is past performance a ground for dismissal of a central bank governor?	1.00

Appendix 4: Goal-CBI index (Briault et al. 1997)

 

1. Whether the statutes of the central bank make it independent of the government	2.00
2. Whether more than half the appointments to the central bank board are made independently of the government	2.00
3. Whether there are government officials on the board	2.00
4. Whether the central bank does in practice set its own goals (for example, monetary or inflation targets).	2.00

Appendix 5: TR index (Eijffinger and Geraats 2006)

 

1. Political
(a) Formal objectives	1.00
(b) Quantitative targets	1.00
(c) Institutional arrangements	1.00
2. Economic
(a) Economic data	1.00
(b) Policy models	1.00
(c) Central bank forecasts	1.00
3. Procedural
(a) Explicit strategy	1.00
(b) Minutes	1.00
(c) Voting records	1.00
4. Policy
(a) Prompt announcement	1.00
(b) Policy explanation	1.00
(c) Policy inclination	1.00
5. Operational
(a) Control errors	1.00
(b) Transmission disturbances	1.00
(c) Evaluation policy outcome	1.00

Appendix 6: Evaluation table

 

Index		CBI Sousa (2002)a				AC Sousa (2002)a	TR Eijffinger and Geraats (2006) for 1998						Goal-CBI Briault et al. (1997)
Country		Personal	Political	Economic and financial	Totalb	Final responsibility	Political	Economic	Procedural	Policy	Operational	Total	Total
1	Argentina	1.25	2.83	1.00	5.08	2
2	Australia	0.50	2.16	0.00	2.66	5	3	1	1	1.5	1.5	8	2
3	Austria	1.66	3.16	1.00	5.82	1
4	Belgium	1.75	1.50	0.00	3.25	4
5	Canada	0.50	1.83	0.00	2.33	4	3	2.5	1	2	2	10.5	2
6	Chile	2.00	1.83	1.00	4.83	3
7	Czech Republic	1.58	3.16	1.00	5.74	2
8	Denmark	2.16	2.83	0.00	4.99	2
9	EMU—ECB	2.50	3.66	1.00	7.16	1
10	England	1.00	2.66	0.00	3.66	4	3	1.5	3	1.5	2	11	1
11	Finland	2.50	3.66	1.00	7.16	2
12	France	1.50	3.16	1.00	5.66	2
13	Germany	1.50	3.16	1.00	5.66	2	1c	1c	1c	2c	1c	6c	6
14	Greece	1.91	3.16	1.00	6.07	1
15	Hungary	1.91	3.66	1.00	6.57	2
16	Iceland	1.75	3.33	0.00	5.08	4
17	Ireland	1.00	3.16	1.00	5.16	2
18	Italy	2.16	3.16	1.00	6.32	1
19	Japan	0.75	3.66	0.00	4.41	3	1.5	1	2	1.5	2	8	4
20	Korea	0.75	2.16	0.00	2.91	4
21	Luxemburg	1.25	3.16	1.00	5.41	2
22	Mexico	1.83	2.33	0.00	4.16	2
23	Netherlands	2.41	3.16	0.00	5.57	2
24	New Zealand	1.83	2.16	1.00	4.99	4	3	2.5	3	1	1	10.5	3
25	Norway	1.58	1.83	0.00	3.41	5
26	Poland	1.25	2.16	0.00	3.41	3
27	Portugal	1.50	3.16	1.00	5.66	1
28	Slovakia	1.00	3.50	1.00	5.50	1
29	Spain	0.75	3.16	1.00	4.91	2
30	Sweden	2.75	3.16	1.00	6.91	1	2	1.5	2	1.5	2	9	2
31	Switzerland	2.08	3.33	1.00	6.41	2	1	1	1	2	1	6	6
32	Turkey	1.66	2.83	0.00	4.49	3
33	USA	2.00	1.83	0.00	3.83	2	1	2.5	2	1.5	1.5	8.5	4

1. ^aAssessment is based on situation in January 2002
2. ^bExcludes aspect 9 due to missing observations
3. ^cEijffinger and Geraats (2006) do not provide a value for the Bundesbank for the year 1998. The literature agrees that Bundesbank’s practices were very similar to the Swiss National Bank’s (see e.g. Bernanke et al. 1999) and we indeed find all the two countries’ criteria scores to coincide. This is further supported by the fact that in Sousa (2002) and Fry et al. (2000) the scores of the two central banks are equal (in the latter index in terms of goal-TR)