Assessing Central Bank Credibility in Emerging Markets

Assessing Central Bank Credibility in Emerging Markets

Giancarlo Perasso
May 23, 2023 3 minutes

Monetary policy in emerging markets (EMs) underwent drastic change at the beginning of the 21st century. Following the Asian (1997) and Russian (1998) financial crises, many central banks in EMs stopped targeting exchange rates and adopted an inflation targeting framework. Broadly speaking, that switch has been successful at reducing inflation and bolstering central banks’ credibility.

But the recent acceleration and persistence of inflation in many EMs raises new challenges for central bankers and investors. In our view, the current high inflation is the result of (mostly) exogenous shocks but also of domestic polices, like a slow retrenchment of pandemic-era fiscal expansion.

 

The recent rise in inflation might, then, put these central banks’ hard-earned credibility at risk. In this note, we look at the experience of five major EM central banks that have adopted an inflation-target framework (table 1).

 

Table 1

Inflation Targets of Select EM Central Banks

This table shows five major EM central banks that have adopted an inflation-target framework (Brazil, Chile, Hungary, Mexico, South Africa).
zoom_in
close
This table shows five major EM central banks that have adopted an inflation-target framework (Brazil, Chile, Hungary, Mexico, South Africa).

Source: IMF, as of April 2023

Anchoring inflation expectations is a key component of the inflation target framework. When expectations are de-anchored, the risk of runaway inflation increases. This, in turn, might require stricter monetary tightening than if the central bank had acted earlier. For that reason, anchoring expectations when inflation surges is an important pre-emptive action for central banks that have adopted an inflation-targeting framework. 

To analyze the behavior of the five central banks in our sample, we first assess how successful they have been in anchoring inflation expectations. We do so by computing a simple “credibility index”, as described in the Appendix. That index is a normalized measure of the difference between inflation expectations and the central bank’s inflation target, where 1 = full credibility of the target.

 

Table 2

Summary Statistics on Inflation, Inflation Expectations and Central Bank Credibility

Table shows each country's inflation target (%), average inflation, average inflation expectations, and credibility index.
zoom_in
close
Table shows each country's inflation target (%), average inflation, average inflation expectations, and credibility index.

Source: PGIM Fixed Income, Haver, as of April 2023.

Table 2 shows that inflation expectations generally converge towards the inflation target even when actual inflation is outside of the target band. Hungary and Brazil are the only countries where average inflation expectations remain above target – albeit slightly – even after our arbitrarily-imposed 10 year “credibility building” period.

Second, we look at descriptive statistics to find regularities in central banks’ policies and outcomes.

Figures 1 to 5 show that, all else equal, the faster and more determined a central bank is in rising rates, the faster inflation expectations drop.

A single large hike may not be enough (as in Hungary’s recent experience) and other sizeable hikes may be needed (as in Chile and Mexico) to lower inflation expectations. Once inflation expectations are anchored within the target band, prudent rate cuts can follow (Chile currently, Chile and South Africa in 2008, Brazil in 2002).

 

Figures 1-5: Changes in Inflation Expectations and Policy Responses

Figure 1

Brazil - Policy Rate Response

This bar chart shows the month-over-month change in policy rate in Brazil from 2001-2022 with a line graph overlayed showing 1-year expectation.
zoom_in
close
This bar chart shows the month-over-month change in policy rate in Brazil from 2001-2022 with a line graph overlayed showing 1-year expectation.

Figure 2

Chile - Policy Rate Response

This bar chart shows the month-over-month change in policy rate in Chile from 2001-2022 with a line graph overlayed showing 1-year expectation.
zoom_in
close
This bar chart shows the month-over-month change in policy rate in Chile from 2001-2022 with a line graph overlayed showing 1-year expectation.

Figure 3

Hungary - Policy Rate Response

This bar chart shows the month-over-month change in policy rate in Hungary from 2001-2022 with a line graph overlayed showing 1-year expectation.
zoom_in
close
This bar chart shows the month-over-month change in policy rate in Hungary from 2001-2022 with a line graph overlayed showing 1-year expectation.

Figure 4

Mexico - Policy Rate Response

This bar chart shows the month-over-month change in policy rate in Mexico from 2001-2022 with a line graph overlayed showing 1-year expectation.
zoom_in
close
This bar chart shows the month-over-month change in policy rate in Mexico from 2001-2022 with a line graph overlayed showing 1-year expectation.

Figure 5

South Africa - Policy Rate Response

This bar chart shows the month-over-month change in policy rate in South Africa from 2001-2022 with a line graph overlayed showing 1-year expectation.
zoom_in
close
This bar chart shows the month-over-month change in policy rate in South Africa from 2001-2022 with a line graph overlayed showing 1-year expectation.

Source: PGIM Fixed Income, Haver, as of April 2023.

APPENDIX: Our central bank credibility index

For each survey, we calculate the difference between 1-year inflation expectations (or 2-year expectations, depending on the inflation target set by the respective central bank) and the central bank’s inflation target. Since these countries have a target band, we calculate the difference between expectations and the mid-value of that band. We then normalize that time series of differences to fall between 0 and 1, using the normalization formula. The index at time “t” then becomes1:

Index(t) = 1- dP(t)/{max[Ex(0…T) – P*] – min[Ep(0…T) – P*]}

Where:

P* = inflation target

Ep(t) = expected inflation at time “t”

dP(t) = Ep(t) -P*

0< t < T

And a value of 1 indicates full credibility.

Sample = 0…T

As an example, Figure A.1 shows the index for South Africa.

 

Figure A.1

South Africa - Central Bank Credibility Index

Line graph showing credibility between 2001-2022. zoom_in
close
Line graph showing credibility between 2001-2022.

 

1  If inflation expectations fall below the target, we assign a value of zero (i.e., target met) to that observation.


Giancarlo Perasso
Giancarlo Perasso
Lead Economist Africa and Former Soviet Union Public & Private Fixed Income

Subscribe to the Bond Blog

Sign up to receive our weekly email alert notifying you when new blog entries are published.

close

*Required Fields

*
*
*
*
*
Consent
Strategies of interest
Topics of Interest