## 1. Bank Transparency Index: South Africa and its Peers

## Source details

**Canonical URL:** [1. Bank Transparency Index: South Africa and its Peers](https://www.imf.org/-/media/files/publications/wp/2018/wp18173.pdf)

## Other formats

- [Markdown version](/-/media/files/publications/wp/2018/wp18173.pdf.md)
- [Structured JSON version](/-/media/files/publications/wp/2018/wp18173.pdf.json)

---

### Introduction and research question
- Examines why exchange rate pass-through (ERPT) to consumer inflation declined in South Africa since adoption of inflation targeting in the early 2000s.
- Hypothesis: improving monetary policy credibility explains the declining ERPT (following literature such as Taylor (2000)).

### Measurement of credibility, ERPT, and data
- Data: quarterly South Africa, 1980Q1–2015Q2; time-varying ERPT estimated with a 10-year rolling-window regression (equations (10) and (11)).
  - Extracted ERPT: decline from 50 percent in 1998 to 20 percent in 2002; thereafter roughly stable between 20 percent and 30 percent.
- Forecast disagreement: Reuters survey available from 2002Q2.
  - Number of respondents: about 13 at the beginning, maximum of 25 respondents in September 2014.
- Credibility proxy construction:
  - Convert Reuters’ quarterly-expressed forecasts using equation (1).
  - Disagreement σ_t^π: standard deviation or interquartile range across forecasters.
  - Credibility index δ_t = 1 / σ_t^π (equation (2)).
- Permanent shock to inflation h_t^π estimated with a UCSV framework (bounded trend between 3 and 6 percent).
  - Bounded trend innovation: ε_t^τ ~ TTN(a−τ_{t−1}, b−τ_{t−1}; σ_τ^2).

### Determinants of monetary policy credibility — empirical findings
- Correlations and co-movements:
  - Two disagreement measures strongly correlated: coefficient of correlation = 0.90.
  - Forecast disagreement co-moves with inflation; high disagreement during 2002Q3–2003Q3 and 2007Q4–2009Q1.
  - Deviation of realized inflation from midpoint 4.5 percent has substantially declined over time.
  - Forecast disagreement closely related to estimated inflation volatility.
  - Forecast disagreement is countercyclical: rises when output gap is negative, falls when output gap is positive.
- Expectations formation (equation (12)): π_t^e = ρ π_{t−1} + (1−ρ) π^* + ε_t, with π^* set at 6 percent.
  - Sample-split estimates (coefficients reported as: Lagged Inflation; Target (6); AR(1)):
    - Pre-2009 (Table 1, N = 32):
      - Total: 0.275***; 0.725***; 0.771***.
      - Analyst: 0.098**; 0.902***; 0.927***.
      - Business: 0.306***; 0.694***; 0.780***.
      - Trade Union: 0.452***; 0.548***; 0.594***.
    - Post-2009 (Table 2, N = 33):
      - Total: -0.024; 1.024***; 0.732***.
      - Analyst: 0.066*; 0.934***; 0.947***.
      - Business: -0.087; 1.087***; 0.828***.
      - Trade Union: -0.098; 1.098***; 0.619***.
  - Interpretation:
    - Pre-2009 agents somewhat backward-looking (lagged inflation coefficients significant).
    - Post-2009 agents largely forward-looking and anchored at 6 percent (lagged inflation not significantly different from zero; Wald test does not reject coefficient of target being one).
- FMOLS cointegrating regressions of equation (3) — determinants of δ_t (Table 3, N = 48 each):
  - Column (1):
    - Permanent inflation shock (h_π): -7.493*** (0.967)
    - Monetary policy uncertainty (Δi^2): -0.897*** (0.321)
    - α_0: 13.627*** (0.592)
    - Adj R^2: 0.596
  - Column (2):
    - Permanent inflation shock (h_π): -11.830*** (1.166)
    - Monetary policy uncertainty (Δi^2): -0.975* (0.522)
    - α_0: 15.010*** (2.420)
    - Adj R^2: 0.599
  - Column (3):
    - Permanent inflation shock (h_π): -7.493*** (0.967)
    - Output gap (y): 1.460*** (0.669)
    - α_0: 10.563*** (0.669)
    - Adj R^2: 0.803
- Summary of determinant effects:
  - Permanent inflation shock (h_π) has a large negative effect on credibility (statistically significant).
  - Inflation volatility negatively affects credibility.
  - Monetary policy uncertainty (squared change in interest rate) negatively affects credibility.
  - Output gap positively associated with credibility.
  - Adjusted R-squared values: 0.596, 0.599, and 0.803 for columns (1)–(3) respectively.

### Monetary policy credibility and ERPT — core empirical relationship
- Visual and correlation evidence:
  - From 2003Q2 to 2015Q1, credibility and ERPT negatively correlated.
  - High ERPT at beginning of sample coincides with low credibility; as credibility improves ERPT trends downward.
  - Episodes:
    - Credibility peaked in 2006 then declined starting 2007Q2; ERPT peaked at 27 percent in 2009Q3 while credibility reached trough two quarters earlier.
    - From 2009 to 2012 negative relationship persists; credibility rises sharply then remains high and stable; ERPT relatively constant from 2012 onward.
- Longer-run perspective using cumulative credibility (Redl, available since 1997):
  - Similar pattern except for a sharp increase in credibility in 2013.
  - Positive correlation between ERPT and credibility from 1997 to 2000; negative relationship evident from 2001 onward.
- Second-stage ERPT model (equation (9)): γ_t = β_0 + β_1 δ_t + ω′ Z_t + ξ_t, authors set Z_t = 0 to avoid multicollinearity.

### Regression evidence from Table 4 (estimation of equation (9))
- Regression (1): authors' credibility measure; sample 2003Q2–2015Q1; lagged credibility (three quarters).
  - Monetary policy credibility index (δ): -0.004*** (0.001)
  - β0: 0.225*** (0.014)
  - N: 47
  - R2: 0.638
- Regression (2): Redl’s credibility measure; sample 1997–2014.
  - Monetary policy credibility index (δ): 0.004 (0.003)
  - β0: 0.343*** (0.051)
  - N: 71
  - R2: 0.468
- Regression (3): Redl’s credibility measure; sample 2001–2014.
  - Monetary policy credibility index (δ): -0.002** (0.001)
  - β0: 0.178*** (0.017)
  - N: 47
  - R2: 0.430
- Interpretation:
  - Regression (1) confirms a negative relationship between credibility and ERPT (coefficient significant at 1 percent; R2 = 0.638).
  - Regression (2) using full sample from 1997 finds credibility does not explain ERPT movements (early positive relationship can be misleading).
  - Regression (3) supports a changing relationship over time.

### Robustness checks (Section V and Table 5)
- Strategies to address endogeneity and measurement:
  - Use lagged credibility (t-3).
  - Use four credibility measures:
    - (i) BER survey of business (eq. 1).
    - (ii) BER survey of trade unions (eq. 2).
    - (iii) Deviations of SARB forecasts from Reuters consensus (eq. 3).
    - (iv) Deviations of SARB forecasts from actual inflation (eq. 4).
  - BER sample sizes: business forecasters vary from 169 to 633; trade unions vary from 7 to 29.
- Table 5 key coefficients (monetary policy credibility lagged t-3):
  - Column (1) = BER business:
    - Monetary policy credibility (t-3): -0.012** (0.006)
    - β0: 0.189*** (0.018)
    - N: 53
    - R2: 0.690
  - Column (2) = BER trade unions:
    - Monetary policy credibility (t-3): -0.004 (0.003)
    - β0: 0.165*** (0.013)
    - N: 53
    - R2: 0.661
  - Column (3) = SARB vs Reuters deviations:
    - Monetary policy credibility (t-3): -0.002*** (0.001)
    - β0: 0.207*** (0.014)
    - N: 47
    - R2: 0.603
  - Column (4) = SARB forecast deviation from actual inflation:
    - Monetary policy credibility (t-3): -0.001 (0.001)
    - β0: 0.190*** (0.016)
    - N: 47
    - R2: 0.476
- Robustness summary:
  - Results broadly similar across credibility measures; some credibility measures produce statistically insignificant coefficients (columns (2) and (4)).

### Policy implications and recommendations
- Core findings with policy relevance:
  - Improving monetary policy credibility has helped reduce ERPT to inflation in South Africa over the sample period.
  - Inflation volatility negatively affects monetary policy credibility.
  - Monetary policy uncertainty (policy-rate volatility) undermines the predictability of inflation and weakens anchoring of inflation expectations.
  - Higher inflationary environment is detrimental to central bank credibility; credibility improves in favorable economic situations and deteriorates during recessions.
- Policy levers identified to increase credibility:
  - Greater transparency.
  - Better communication.
  - Lower volatility in interest rates.
- Operational implication:
  - Monetary authorities need not react strongly to exchange rate fluctuations and can focus on domestic determinants of inflation.

### Open questions and areas for further research
- Whether improving monetary policy credibility alone would lower inflation expectations, which remain persistently at the high end of the 3–6 percent target range, is unresolved.
- Structural factors potentially keeping inflation expectations high:
  - Market concentration in price formation among producers and trade unions (conjectured by Klein (2011) and Fedderke et al. (2007)).
- Interaction of credibility improvements with structural features of the economy (e.g., markup behavior) is a topic for further investigation.

*Source: wp18173 - 1. Bank Transparency Index: South Africa and its Peers (excerpt) and Sections IV and 4 (IMF working paper text provided).*

### 1. Bank Transparency Index: South Africa and its Peers ..........................................................8

### 1. Bank Transparency Index: South Africa and its Peers

### Introduction
- Following Kabundi and Mbelu (2018), exchange rate pass-through to consumer inflation has declined considerably in South Africa since the adoption of the inflation targeting framework in the early 2000.
- This paper examines the role of monetary policy credibility in explaining the decline in pass-through.

### Importance for Policy
- Understanding exchange rate pass-through in an emerging market like South Africa is important because such economies are prone to shocks that can materially impact inflation.
- Since adoption of inflation targeting in 2000, the central bank has allowed the rand to fluctuate with hardly any interventions, leaving limited direct control over exchange rate effects on inflation.
- A clear assessment of pass-through is necessary because monetary authorities can primarily affect the domestic component of price formation.

### Prior Findings on Pass-Through
- It is well documented that in South Africa—and in line with other emerging markets—the pass-through has declined over the past fifteen years or so.
- Kabundi and Mbelu (2018) assess pass-through for 1994–2014 and split it into two phases:
  - First phase (impact of exchange rate fluctuations on import prices): the pass-through is nearly complete.
  - Second phase (transmission of import prices to headline consumer prices): the pass-through is incomplete and has declined from about 70 percent in the late nineties to about 20 percent at the end of the period.
- They also find asymmetry by business cycle stage: both the level and speed of pass-through in first and second stages rise during cyclical upturns and decline when the economy is weak.

### Hypothesis: Role of Monetary Policy Credibility
- The paper conjectures, following literature such as Taylor (2000), that improved monetary policy credibility largely explains the declining pass-through.
- Monetary policy credibility is described as the degree to which agents believe the central bank will act to meet key policy objective(s), especially inflation.
- Credibility is unobservable and must be inferred; Dovern, Fritsche and Slacalek (2012) argue that the degree of (dis)agreement among professional forecasters is a good proxy for credibility.
- Monetary policy transparency, predictability, and perceived central bank independence help coordinate forecasters’ inflation expectations.

### Measuring Credibility and Related Factors
- When monetary policy is credible, uncertainty around objectives and tools is reduced; keeping inflation broadly within the target range, despite shocks, increases credibility.
- Another measure of credibility is the degree to which the central bank meets its own inflation objective, controlling for shocks.
- In the baseline estimations this paper uses the standard deviation or the coefficient of variation of individual inflation forecasts as measures of monetary policy credibility.

### Empirical Findings (from the paper’s approach and preview)
- The paper shows that monetary policy credibility (measured by forecast dispersion) is negatively affected by:
  - the level of inflation itself,
  - monetary policy uncertainty (the degree to which interest rates fluctuate),
  - and a measure of the unobserved stochastic volatility of inflation.
- Using a monetary policy credibility index derived from the first phase, the paper analyzes determinants of the pass-through and confirms that the declining pass-through is explained by improving monetary policy credibility.
- Results are confirmed by several robustness checks using different measures of credibility and after controlling for possible endogeneity.

### Paper Structure (preview)
- Section II: literature review.
- Section III: specification framework to evaluate monetary policy credibility and the model for analyzing the relationship between monetary policy credibility and exchange rate pass-through in South Africa.

*Source: wp18173 - 1. Bank Transparency Index: South Africa and its Peers (excerpt).*

### Section IV discusses the empirical results, focusing first on determinants of monetary policy

### wp18173 - Section IV discusses the empirical results, focusing first on determinants of monetary policy credibility

### Empirical strategy and data
- Quarterly data from South Africa observed from 1980Q1 to 2015Q2 are used to obtain time-varying ERPT based on a 10-year rolling window regression (equations (10) and (11)).
- Extracted ERPT: decline from 50 percent in 1998 to 20 percent in 2002; thereafter roughly stable between 20 percent and 30 percent.
- Measure of forecast disagreement constructed from Reuters’ survey available from 2002Q2.
  - Number of respondents: about 13 at the beginning of the sample, maximum of 25 respondents in September 2014.
- For the pre-IT period an alternative credibility series from Redl (2015) is used.
- Credibility proxy construction:
  - Convert Reuters’ quarterly-expressed forecasts using equation (1) (weighted moving average of fixed-event forecasts).
  - Disagreement measure σ_t^π: standard deviation or interquartile range across forecasters.
  - Index of monetary policy credibility δ_t = 1 / σ_t^π (equation (2)).
- Permanent shock to inflation h_t^π estimated with a UCSV framework (bounded trend between 3 and 6 percent).
  - Bounded trend inflation innovation assumed truncated normal: ε_t^τ ~ TTN(a−τ_{t−1}, b−τ_{t−1}; σ_τ^2) where a and b are lower and upper bounds of the target band.

### Determinants of monetary policy credibility — empirical findings
- Visual and correlation evidence:
  - Two disagreement measures (standard deviation and interquartile range) strongly correlated with coefficient of correlation of 0.90.
  - Forecast disagreement co-moves with inflation; periods of high inflation (2002Q3–2003Q3 and 2007Q4–2009Q1) show high disagreement.
  - Deviation of realized inflation from midpoint 4.5 percent has substantially declined over time.
  - Forecast disagreement closely related to estimated inflation volatility (same estimated inflation volatility as in Kabundi, Schaling, and Some (2018)).
  - Forecast disagreement countercyclical: disagreement rises when output gap is negative; falls when output gap is positive (Figure 8).
- Inflation expectations formation (equation (12)): π_t^e = ρ π_{t−1} + (1−ρ) π^* + ε_t, where π^* is set at 6 percent in the estimation.
  - Sample split results:
    - Pre-2009 (Table 1, N = 32):
      - Total: Lagged Inflation 0.275***; Target (6) 0.725***; AR(1) 0.771***.
      - Analyst: Lagged Inflation 0.098**; Target (6) 0.902***; AR(1) 0.927***.
      - Business: Lagged Inflation 0.306***; Target (6) 0.694***; AR(1) 0.780***.
      - Trade Union: Lagged Inflation 0.452***; Target (6) 0.548***; AR(1) 0.594***.
      - Note: Robust standard errors reported in parentheses; significance: ***p < 0.01, **p < 0.05.
    - Post-2009 (Table 2, N = 33):
      - Total: Lagged Inflation -0.024; Target (6) 1.024***; AR(1) 0.732***.
      - Analyst: Lagged Inflation 0.066*; Target (6) 0.934***; AR(1) 0.947***.
      - Business: Lagged Inflation -0.087; Target (6) 1.087***; AR(1) 0.828***.
      - Trade Union: Lagged Inflation -0.098; Target (6) 1.098***; AR(1) 0.619***.
      - Note: Robust standard errors reported in parentheses; significance: ***p < 0.01, *p < 0.1.
  - Interpretation:
    - Pre-2009: agents somewhat backward-looking (lagged inflation coefficients statistically significant).
    - Post-2009: agents largely forward-looking and expectations anchored at 6 percent (coefficients on lagged inflation mostly not different from zero; Wald test does not reject coefficient of target being one for post-2009).
- FMOLS cointegrating regressions of equation (3) (Table 3) — determinants of credibility δ_t:
  - Regression outcomes (N = 48 for each):
    - Column (1):
      - Permanent inflation shock (h_π): -7.493*** (0.967)
      - Monetary policy uncertainty (Δi^2): -0.897*** (0.321)
      - α_0: 13.627*** (0.592)
      - Adj R^2: 0.596
    - Column (2):
      - Permanent inflation shock (h_π): -11.830*** (1.166)
      - Monetary policy uncertainty (Δi^2): -0.975* (0.522)
      - α_0: 15.010*** (2.420)
      - Adj R^2: 0.599
    - Column (3):
      - Permanent inflation shock (h_π): -7.493*** (0.967) [same numeric value reported in table heading layout]
      - Output gap (y): 1.460*** (0.669)
      - α_0: 10.563*** (0.669)
      - Adj R^2: 0.803
  - Note: Standard errors in parentheses; significance notation: ***p < 0.01, **p < 0.05, *p < 0.1.
- Summary of determinant effects (textual and econometric):
  - Permanent inflation shock (h_π) has a large negative effect on credibility (statistically significant).
  - Inflation volatility negatively affects credibility.
  - Monetary policy uncertainty (squared change in interest rate) negatively affects credibility.
  - Output gap positively associated with credibility (credibility increases in favorable economic situations and deteriorates during recessions).
  - Adjusted R-squared values: 60 percent for the two first regressions, and 80 percent for the last regression.

### Monetary policy credibility and exchange rate pass-through (ERPT)
- Visual and correlation evidence:
  - From 2003Q2 to 2015Q1, credibility and ERPT are negatively correlated (Figure 9).
  - High ERPT at beginning of sample coincides with low credibility; as credibility improves ERPT trends downward.
  - Specific episodes:
    - Credibility peaked in 2006 before starting decline in 2007Q2; ERPT peaked at 27 percent in 2009Q3 while credibility reached trough two quarters earlier.
    - From 2009 to 2012 negative relationship persists; credibility rises sharply then remains high and stable; ERPT remains relatively constant from 2012 onward.
- Longer-run perspective using cumulative credibility (Redl, available since 1997):
  - Pattern similar to shorter sample except for a sharp increase in credibility in 2013.
  - Positive correlation between ERPT and credibility from 1997 to 2000 (both increased until 1998Q3 then declined; credibility bottom in 2000 while ERPT still decreasing).
  - Negative relationship becomes evident only from 2001 onward.
- Model specification for second-stage ERPT (equation (9)):
  - γ_t = β_0 + β_1 δ_t + ω′ Z_t + ξ_t, with Z_t set to 0 in the authors’ model to avoid multicollinearity with variables already in equation (3).

### Key policy implications and conclusions drawn from Section IV
- Three main policy implications derived from the results:
  - Inflation volatility negatively affects monetary policy credibility in South Africa.
  - Uncertainty in monetary policy (policy-rate volatility) undermines the predictability of inflation and weakens anchoring of inflation expectations.
  - Higher inflationary environment is detrimental to central bank credibility; credibility improves in favorable economic situations and deteriorates during recessions.
- Additional observations:
  - Adoption of inflation-targeting framework and improvements in central bank transparency (CBT) are linked to gains in credibility; South Africa’s CBT increased from 5 to 8.5 during 1999–2001 (Dincer and Eichengreen (2014) index range 0–15).
  - Communication improvements by SARB (publication of macroeconomic assumptions and forecasts, six-monthly monetary policy review, publication of core Quarterly Projection Model in 2017) likely further improved credibility after 2012.
  - Anchoring of inflation expectations across agents strengthened after 2009, consistent with higher credibility and a decline in ERPT.
- Methodological notes:
  - Credibility index δ_t defined as inverse of forecast disagreement; interquartile range preferred in some analyses for robustness to outliers.
  - Use of FMOLS accounts for serial correlation and heteroscedasticity.
  - Use of core instead of headline inflation does not materially change basic results.

*Source: wp18173 - Section IV discusses the empirical results, focusing first on determinants of monetary policy credibility (IMF working paper text provided).*

### 4. Table 4 is based on the estimation of equation (9).

### 4. Table 4 is based on the estimation of equation (9).

### Regression specifications and sample coverage
- Regression (1): uses the authors' measure of credibility; sample from 2003Q2 to 2015Q1; lagged credibility used (three quarters) to mitigate endogeneity.
- Regression (2): uses Redl’s credibility measure; sample from 1997 to 2014 (entire sample).
- Regression (3): uses Redl’s credibility measure on a short sample to account for change in the nature of the relationship; sample from 2001 to 2014.
- Note: Lagged credibility is used to account for the possibility of endogeneity arising from the reaction of monetary policy credibility to changes in the ERPT.

### Key empirical findings (Table 4)
- Regression (1):
  - Monetary policy credibility index (δ): -0.004***
  - Standard error: (0.001)
  - β0: 0.225***
  - Standard error: (0.014)
  - N: 47
  - R2: 0.638
- Regression (2):
  - Monetary policy credibility index (δ): 0.004
  - Standard error: (0.003)
  - β0: 0.343***
  - Standard error: (0.051)
  - N: 71
  - R2: 0.468
- Regression (3):
  - Monetary policy credibility index (δ): -0.002**
  - Standard error: (0.001)
  - β0: 0.178***
  - Standard error: (0.017)
  - N: 47
  - R2: 0.430
- Statistical significance notation: ***p < 0.01 and **p < 0.05.
- Interpretation highlights:
  - Regression (1) confirms a negative relationship between monetary policy credibility and exchange rate pass-through (ERPT); coefficient statistically significant at 1 percent and R2 of 64 percent.
  - Using the full sample from 1997 (regression (2)), credibility does not explain movement in the ERPT; early positive relationship in the sample can be misleading.
  - Regression (3) supports a changing relationship over time between credibility and ERPT.

### Dynamics of credibility and inflation expectations
- Decline in credibility after 2007 coincides with a remarkable surge in inflation.
- Both credibility and inflation rally from 2009 to the end of the sample, coinciding with stability in inflation expectations (Figure 5 referenced).
- From 2009 to 2011 the two credibility series (businesses and trade unions) move closely; after 2011 the trade unions' credibility measure becomes steeper than that of businesses.
- Regression (1) uses business-based credibility; regression (2) uses trade-union-based credibility.
  - Credibility of business explains movement in the ERPT, while credibility of trade unions yields an insignificant coefficient—attributed to the steeper rise in trade unions' credibility from 2011.
  - When estimating the model until 2011, the coefficient moves from -0.005 to -0.016 and becomes significant at 1 per cent.

### Robustness checks (Section V and Table 5)
- Approaches to robustness:
  - Use of lagged credibility (three quarters) to mitigate endogeneity.
  - Use of four different measures of credibility:
    - (i) BER survey of business (eq. 1).
    - (ii) BER survey of trade unions (eq. 2).
    - (iii) Deviations of SARB forecasts from Reuters consensus forecasts (eq. 3).
    - (iv) Deviations of SARB forecasts from actual inflation (eq. 4).
  - For BER data, standard deviation of two-year-ahead forecasts of inflation by different agents; business forecasters sample varies from 169 to 633; trade unions sample varies from 7 to 29.
- Table 5 results (monetary policy credibility lagged t-3):
  - Column (1) = Credibility from BER survey of business:
    - Monetary policy credibility (t-3): -0.012**
    - Standard error: (0.006)
    - β0: 0.189***
    - Standard error: (0.018)
    - N: 53
    - R2: 0.690
  - Column (2) = Credibility based on BER survey of trade unions:
    - Monetary policy credibility (t-3): -0.004
    - Standard error: (0.003)
    - β0: 0.165***
    - Standard error: (0.013)
    - N: 53
    - R2: 0.661
  - Column (3) = Credibility based on deviation of SARB forecasts from Reuters forecast:
    - Monetary policy credibility (t-3): -0.002***
    - Standard error: (0.001)
    - β0: 0.207***
    - Standard error: (0.014)
    - N: 47
    - R2: 0.603
  - Column (4) = Credibility based on SARB forecast deviation from actual inflation:
    - Monetary policy credibility (t-3): -0.001
    - Standard error: (0.001)
    - β0: 0.190***
    - Standard error: (0.016)
    - N: 47
    - R2: 0.476
- Robustness summary:
  - Results are generally similar across credibility measures.
  - Monetary policy credibility coefficients for equations (2) and (4) are not statistically significant.

### Policy implications and recommendations
- Improving monetary policy credibility has helped reduce the exchange rate pass-through to inflation in South Africa over the sample period examined.
- Consequence for policy:
  - Monetary policy authorities need not react strongly to exchange rate fluctuations (a policy choice not to influence the exchange rate), allowing focus on domestic factors under their control.
- Policy levers to increase credibility identified in the paper:
  - Greater transparency.
  - Better communication.
  - Lower volatility in interest rates.

### Open questions and areas for further research
- The paper does not resolve whether improving monetary policy credibility alone would lower inflation expectations, which appear persistently at the high end of the 3–6 percent target range.
- Possible structural factors keeping inflation expectations high include market concentration in price formation among producers and trade unions (as conjectured by Klein (2011) and Fedderke et al. (2007)).
- The interaction of credibility improvements with structural features of the economy (e.g., markup behavior) is identified as a worthy topic for further investigation.

*Source: wp18173 - 4. Table 4 is based on the estimation of equation (9).*

---


_Source: https://www.imf.org/-/media/files/publications/wp/2018/wp18173.pdf_