## _wp1209

## Source details

**Canonical URL:** [_wp1209](https://www.imf.org/-/media/websites/imf/imported-full-text-pdf/external/pubs/ft/wp/2012/_wp1209.pdf)

## Other formats

- [Markdown version](/-/media/websites/imf/imported-full-text-pdf/external/pubs/ft/wp/2012/_wp1209.pdf.md)
- [Structured JSON version](/-/media/websites/imf/imported-full-text-pdf/external/pubs/ft/wp/2012/_wp1209.pdf.json)

---

### I. Introduction — scope and motivation
- Sample period: 1997 to 2010.
- Sample of central banks: Banco Central de Chile; Bank of England (BOE); Bank of Thailand; Czech National Bank (CNB); European Central Bank (ECB); National Bank of Poland (NBP); Sveriges Riksbank.
- Main aims:
  - Quantify clarity using an objective textual metric and document length.
  - Document how clarity evolved for seven central banks over the last decade.
  - Analyze whether clarity is related to the inflation outlook and uncertainty around the outlook.

### II. Data, content analysis, and measurement
- Communication types:
  - Long documents for specialists: inflation reports, monetary policy reports, bulletins.
  - Short documents for the general public: press releases and executive summaries.
- Content analysis:
  - Extracted text fragments containing the central bank’s view on inflation risks (demand, supply, external).
  - Ternary coding scheme for each factor: inflationary (+1), neutral (0), deflationary (-1).
  - Overall verbal assessment S = sum of individual coded inflation factors (equal weights).
  - Empirical measures used: |S| and V = |AD – AS| + |AD – FOR| + |AS – FOR|.
- Frequency: most communications data available quarterly; for the ECB quarterly averages of Monthly Bulletins used.
- Readability metric: Flesch–Kincaid (FK) grade level with formula:
  - 0.39*(# words / # sentences) + 11.8 * (# syllables / # words) - 15.59
  - FK range in sample: varies from 10 to 19 years.
- Alternative dimension: document length measured by number of words.

### III. Empirical specification
- Reduced-form regression (informal): FK_it regressed on country fixed effects and trend (T_it; 1998q1 = 1), |S_it|, V_it, |E_it[π_i,t+4] − π_i,t*|, |π_i,t − π_i,t*|, and DIS_it (fraction of dissenting votes when available).
- Prior expectation: higher values of the five covariates imply more to explain and thus expected positive effects on FK (less clarity), though increased drafting effort could offset this.

### IV. Key empirical findings — overview
- Readability is largely country-specific and driven by deterministic variables (constants and trends).
- Significant and persistent differences in clarity over time and across countries; country-specific trends and idiosyncratic shocks matter.
- Short-term fluctuations in clarity are difficult to account for with included covariates.
- Global financial crisis had a measurable adverse effect on clarity in some countries.

### V. Detailed results — clarity of inflation reports (selected findings)
- FK trends (coefficient per year, robust SE in parentheses):
  - Chile: -0.03* (0.02)
  - Euro area: 0.03** (0.01)
  - Poland: 0.04 (0.03)
  - Sweden: -0.04*** (0.01)
  - Thailand: 0.11** (0.04)
  - United Kingdom: -0.02** (0.01)
- Constant terms (FK levels, robust SE in parentheses; all significant at *** p<0.01):
  - Chile: 16.39*** (0.55)
  - Czech Republic: 14.09*** (0.79)
  - Euro area: 14.81*** (0.39)
  - Poland: 14.26*** (1.19)
  - Sweden: 13.88*** (0.56)
  - Thailand: 13.12*** (1.70)
  - United Kingdom: 12.99*** (0.60)
- Selected covariate effects (coefficient (robust SE)):
  - Czech Republic: |S| = -0.24** (0.10)
  - United Kingdom: |E_t[π_{t+4}] - π*| = 0.84* (0.46)
  - Czech Republic: |π - π*| = 0.37** (0.16)
  - Sweden: Dissent = -2.77* (1.38) [when available]
- Sample sizes and fit (selected):
  - Obs.: Chile 30, Czech Republic 26, Euro area 35, Poland 18, Sweden 31, Thailand 22, United Kingdom 48
  - Adj. R^2: Chile -0.01, Czech Republic 0.23, Euro area 0.05, Poland -0.03, Sweden 0.30, Thailand 0.28, United Kingdom 0.23

### VI. Detailed results — clarity of press releases / statements (selected findings)
- Trend coefficients (coefficient (robust SE)):
  - Chile: -0.07*** (0.02)
  - Euro area: 0.04*** (0.01)
  - Poland: 0.07** (0.02)
  - Sweden: -0.06*** (0.01)
  - Thailand: 0.12** (0.05)
  - United Kingdom: 0.02* (0.01)
- Constant terms (FK levels, robust SE in parentheses; all significant at *** p<0.01):
  - Chile: 16.80*** (0.71)
  - Euro area: 14.24*** (0.42)
  - Poland: 13.09*** (0.69)
  - Sweden: 14.16*** (0.45)
  - Thailand: 12.03*** (1.29)
  - United Kingdom: 11.20*** (0.41)
- Selected covariate effects (coefficient (robust SE)):
  - Thailand: V = 0.18* (0.09)
  - United Kingdom: V = 0.10** (0.04)
  - Euro area: |π - π*| = 0.73* (0.36)
  - Sweden: |π - π*| = -0.44** (0.19)
  - United Kingdom: |π - π*| = -0.57** (0.21)
- Sample sizes and fit (selected):
  - Obs.: Chile 30, Euro area 35, Poland 18, Sweden 38, Thailand 20, United Kingdom 36
  - Adj. R^2: Chile 0.42, Euro area 0.10, Poland 0.40, Sweden 0.42, Thailand 0.32, United Kingdom 0.28

### VII. Detailed results — length of inflation reports (executive summaries)
- Trend coefficients (word count per year, coefficient (robust SE)):
  - Euro area: 11*** (3)
  - Poland: 53*** (9)
  - Thailand: 38*** (7)
  - United Kingdom: 11*** (1)
- Constant terms (word counts, robust SE in parentheses):
  - Chile: 2,918*** (535)
  - Czech Republic: 985*** (204)
  - Euro area: 1,139*** (123)
  - Poland: -667* (337)
  - Sweden: 1,437** (539)
  - Thailand: 994*** (183)
  - United Kingdom: 1,162*** (63)
- Selected covariate effects (coefficient (robust SE)):
  - Czech Republic: |S| = -63* (32)
  - Euro area: V = -43*** (15)
  - Thailand: V = -36* (18)
  - United Kingdom: V = -18*** (6)
  - Euro area: |π - π*| = 287** (127)
  - Thailand: |π - π*| = -204** (86)
  - United Kingdom: |π - π*| = 67** (28)
- Sample sizes and fit (selected):
  - Obs.: Chile 30, Czech Republic 26, Euro area 35, Poland 18, Sweden 31, Thailand 22, United Kingdom 48
  - Adj. R^2: Chile -0.10, Czech Republic 0.19, Euro area 0.41, Poland 0.64, Sweden -0.10, Thailand 0.68, United Kingdom 0.61

### VIII. The Global Financial Crisis effect
- Crisis dummy: 1 for 2008q4 – 2009q4, 0 otherwise.
- Crisis dummy coefficients (dependent variable: Flesch-Kincaid grade level, coefficient (robust SE)):
  - Poland: 1.75** (0.73)
  - Sweden: 1.46*** (0.38)
  - United Kingdom: 1.15*** (0.24)
- Selected other coefficients (coefficient (robust SE)):
  - Poland: |S| = 0.19** (0.08); |E_t[π_{t+4}] - π*| = -0.50** (0.21)
  - Sweden: |E_t[π_{t+4}] - π*| = 0.33 (0.43)
  - United Kingdom: |E_t[π_{t+4}] - π*| = 0.63* (0.34)
- Sample sizes and fit:
  - Obs.: Poland 18, Sweden 31, United Kingdom 48
  - Adj. R^2: Poland 0.33, Sweden 0.46, United Kingdom 0.36
- Interpretation:
  - For Poland, Sweden, and the United Kingdom the crisis dummy is statistically significant and positive.
  - Estimated effect: on average the financial crisis added about 1-2 additional years of schooling needed to understand the reports in these countries.
  - Pooled regression coefficient for the crisis dummy around 0.5 but not statistically significant.

### IX. Sensitivity analysis and robustness
- Adding counts of individual inflation factors (AD_it, AS_it, FOR_it) as regressors: statistically insignificant.
- Adding lagged values of past inflation deviations and voting dissent: results do not differ materially from main results.
- Automated general-to-specific procedure (Autometrics in Oxmetrics): qualitative findings unaffected; routine generally selected variables identical to those in main analysis after identifying country-specific time dummies for the crisis period.

### X. Interpretation, conclusions, and suggested extensions
- Deterministic variables (constant and trend) are consistently among the most relevant predictors of readability across specifications; country-specific long-term trends dominate readability patterns.
- Little evidence that central banks were systematically less able to clearly explain policies when faced with higher uncertainty or a less favorable inflation outlook: few non-deterministic covariates have consistent effects across countries.
- Specific relationships found in some countries:
  - Forward-looking inflation gaps and contemporaneous gaps affect clarity in some countries (notably the United Kingdom and the Czech Republic).
  - Voting dissent and signal counts affect clarity in some country-specific ways (BOE, CNB).
- The global financial crisis posed a major communication challenge, reducing clarity in several countries.
- Suggested extensions and follow-up research:
  - Extend analysis beyond the seven central banks to examine how central banks adjust monetary policy communications to external developments.
  - Examine whether changes in central bank communication clarity affect market and public expectations of interest rates and inflation.
  - Expand analysis to other forms of central bank communication: oral statements by central bank officials, contents of monetary policy committee minutes (where published), and communication on topics other than the inflation outlook (in particular, financial stability).

*Source: IMF working paper section "2. Summary Statistics for Readability of Inflation Reports and Press Statements" (excerpt provided).*

### References .............................................................................................................

### _wp1209 - References .............................................................................................................

### Tables
- 1. Data Overview .....................................................................................................................16
- 2. Sample Correlations .............................................................................................................17
- 3. Explaining Clarity of Inflation Reports ...............................................................................18
- 4. Explaining Clarity of Press Releases / Statements ..............................................................19
- 5. Explaining the Length of Inflation Reports .........................................................................20
- 6. The Global Financial Crisis and Clarity of Inflation Reports ..............................................21

### Figures
- 1. Readability of Inflation Reports and Press Statements, 1997–2010 ....................................14

*Source: https://www.imf.org/-/media/websites/imf/imported-full-text-pdf/external/pubs/ft/wp/2012/_wp1209.pdf*

### 2. Summary Statistics for Readability of Inflation Reports and Press Statements ..................15

### 2. Summary Statistics for Readability of Inflation Reports and Press Statements ..................15

### I. Introduction — scope and motivation
- Paper studies clarity of central bank communication on the inflation outlook and asks whether clarity depends on context (inflation outlook, uncertainty).
- Sample period: 1997 to 2010.
- Main aims:
  - Quantify clarity using an objective textual metric and document length.
  - Document how clarity evolved for seven central banks over the last decade.
  - Analyze whether clarity is related to the inflation outlook and uncertainty around the outlook.
- Sample of central banks: Banco Central de Chile; Bank of England (BOE); Bank of Thailand; Czech National Bank (CNB); European Central Bank (ECB); National Bank of Poland (NBP); Sveriges Riksbank.

### II. Data and content analysis
- Communication types analyzed:
  - Long documents for specialists: inflation reports, monetary policy reports, bulletins.
  - Short documents for the general public: press releases and executive summaries.
- Content analysis:
  - Text fragments containing the central bank’s view on inflation risks (demand, supply, external) were extracted.
  - Ternary coding scheme for each factor: inflationary (+1), neutral (0), deflationary (-1).
  - Overall verbal assessment S = sum of individual coded inflation factors (equal weights).
  - Empirical analyses use |S| (absolute value) and dispersion measure V = |AD – AS| + |AD – FOR| + |AS – FOR|.
- Frequency: most communications data available quarterly; for the ECB quarterly averages of Monthly Bulletins used.

### III. Measuring clarity (Flesch-Kincaid and length)
- Clarity metric: Flesch–Kincaid (FK) grade level (years of education needed to comprehend text).
- FK formula (as used in the paper):
  - 0.39*(# words / # sentences) + 11.8 * (# syllables / # words) - 15.59
- FK range in sample: varies from 10 to 19 years.
- Alternative dimension: document length (number of words). Length treated separately from clarity because longer communications can both aid precision and bury the message.

### IV. Empirical specification for clarity
- Reduced-form regression (informal description of equation (2)):
  - FK_it (FK grade level of communication by central bank i in quarter t) is regressed on:
    - country fixed effects and a linear trend (T_it; 1998q1 = 1 used for trend term),
    - absolute count of inflation factors |S_it|,
    - dispersion of inflation signals V_it,
    - forward-looking inflation gap: |E_it[π_i,t+4] − π_i,t*|,
    - contemporaneous inflation gap: |π_i,t − π_i,t*|,
    - fraction of dissenting votes DIS_it (for countries publishing voting records).
- Priors: higher values of the five covariates imply more to explain and thus expected positive effects on FK (less clarity), though increased drafting effort could offset this.

### V. Key empirical findings — overview
- Readability is largely country-specific and driven by deterministic variables (constants and trends).
- There are significant and persistent differences in clarity over time and across countries; readability exhibits country-specific trends and idiosyncratic shocks.
- Short-term fluctuations in clarity are difficult to account for with the included covariates.
- Global financial crisis had a measurable adverse effect on clarity in some countries.

### VI. Detailed results by communication type
- A. Clarity of inflation reports
  - FK trends:
    - Improvements over time: Chile, Sweden, United Kingdom — improving by almost 1/5 of a year of schooling per year.
    - Worsening over time: euro area Monthly Bulletins (about 1/10 of a year per year) and Thailand (about 2/5 of a year per year).
    - Czech Republic and Poland: no statistically significant trends.
  - Covariate effects (country-specific, few consistent results):
    - |S| (number of inflation factors): associated with a reduction in FK (more clarity) only for the Czech Republic.
    - V (signal dispersion): no significant impact on report clarity in most countries.
    - Forward-looking inflation gap |E[π_{t+4}] − π*|: increases FK (reduces clarity) in three out of seven countries; statistically significant only for the United Kingdom (decline in readability of about 3/4 of a year of schooling).
    - Contemporaneous inflation gap |π_t − π*|: associated with lower clarity for most countries, statistically significant only for the Czech Republic.
    - Dissent (DIS): higher fraction of dissent associated with increased clarity only in the United Kingdom.

- B. Clarity of press releases and statements
  - Trends:
    - Communication became clearer over time in Chile and Sweden (by about 1/4 of a year of schooling per year).
    - Communication became less clear over time in the euro area, Poland, Thailand, and the United Kingdom (decline in Thailand six times larger than in the United Kingdom).
  - Covariate effects:
    - V (signal dispersion) reduced clarity in Thailand and the United Kingdom.
    - Past inflation deviations from the target associated with increases in clarity in Sweden and the United Kingdom.
    - Overall, non-deterministic variables play no systematic role.

- C. Length of inflation reports (executive summaries)
  - Report length is very country-specific and trended upward in most countries (euro area, Poland, Thailand, United Kingdom).
  - Covariate effects:
    - |S| reduced report length in the Czech Republic.
    - Higher dispersion V associated with shorter reports in the euro area, Thailand, and the United Kingdom (possible effort to be more concise when signals are dispersed).
    - Past inflation deviations associated with longer reports in the euro area and the United Kingdom, but shorter in Thailand.
    - Dissent increases report length in Poland.
  - Pooled regressions suggest no robust effect of the considered variables across all countries.

- D. Impact of the global financial crisis
  - Crisis dummy defined as 1 from 2008q4 to 2009q4, 0 otherwise.
  - For Poland, Sweden, and the United Kingdom the crisis dummy is statistically significant and positive.
  - Estimated effect: on average the financial crisis added about 1-2 additional years of schooling needed to understand the reports in these countries.
  - Pooled regression coefficient for the crisis dummy around 0.5 but not statistically significant.
  - Crisis dummy leaves earlier country estimates broadly unaffected.

- E. Sensitivity analysis
  - Adding counts of individual inflation factors (AD_it, AS_it, FOR_it) as regressors: statistically insignificant.
  - Adding lagged values of past inflation deviations and voting dissent: results do not differ materially from main results.
  - Automated general-to-specific procedure (Autometrics in Oxmetrics): qualitative findings unaffected; routine generally selected variables identical to those identified in main analysis after identifying country-specific time dummies for the crisis period.

### VII. Interpretation and principal conclusions
- Readability (Flesch–Kincaid) varies across countries and over time; part of the variation reflects language and translation issues (many documents prepared in native languages; English versions used here).
- Overall, little evidence that central banks were systematically less able to clearly explain policies when faced with higher uncertainty or a less favorable inflation outlook: few non-deterministic covariates have consistent effects across countries.
- Some specific relationships found:
  - Forward-looking inflation gaps and contemporaneous gaps affect clarity in some countries (notably the United Kingdom and the Czech Republic).
  - Voting dissent and signal counts affect clarity in some country-specific ways (BOE, CNB).
- The global financial crisis posed a major communication challenge, reducing clarity in several countries.
- Policy-relevant implication: clarity is an important pre-condition for transparency and carries potential direct benefits (higher signal-to-noise ratio), but drivers of clarity are largely country-specific and not easily summarized by a small set of macroeconomic covariates.

*Source: IMF working paper section "2. Summary Statistics for Readability of Inflation Reports and Press Statements" (excerpt provided).*

### section III.B: the deterministic variables (constant and trend) remained the most relevant for

### section III.B: the deterministic variables (constant and trend) remained the most relevant for

### Data and methodology
- Sample covers communication on the inflation outlook by seven central banks between 1997 and 2010.
- Readability measure: Flesch-Kincaid grade level (interpreted as the number of years of education needed to comprehend a text).
- Communication channels analyzed: inflation reports (executive/summary sections), press releases, and statements.
- Content analysis combined with the Flesch-Kincaid grade level; missing observations were interpolated.
- Regression independent variables (as described for Tables 3–5): the absolute value of the evaluation on risks to price stability given by the central bank (|S|), the uncertainty regarding this signal (V), the forward-looking inflation gap (|E_t[π_{t+4}] - π*|), the contemporaneous inflation gap (|π_t - π*|), and, when available, the fraction of dissenting votes regarding monetary policy decisions. Some specifications include a crisis dummy equal to 1 for 2008q4 – 2009q4.

### Main empirical findings
- Readability is country-specific and follows long-term trends.
- Overall, little evidence that central banks had greater difficulty explaining policy when faced with higher uncertainty or a less favorable inflation outlook; short-term fluctuations in clarity are difficult to account for.
- The central bank’s assessment of inflation and dissent in voting on interest rates explain some variation in clarity.
- The global financial crisis is generally associated with less clarity in central bank communications.

### Key regression results — Clarity of inflation reports (Table 3, dependent variable: Flesch-Kincaid grade level)
- Trend coefficients (selected countries, coefficient (robust SE)):
  - Chile: -0.03* (0.02)
  - Euro area: 0.03** (0.01)
  - Poland: 0.04 (0.03)
  - Sweden: -0.04*** (0.01)
  - Thailand: 0.11** (0.04)
  - United Kingdom: -0.02** (0.01)
- Constant terms (all statistically significant at *** p<0.01):
  - Chile: 16.39*** (0.55)
  - Czech Republic: 14.09*** (0.79)
  - Euro area: 14.81*** (0.39)
  - Poland: 14.26*** (1.19)
  - Sweden: 13.88*** (0.56)
  - Thailand: 13.12*** (1.70)
  - United Kingdom: 12.99*** (0.60)
- Selected coefficients on other variables (coefficient (robust SE)):
  - Czech Republic: |S| = -0.24** (0.10)
  - United Kingdom: |E_t[π_{t+4}] - π*| = 0.84* (0.46)
  - Czech Republic: |π - π*| = 0.37** (0.16)
  - Sweden: Dissent = -2.77* (1.38) [when available]
- Sample sizes and fit (selected):
  - Obs.: Chile 30, Czech Republic 26, Euro area 35, Poland 18, Sweden 31, Thailand 22, United Kingdom 48
  - Adj. R^2: Chile -0.01, Czech Republic 0.23, Euro area 0.05, Poland -0.03, Sweden 0.30, Thailand 0.28, United Kingdom 0.23

### Key regression results — Clarity of press releases / statements (Table 4, dependent variable: Flesch-Kincaid grade level)
- Trend coefficients (coefficient (robust SE)):
  - Chile: -0.07*** (0.02)
  - Euro area: 0.04*** (0.01)
  - Poland: 0.07** (0.02)
  - Sweden: -0.06*** (0.01)
  - Thailand: 0.12** (0.05)
  - United Kingdom: 0.02* (0.01)
- Constant terms (all statistically significant at *** p<0.01):
  - Chile: 16.80*** (0.71)
  - Euro area: 14.24*** (0.42)
  - Poland: 13.09*** (0.69)
  - Sweden: 14.16*** (0.45)
  - Thailand: 12.03*** (1.29)
  - United Kingdom: 11.20*** (0.41)
- Selected coefficients on other variables (coefficient (robust SE)):
  - Thailand: V = 0.18* (0.09)
  - United Kingdom: V = 0.10** (0.04)
  - Euro area: |π - π*| = 0.73* (0.36)
  - Sweden: |π - π*| = -0.44** (0.19)
  - United Kingdom: |π - π*| = -0.57** (0.21)
- Sample sizes and fit (selected):
  - Obs.: Chile 30, Euro area 35, Poland 18, Sweden 38, Thailand 20, United Kingdom 36
  - Adj. R^2: Chile 0.42, Euro area 0.10, Poland 0.40, Sweden 0.42, Thailand 0.32, United Kingdom 0.28

### Length of inflation reports (Table 5, dependent variable: word count)
- Trend coefficients (coefficient (robust SE)):
  - Euro area: 11*** (3)
  - Poland: 53*** (9)
  - Thailand: 38*** (7)
  - United Kingdom: 11*** (1)
- Constant terms (selected):
  - Chile: 2,918*** (535)
  - Czech Republic: 985*** (204)
  - Euro area: 1,139*** (123)
  - Poland: -667* (337)
  - Sweden: 1,437** (539)
  - Thailand: 994*** (183)
  - United Kingdom: 1,162*** (63)
- Selected coefficients on other variables (coefficient (robust SE)):
  - Czech Republic: |S| = -63* (32)
  - Euro area: V = -43*** (15)
  - Thailand: V = -36* (18)
  - United Kingdom: V = -18*** (6)
  - Euro area: |π - π*| = 287** (127)
  - Thailand: |π - π*| = -204** (86)
  - United Kingdom: |π - π*| = 67** (28)
- Sample sizes and fit (selected):
  - Obs.: Chile 30, Czech Republic 26, Euro area 35, Poland 18, Sweden 31, Thailand 22, United Kingdom 48
  - Adj. R^2: Chile -0.10, Czech Republic 0.19, Euro area 0.41, Poland 0.64, Sweden -0.10, Thailand 0.68, United Kingdom 0.61

### The Global Financial Crisis effect (Table 6: crisis dummy = 1 for 2008q4 – 2009q4)
- Crisis dummy coefficients (coefficient (robust SE), dependent variable: Flesch-Kincaid grade level):
  - Poland: 1.75** (0.73)
  - Sweden: 1.46*** (0.38)
  - United Kingdom: 1.15*** (0.24)
- Selected other coefficients (coefficient (robust SE)):
  - Poland: |S| = 0.19** (0.08); |E_t[π_{t+4}] - π*| = -0.50** (0.21)
  - Sweden: |E_t[π_{t+4}] - π*| = 0.33 (0.43)
  - United Kingdom: |E_t[π_{t+4}] - π*| = 0.63* (0.34)
- Sample sizes and fit:
  - Obs.: Poland 18, Sweden 31, United Kingdom 48
  - Adj. R^2: Poland 0.33, Sweden 0.46, United Kingdom 0.36

### Interpretation and implications
- Deterministic variables (constant and trend) are consistently among the most relevant predictors of readability across specifications.
- Country-specific long-term trends dominate readability patterns, making short-term explanatory power of uncertainty and inflation outlook limited.
- The global financial crisis coincided with higher Flesch-Kincaid grade levels (less clarity) in the three countries where the crisis dummy is significant.
- Some evidence that dissent in monetary policy committees and the central bank’s own inflation assessment explain part of the variation in clarity.

### Suggested extensions and follow-up research (from the paper)
- Extend analysis beyond the seven central banks in the sample to examine factors influencing how central banks adjust monetary policy communications in response to external developments.
- Examine whether changes in central bank communication clarity affect market and public expectations of interest rates and inflation.
- Expand analysis to other forms of central bank communication: oral statements by central bank officials, contents of monetary policy committee minutes (where published), and communication on topics other than the inflation outlook (in particular, financial stability).

*Source: Authors’ calculations and analysis in section III.B of the provided IMF working paper excerpt.*

---


_Source: https://www.imf.org/-/media/websites/imf/imported-full-text-pdf/external/pubs/ft/wp/2012/_wp1209.pdf_