## _wp1045 - 2.  OLS Regressions for Cross-Section of Countries

## Source details

**Canonical URL:** [_wp1045 - 2.  OLS Regressions for Cross-Section of Countries](https://www.imf.org/-/media/websites/imf/imported-full-text-pdf/external/pubs/ft/wp/2010/_wp1045.pdf)

## Other formats

- [Markdown version](/-/media/websites/imf/imported-full-text-pdf/external/pubs/ft/wp/2010/_wp1045.pdf.md)
- [Structured JSON version](/-/media/websites/imf/imported-full-text-pdf/external/pubs/ft/wp/2010/_wp1045.pdf.json)

---

### I. Introduction
- Purpose: evaluate performance of the inflation targeting (IT) monetary framework during the global sudden stop initiated in mid-September 2008; focus on uncovering stylized facts rather than causal identification.
- Framing and critiques:
  - IT may be “conservative window dressing” (Romer, 2006); alternatively, IT’s emphasis on credibility and communication matters for inflation expectations.
  - Criticisms: narrow focus on inflation may neglect unemployment and financial stability (Stiglitz, 2008; Buiter, 2009); vulnerability with liability dollarization (Calvo and Reinhart, 2002).
- Potential advantages of IT during the 2008-style crisis:
  - Credibility helps avoid deflation and a liquidity trap (Governor Carney quote cited).
  - IT adopters may have greater scope to ease policy without compromising inflation outlooks (Ghosh and others, 2009).
  - IT countries tended to have higher interest rates during pre-crisis expansion, providing more room for rate cuts.
  - Correlation between IT and flexible exchange rates, which may act as shock-absorbers (Broda, 2004; Edwards and Levy Yeyati, 2005; Mendoza, 1995).

### II. Methodology
- Empirical framework:
  - Panel regressions with country fixed effects (λi) and time effects (λt).
  - Main specification: y_it = λ_i + λ_t + φ_t × IT_it + ε_it, where φ_t measures the difference in variable y_it between IT and non-IT countries.
- Estimation approaches:
  - OLS for mean IT effect; median (quantile) regressions for median effects.
  - Examine coefficients of time dummies and IT × time interactions.
- Event study timing and samples:
  - Time zero: August 2008.
  - Monthly variables: sample starts January 2006.
  - Quarterly variables: sample starts 2002Q1.
  - Exclusions: countries with ngdpd in 2002 < USD 10 billion; Zimbabwe; Angola; Qatar; Sudan; United Arab Emirates; United States. Full sample after exclusions: 84 countries.
  - Iceland and Latvia excluded due to GDP cutoff.
  - Emerging markets defined as countries with available J.P. Morgan EMBI spread data.
- Outcome variables examined (in order): nominal policy rates, real policy rates, inflation, real effective exchange rate (REER), EMBI spreads, 5-year CDS premia, unemployment rates, industrial production, GDP growth.

### III. Results — Policy rates and real policy rates
- Policy rates (sample 49 countries with 21 IT countries; data through November 2009):
  - Both IT and non-IT countries increased nominal rates from early 2008 to August 2008; IT countries tightened slightly more pre-crisis.
  - In September 2008, 6 out of 22 IT countries tightened policy rates (Brazil, Chile, Indonesia, Israel, Peru, Sweden).
  - After crisis deepened, IT countries cut policy rates by 2 percentage points more than non-IT countries.
  - Difference in mean policy rates across groups was statistically significant and persistent.
- Real policy rates (calculated using 12-month backward-looking inflation):
  - IT countries avoided an increase in real policy rates since the crisis.
  - Loosening of real policy rates for IT countries relative to non-IT countries was about 4 percentage points.

### III. Results — Inflation
- Sample: n=84, of which 25 IT countries; data through November 2009.
- Key findings:
  - Price levels dropped across the board in the last two months of 2008.
  - Average annualized inflation rate was lower by 5 percent for IT countries and 6½ percent for non-IT countries in December 2009 relative to August 2009.
  - Median IT and non-IT countries had negative monthly inflation in December 2009.
  - IT countries were less likely to experience a deflation scare (three consecutive months of negative monthly inflation); statistically significant differences for 2008M12-2009M1 and 2009M6.
  - Footnote data (August 2009): 2 IT countries had 3 consecutive negative monthly inflation readings (Chile and Republic of Serbia); 7 non-IT countries did (Bulgaria, China, P.R.: Mainland, Ireland, Japan, Libya, Lithuania and Taiwan Province of China).

### III. Results — Real effective exchange rate (REER)
- Sample: n=84, 25 IT countries; data through August 2009.
- Key findings:
  - REER depreciation for the average IT country was more than 15 percentage points by 2009Q1 (base period August 2008).
  - By August 2009, IT countries on average were about 5 percent weaker than at crisis outset.
  - Finding holds for full sample of 89 countries, robust to subset of 31 emerging markets, and to excluding countries with pegs/heavily managed floats pre-crisis.
  - Among an 85-country universe with 25 IT countries, the 12 countries with largest real depreciations six months into the crisis (February 2009) were all IT countries; examples include Poland at -31½ percent and Turkey at -17½ percent.
  - No IT countries among top 30 countries with real appreciations during that period (e.g., Japan appreciated by 26½ percent; Venezuela, Rep. Bol. by 24 percent).
  - Definition: pegs/heavily managed floats = monthly exchange rate standard deviation < 1% from 2006M1 through 2008M8; 18 such countries identified.

### III. Results — Risk premia: EMBI spreads and 5-year CDS spreads
- EMBI spreads (emerging market balanced panel n=29, 12 IT countries; data through December 2009):
  - EMBI spreads rose sharply at crisis outset for both groups; increases were sharper for non-IT countries.
  - In November 2008, average EMBI spread increase: IT countries about 400 basis points wider than two months before; non-IT emerging markets about 700 basis points.
  - EMBI spreads for IT emerging markets began to narrow in November 2008; non-IT spreads widened until December 2008 (more than 800 basis points above pre-crisis levels).
  - EMBI spreads for both groups returned to August 2008 levels by October 2009.
- 5-year sovereign CDS spreads (n=49, 19 IT countries; through December 2009):
  - CDS spreads increased across the board in October 2008.
  - Starting November 2008, average CDS spread for IT countries at least 75 basis points lower than for other countries; difference peaked at 270 basis points in December 2008.
  - Median CDS spreads suggest difference driven by large increases in outliers: Argentina, Ukraine, Republica Bolivariana de Venezuela.

### III. Results — Labor market, industrial production, and GDP growth
- Unemployment rates (n=50, 22 IT countries; data through September 2009):
  - Initially similar across frameworks; unemployment spiked everywhere.
  - Up to first 2 months of 2009, unemployment rose more for median IT country.
  - From then on, unemployment stabilized for IT countries; by June 2009 a statistically significant difference emerged favoring IT countries: unemployment was increasing annually by about ½ percentage point less than other countries.
  - Effect driven by advanced economies sample; for emerging markets (n=14, 9 IT countries) no IT effect found.
  - Results qualitatively similar when using seasonally adjusted unemployment levels, but with less statistical significance.
- Industrial production (n=55, 21 IT countries; data through September 2009):
  - In December 2008, both IT and non-IT countries had industrial production on average 9 percent lower than at crisis outset.
  - By September 2009, non-IT countries’ loss in industrial output was about 9 percent of August 2008 levels; IT countries’ loss was about 5½ percent.
  - Performance advantage largely driven by the advanced countries subsample (n=26, 7 IT countries: Canada, Czech Republic, Israel, Republic of Korea, Norway, Sweden, United Kingdom).
  - No statistical difference in industrial production performance between IT and non-IT emerging countries.
- GDP growth (n=49, 21 IT countries; data 2001 through 2009Q2):
  - Use year-over-year quarter growth rates.
  - Pattern: IT countries do relatively better when world economy is doing worse (post-9/11 and post-2008).
  - Median IT vs non-IT differential positive in aftermath of financial turmoil; small advantage for non-IT countries in pre-crisis boom period.
  - No causal claim; differences may reflect other country characteristics or omitted variables.

### IV. Robustness and other considerations
- Time aggregation (semiannual data: sample of 51 countries, 2003H1 to 2009H1):
  - Marginally significant difference in GDP growth in 2009 favoring IT countries, concentrated among advanced IT countries.
  - Industrial production effect in favor of IT countries economically significant (~7 percentage points) and marginally significant for full sample and non-advanced sub-sample.
  - No statistically significant difference identified for unemployment rates.
- Omitted variables and cross-section regressions (growth 2008H1-2009H1):
  - Controls considered: short-term external debt/GDP ratio; openness (average exports+imports to GDP, 2003-2007); change in commodity terms of trade; average GDP growth of trading partners.
  - Column 1 (sample of 56 countries): positive but statistically insignificant relation between IT and growth in 2009H1.
  - Column 2 (split sample): any positive effect driven by subset of advanced countries.
  - Column 3 (non-advanced countries): replicate Blanchard and Faruqee (2010) result — an additional 10 percentage points of GDP in short-term external debt in 2007 implies on average a 3 percent reduction in growth in 2009H1 (coefficient: -0.306 with standard error [0.091]***).
  - Short-term external debt/GDP ratio accounts for more variation in crisis aftermath than other variables analyzed.
  - Column 4: openness to trade has a negative effect (marginally insignificant) controlling for IT; changes in commodity terms of trade and trading partners’ GDP growth not significant once controlling for IT.

### V. Conclusion
- Summary assessment:
  - IT countries lowered nominal policy rates by more than other countries; this translated into an even larger differential in real interest rates.
  - IT countries on average better avoided persistent deflation.
  - Flexible exchange rate regimes associated with IT saw sharp real depreciations not accompanied by greater market risk perceptions.
  - Some weak evidence of better outcomes on unemployment rates and stronger industrial production performance among advanced IT countries.
  - Advanced IT countries showed higher GDP growth rates than non-IT peers; no such difference for emerging countries or full sample once controls (e.g., short-term external debt to GDP) are included.
- Caveat: results are descriptive and not causal; differences between IT and non-IT may reflect other country characteristics or omitted policies.

*Source: _wp1045 - 2.  OLS Regressions for Cross-Section of Countries (IMF working paper content provided).*

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

### _wp1045 - References

### Figures
- 1. (a) Median Policy Rate for IT and Non –IT Countries.................................................8  
  (b) Difference in the Time Effects of Policy Rates for IT and Non-IT Countries.................8
- 2. (a) Median Real Policy Rates for IT and Non-IT Countries..........................................8  
  (b) Difference in the Time Effects of Real Policy Rates for IT and Non-IT Countries............8
- 3. (a) Median Monthly Annualized Inflation for IT and Non-IT Countries   ........................9  
  (b) Probability of 3 Month Deflation....................................................................9
- 4. (a) Changes in Log REER for IT and Non –IT Countries, Full Sample ............................10  
  (b) Changes in Log REER for IT and Non –IT Countries, Emerging Market Countries........ 10
- 5. (a) Median EMBI Spread for IT and Non –IT Countries  .......................................... 11  
  (b) Difference Between Time Effects of IT and Non-IT Countries................................11
- 6. (a) Median 5-Year CDS Spreads for IT and Non –IT Countries   .................................12  
  (b) Difference in the Time Effects of CDS premia for IT and Non-IT Countries................12
- 7. (a) Median Unemployment Rates for IT and Non –IT Countries     ..............................13  
  (b) Difference in the Time Effects of Unemployment for IT and Non-IT Countries   .......... 13
- 8. (a) Median Growth in Industrial Production since Jan-06, for IT and Non-IT Countries  ......14  
  (b) Difference in Time Effects of Industrial Production for IT and Non-IT Countries..........14
- 9. (a) Median GDP Growth Rate since 2001Q1 for IT and Non-IT Countries.......................15  
  (b) Difference in Time Effects of GDP Growth Rate for IT and Non-IT Countries..............15

### Tables
- 1.  OLS  Regressions with Time and Country Effects ...................................................................20

*Source: _wp1045 - References*

### 2.  OLS Regressions for Cross-Section of Countries .....................................................................

### 2.  OLS Regressions for Cross-Section of Countries

### I. INTRODUCTION
- Purpose: evaluate performance of the inflation targeting (IT) monetary framework during the global sudden stop initiated in mid-September 2008; focus on uncovering stylized facts rather than causal identification.
- Framing:
  - IT may be “conservative window dressing” (Anna Schwartz interpretation via Romer, 2006) — no difference between IT and non-IT if both commit to low inflation.
  - Counterarguments: IT’s emphasis on credibility and communication matters for inflation expectations; IT is typically accompanied by (somewhat) flexible exchange rates which may increase resilience to external shocks.
- Criticisms of IT noted:
  - Narrow focus on inflation may neglect unemployment and financial stability (Stiglitz, 2008; Buiter, 2009).
  - Potential vulnerabilities similar to flexible exchange rate regimes in presence of liability dollarization (Calvo and Reinhart, 2002).
- Potential advantages of IT during the 2008-style crisis:
  - Credibility helps avoid deflation and a liquidity trap; quote from Governor Carney: “Just as inflation targeting has proven its ability to prevent the entrenchment of high and volatile inflation, it also has the power to prevent the onset of persistent deflation.”
  - Credibility allows emerging market IT adopters greater scope to ease policy without compromising inflation outlooks (Ghosh and others, 2009).
  - IT countries tended to have higher interest rates during the pre-crisis expansion, providing more room for rate cuts when crisis hit.
  - Correlation between IT and flexible exchange rates, which act as shock-absorbers (Broda, 2004; Edwards and Levy Yeyati, 2005; Mendoza, 1995).

### II. METHODOLOGY
- Empirical framework: panel regressions with country fixed effects (λi) and time effects (λt); main specification:
  - y_it = λ_i + λ_t + φ_t × IT_it + ε_it
  - φ_t measures difference in variable y_it between IT and non-IT countries.
- Estimation approaches: OLS for mean IT effect; median (quantile) regressions for median effects; examine coefficients of time dummies and IT × time interactions.
- Event study timing:
  - Time zero: August 2008 (last period before crisis intensifies following Lehman failure in mid-September 2008).
  - Monthly variables sample starts January 2006; quarterly variables sample starts 2002Q1.
- Sample selection and exclusions:
  - Exclude countries with nominal GDP in dollars (WEO variable ngdpd) in 2002 less than USD 10 billion.
  - Excluded: Zimbabwe (outlier), Angola, Qatar, Sudan, United Arab Emirates (constant inflation series in IMF/INS dataset), and United States (source of initial shock).
  - Full sample after exclusions: 84 countries.
  - Iceland excluded due to GDP cutoff despite severe crisis and suspension of IT; Latvia also excluded by same GDP cutoff.
- Outcome variables examined (in order): nominal policy rates, real policy rates, inflation, real effective exchange rate (REER), EMBI spreads, 5-year CDS premia, unemployment rates, industrial production, GDP growth.
- Emerging markets defined as countries with available J.P. Morgan EMBI spread data.

### III. RESULTS
- Policy (policy rates; sample 49 countries with 21 IT countries; data through November 2009):
  - Both IT and non-IT countries increased nominal rates from early 2008 to August 2008; IT countries tightened slightly more pre-crisis.
  - In September 2008, 6 out of 22 IT countries tightened policy rates (Brazil, Chile, Indonesia, Israel, Peru, Sweden).
  - After crisis deepened, IT countries cut policy rates by 2 percentage points more than non-IT countries.
  - Difference in mean policy rates across groups was statistically significant and persistent.
- Real policy rates (calculated using 12-month backward-looking inflation; n unspecified in this section for real rates):
  - IT countries avoided an increase in real policy rates since the crisis.
  - Loosening of real policy rates for IT countries relative to non-IT countries was about 4 percentage points.
- Inflation (n=84, of which 25 IT countries; data through November 2009):
  - Price levels dropped across the board in the last two months of 2008.
  - Average annualized inflation rate was lower by 5 percent for IT countries and 6½ percent for non-IT countries in December 2009 relative to August 2009.
  - Median IT and non-IT countries had negative monthly inflation in December 2009.
  - IT countries were less likely to experience a deflation scare (three consecutive months of negative monthly inflation); statistically significant differences for 2008M12-2009M1 and 2009M6.
  - Footnote: In August 2009, 2 IT countries had 3 consecutive negative monthly inflation readings (Chile and Republic of Serbia); 7 non-IT countries did (Bulgaria, China, P.R.: Mainland, Ireland, Japan, Libya, Lithuania and Taiwan Province of China).
- Real effective exchange rate (REER) (n=84, 25 IT countries; data through August 2009):
  - REER depreciation for the average IT country was more than 15 percentage points by 2009Q1 (base period August 2008).
  - By August 2009, IT countries on average were about 5 percent weaker than at crisis outset.
  - Finding holds for full sample of 89 countries, robust to subset of 31 emerging markets, and to excluding countries with pegs/heavily managed floats against the U.S. dollar pre-crisis.
  - Among 85-country universe with 25 IT countries, the 12 countries with largest real depreciations six months into the crisis (February 2009) were all IT countries; listed depreciations including Poland at -31½ percent, and Turkey at -17½ percent among others.
  - No IT countries among top 30 countries with real appreciations during that period (e.g., Japan appreciated by 26½ percent; Venezuela, Rep. Bol. by 24 percent).
  - Definition note: pegs/heavily managed floats = monthly exchange rate standard deviation < 1% from 2006M1 through 2008M8; 18 such countries identified.
- Risk premia: EMBI spreads (emerging market balanced panel n=29, 12 IT countries; data through December 2009):
  - EMBI spreads rose sharply at crisis outset for both groups, but increases were sharper for non-IT countries.
  - In November 2008, average EMBI spread increase: IT countries about 400 basis points wider than two months before; non-IT emerging markets about 700 basis points.
  - EMBI spreads for IT emerging markets began to narrow in November 2008; non-IT spreads widened until December 2008 (more than 800 basis points above pre-crisis levels).
  - EMBI spreads for both groups returned to August 2008 levels by October 2009.
- 5-year sovereign CDS spreads (n=49, 19 IT countries; through December 2009):
  - CDS spreads increased across the board in October 2008.
  - Starting November 2008, average CDS spread for IT countries at least 75 basis points lower than for other countries; difference peaked at 270 basis points in December 2008.
  - Median CDS spreads suggest difference driven by large increases in outliers: Argentina, Ukraine, Republica Bolivariana de Venezuela.
- Unemployment rates (n=50, 22 IT countries; data through September 2009):
  - Initially similar across monetary frameworks; unemployment spiked everywhere.
  - Up to first 2 months of 2009, unemployment rose more for median IT country.
  - From then on, unemployment stabilized for IT countries; by June 2009 a statistically significant difference emerged favoring IT countries: unemployment was increasing annually by about ½ percentage point less than other countries.
  - Effect driven by advanced economies sample; for emerging markets (n=14, 9 IT countries) no IT effect found.
  - Note: results qualitatively similar when using seasonally adjusted unemployment levels, but with less statistical significance.
- Industrial production (n=55, 21 IT countries; data through September 2009):
  - In December 2008, both IT and non-IT countries had industrial production on average 9 percent lower than at crisis outset.
  - By September 2009, non-IT countries’ loss in industrial output was about 9 percent of August 2008 levels; IT countries’ loss was about 5½ percent.
  - Performance advantage largely driven by the advanced countries subsample (n=26, 7 IT countries: Canada, Czech Republic, Israel, Republic of Korea, Norway, Sweden, United Kingdom).
  - No statistical difference in industrial production performance between IT and non-IT emerging countries.
- GDP growth (n=49, 21 IT countries; data 2001 through 2009Q2):
  - Use year-over-year quarter growth rates to avoid seasonality.
  - Pattern: IT countries do relatively better when world economy is doing worse (post-9/11 and post-2008).
  - Median IT vs non-IT differential positive in aftermath of financial turmoil; small advantage for non-IT countries in pre-crisis boom period.
  - Cannot claim causality; differences may reflect other country characteristics or omitted variables.

### IV. ROBUSTNESS AND OTHER CONSIDERATIONS
- Time aggregation:
  - Using semiannual data (sample of 51 countries, 2003H1 to 2009H1) with double fixed effects and robust standard errors:
    - Marginally significant difference in GDP growth in 2009 favoring IT countries, concentrated among advanced IT countries.
    - Industrial production effect in favor of IT countries economically significant (~7 percentage points) and marginally significant for full sample and non-advanced sub-sample.
    - No statistically significant difference identified for unemployment rates.
- Omitted variables:
  - Concern: IT adoption is not random; IT countries may share common risk factors.
  - Variables known to correlate with GDP contractions (Blanchard and Faruqee, 2010) included in robustness checks:
    - Short-term external debt to GDP ratio.
    - Degree of openness to trade (average exports+imports to GDP, 2003-2007).
    - Change in commodity terms of trade.
    - Average GDP growth of trading partners.
  - Regressions of growth in first semester of 2009 (relative to first semester previous year):
    - Column 1 (sample of 56 countries): positive but statistically insignificant relation between IT and growth in 2009H1.
    - Column 2 (split sample): any positive effect driven by subset of advanced countries.
    - Column 3 (non-advanced countries): replicate Blanchard and Faruqee (2010) result — an additional 10 percentage points of GDP in short-term external debt in 2007 implies on average a 3 percent reduction in growth in 2009H1.
    - Short-term external debt/GDP ratio accounts for more variation in crisis aftermath than other variables analyzed.
    - Column 4: openness to trade has a negative effect (marginally insignificant) controlling for IT; changes in commodity terms of trade and trading partners’ GDP growth not significant once controlling for IT.

### V. CONCLUSION
- Summary assessment: IT has had a positive scorecard thus far in the crisis period analyzed.
  - IT countries lowered nominal policy rates by more than other countries; this translated into an even larger differential in real interest rates.
  - IT countries on average better avoided persistent deflation.
  - Flexible exchange rate regimes associated with IT saw sharp real depreciations not accompanied by greater market risk perceptions.
  - Some weak evidence of better outcomes on unemployment rates and stronger industrial production performance among advanced IT countries.
  - Advanced IT countries showed higher GDP growth rates than non-IT peers; no such difference for emerging countries or full sample once controls (e.g., short-term external debt to GDP) are included.
- Caveat: results are descriptive and not causal; differences between IT and non-IT may reflect other country characteristics or omitted policies.

*Source: _wp1045 - 2.  OLS Regressions for Cross-Section of Countries (IMF working paper content provided).*

### REFERENCES

### REFERENCES

### Citations and literature
- Ball L, Sheridan N. Does inflation targeting matter? In: Bernanke B, Woodford M (Eds), The inflation targeting debate. The University of Chicago Press: Chicago; 2005. p. 249-276.
- Batini, Nicoletta, Kenneth Kuttner and Doug Laxton, 2005. “Does Inflation Targeting Work in Emerging Markets?,” IMF World Economic Outlook, September 2005.
- Batini, Nicoletta and Doug Laxton, 2007. “Under What Conditions Can Inflation Targeting Be Adopted? The Experience of Emerging Markets,” In: Mishkin F. and Schmidt-Hebbel, K. (Eds.), Monetary Policy Under Inflation Targeting, Central Bank of Chile: Santiago, pp. 1-38.
- Blanchard, Olivier and Jordi Galí, 2007, “Real Wage Rigidities and the New Keynesian Model,” Journal of Money, Credit and Banking, Supplement to Vol. 39, No. 1 (February 2007), pp. 35-65.
- Blanchard, Olivier, 2009, “The Crisis: Basic Mechanisms and Appropriate Policies,” IMF Working Paper 09/80.
- Blanchard, Olivier and Hamid Faruqee, 2010, “The Impact Effect of the Crisis on Emerging Market Countries,” mimeo.
- Blanchard, Olivier, Giovanni dell’Ariccia and Paolo Mauro, 2010, “Rethinking Macroeconomic Policy”, IMF Staff Position Note 10/03.
- Brito, Ricardo D. and Brianne Bystedt, 2009, “Inflation Targeting in Emerging Economies: Panel Evidence,” Journal of Development Economics, forthcoming.
- Broda, Christian, 2004, “Terms of Trade and Exchange Rate Regimes in Developing Countries,” Journal of International Economics, vol. 63, pp. 31-58.
- Buiter, Willem, 2009, “The Unfortunate Uselessness of Most ’State of the Art’ Academic Monetary Economics,” in http://www.voxeu.org/index.php?q=node/3210
- Calvo, Guillermo A. and Carmen M. Reinhart, 2002, “Fear of Floating,” Quarterly Journal of Economics, v. 107(2,May), pp. 379-408.
- Decressin, Jorg and Douglas Laxton, 2009, “Gauging Risks for Deflation,” IMF Staff Position Note, SPN/09/01.
- Ghosh, Atish R., Marcos Chamon, Christopher Crowe, Jun I. Kim, and Jonathan D. Ostry, 2009, “Coping with the Crisis: Policy Options for Emerging Market Countries,” IMF Staff Position Note 09/08.
- Gonçalves, Carlos Eduardo and João Salles, 2008, “Inflation Targeting in Emerging Economies: What Do the Data Say?” Journal of Development Economics, vol. 85, pp. 312-318.
- Edwards, Sebastian and Eduardo Levy Yeyati, 2005, “Flexible exchange rates as shock absorbers,” European Economic Review, vol. 49, pp. 2079-2105.
- Friedman, Charles and Douglas Laxton, 2009, “Why inflation targeting?” IMF Working Paper 09/86.
- Habermeier , Karl, İnci Ötker-Robe, Luis Jacome, Alessandro Giustiniani, Kotaro Ishi, David Vávra, Turgut Kışınbay, and Francisco Vazquez, 2009, “Inflation Pressures and Monetary Policy Options in Emerging and Developing Countries: A Cross Regional Perspective,” IMF Working Paper WP/09/1.
- Eggertsson, Gauti and Michael Woodford, 2003, “The Zero Bound on Interest Rates and Optimal Monetary Policy,” Brookings Papers on Economic Activity, Vol. 2003, No. 1 (2003), pp. 139-211.
- Mendoza, E.G., 1995, “The Terms of Trade, the Real Exchange Rate, and Economic Fluctuations,” International Economic Review, vol. 36, pp. 101-137.
- Roger, Scott, 2009, “Inflation Targeting at 20: Achievements and Challenges,” IMF Working Paper 09/236.
- Romer, David, 2006, “Advanced Macroeconomics”, McGraw-Hill: Irwin, New York.
- Stiglitz, Joseph, 2008, “The Failure of Inflation Targeting,” in http://www.voxeu.org/index.php?q=node/2549
- Stone, Mark, Scott Roger, Seiichi Shimizu, Anna Nordstrom, Turgut Kisinbay, and Jorge Restrepo, 2009, IMF Occasional Paper 267, mimeo.

### Regression tables — key reported statistics
- Table 1. OLS Regressions with time and country effects: 2003H1-2009H1
  - Dependent variables indicated: GDP growth, Unemployment, Ind. Production.
  - Regressors shown (selected):
    - IT x 2009 coefficients: 0.02, 0.07, -0.236 with t-statistics [1.72], [2.41]*, [0.35]
    - IT x 2009 x Not advanced coefficients: 0.009, 0.074, -0.720 with t-statistics [0.63], [1.98]*, [0.85]
    - IT x 2009 x Advanced coefficients: 0.033, 0.065, 0.343 with t-statistics [2.86] **, [1.89], [0.54]
  - Sample and fit:
    - Observations: 642, 642, 640, 640, 642, 642 (as listed)
    - Number of countries: 51 repeated across columns
    - R-squared: 0.61, 0.61, 0.54, 0.54, 0.36, 0.36
  - Note: Robust t statistics in brackets; * significant at 5%; ** significant at 1%.

- Table 2. OLS Regression for cross-section of countries (dependent variable: Growth 2008H1-2009H1)
  - Regressors and coefficients (columns 1–6 as listed):
    - Inflation targeting: 0.011, -0.018 with standard errors [0.012], [0.020]
    - Inflation targeting x Advanced: 0.024, 0.018, 0.032, 0.024 with standard errors [0.017], [0.018], [0.016]*, [0.017]
    - Inflation targeting x Not advanced: -0.009, -0.012, -0.026, -0.009 with standard errors [0.017], [0.017], [0.020], [0.018]
    - Short-term external debt/GDP, 2007: -0.306 [0.091]***
    - Openness/GDP, 2003-2007 average: -0.013 [0.008]
    - Change in terms of trade, 2008-09: 0.441 [0.436]
    - Growth 2008H1-2009H1, trading partners: -0.004 [0.756]
    - Advanced indicator coefficients: -0.035, -0.030, -0.062, -0.035 with standard errors [0.016]**,[0.016]*,[0.021]***,[0.016]**
  - Observations across columns: 56, 56, 26, 56, 47, 56 (as listed)
  - R-squared: 0.015, 0.1, 0.334, 0.138, 0.217, 0.1
  - Notes: Standard errors in brackets. Constant not reported. * significant at 10%; ** significant at 5%; *** significant at 1%.

### Data appendix — sample definitions, country lists, and data sources
- Sample construction and exclusions:
  - Full sample excludes countries with nominal GDP in dollars (variable ngdpd in the WEO) in 2002 less than USD 10 billion; excludes Zimbabwe; excludes Angola, Qatar, Sudan and United Arab Emirates because their inflation rates at the INS dataset are constant over several months; and excludes the United States.
  - This leaves a full sample of 84 countries.
  - Iceland is excluded because it does not meet the nominal GDP criterion.
- Full sample of 84 countries (as listed): Algeria, Argentina, Australia, Austria, Bangladesh, Belarus, Belgium, Brazil, Bulgaria, Cameroon, Canada, Chile, China, P.R.: Mainland, China, P.R.: Hong Kong, Colombia, Costa Rica, Croatia, Cyprus, Czech Republic, Côte d'Ivoire, Denmark, Dominican Republic, Ecuador, Egypt, El Salvador, Finland, France, Germany, Greece, Guatemala, Hungary, India, Indonesia, Iran, I.R. of, Ireland, Israel, Italy, Japan, Kazakhstan, Kenya, Korea, Republic of, Kuwait, Lebanon, Libya, Lithuania, Luxembourg, Malaysia, Mexico, Morocco, Netherlands, New Zealand, Nigeria, Norway, Oman, Pakistan, Panama, Peru, Philippines, Poland, Portugal, Romania, Russian Federation, Saudi Arabia, Serbia, Republic of, Singapore, Slovak Republic, Slovenia, South Africa, Spain, Sri Lanka, Sweden, Switzerland, Syrian Arab Republic, Taiwan Prov. of China, Tanzania, Thailand, Tunisia, Turkey, Ukraine, United Kingdom, Uruguay, Venezuela, Rep. Bol., Vietnam, Republic of Yemen.
- Inflation targeting (IT) classification:
  - There are 25 IT countries in the sample.
  - IT countries (based on Friedman and Laxton (2009) and Roger (2009)): Australia, Brazil, Canada, Chile, Colombia, Czech Republic, Guatemala, Hungary, Indonesia, Israel, Republic of Korea, Mexico, New Zealand, Norway, Peru, Philippines, Poland, Romania, Republic of Serbia, South Africa, Sweden, Switzerland, Thailand, Turkey, United Kingdom.
  - 4 countries with fully-fledged inflation targeting (publish their endogenous interest rate forecast): Czech Republic, New Zealand, Norway and Sweden.
- Emerging market countries definition:
  - Emerging market countries (n=31, of which 12 IT countries, with data through December 2009) are those for which J.P. Morgan EMBI spread could be gathered: Argentina, Brazil, Bulgaria, Chile, China, P.R.: Mainland, Colombia, Dominican Republic, Ecuador, Egypt, El Salvador, Hungary, Indonesia, Kazakhstan, Lebanon, Malaysia, Mexico, Pakistan, Panama, Peru, Philippines, Poland, Russian Federation, Republic of Serbia, South Africa, Sri Lanka, Tunisia, Turkey, Ukraine, Uruguay, Rep. Bol. Venezuela, and Vietnam.
- Definition of advanced economies:
  - Same as the WEO (n=30): Australia, Austria, Belgium, Canada, China, P.R.: Hong Kong, Cyprus, Czech Republic, Denmark, Finland, France, Germany, Greece, Ireland, Israel, Italy, Japan, Republic of Korea, Luxembourg, Netherlands, New Zealand, Norway, Portugal, Singapore, Slovak Republic, Slovenia, Spain, Sweden, Switzerland, Taiwan Prov. of China, and United Kingdom.
  - Note: United States excluded; Czech Republic belongs to both advanced and emerging samples.
- Exchange rate and inflation data:
  - REER, NEER and inflation rate come from the IMF/INS database.
  - The REER is based on trade-weighted CPI indexes.
- 5-year sovereign CDS spread data:
  - Coverage n=49 (of which 19 IT countries); obtained from Datastream; coverage through December 2009.
  - Countries listed for CDS: Argentina, Australia, Austria, Belgium, Brazil, Bulgaria, Canada, Chile, China, P.R.: Mainland, China, P.R.: Hong Kong, Colombia, Croatia, Czech Republic, Denmark, France, Germany, Greece, Hungary, Indonesia, Ireland, Israel, Italy, Japan, Kazakhstan, Korea, Republic of, Lebanon, Lithuania, Malaysia, Morocco, Netherlands, Norway, Panama, Peru, Philippines, Poland, Portugal, Romania, Russian Federation, Slovak Republic, Slovenia, South Africa, Spain, Sweden, Thailand, Tunisia, Turkey, Ukraine, Venezuela and Vietnam.
- Unemployment rate data:
  - n=50 (of which 22 IT countries; data through September 2009).
  - Sources: IMF/IFS, IMF/GDS, Haver and Datastream/Eurostat.
  - Country lists by source provided as in the appendix.
- Industrial production data:
  - n=55 (of which 21 IT countries; data through September 2009).
  - Sources: IFS, Haver or Datastream (country lists provided in appendix).
- Policy rate data:
  - n=47 (of which 21 IT countries) from the IMF/GDS database.
  - Country coverage listed in the appendix.

*Source: _wp1045 - REFERENCES (PDF chapter/section).*

---


_Source: https://www.imf.org/-/media/websites/imf/imported-full-text-pdf/external/pubs/ft/wp/2010/_wp1045.pdf_