## _wp12262

## Source details

**Canonical URL:** [_wp12262](https://www.imf.org/-/media/websites/imf/imported-full-text-pdf/external/pubs/ft/wp/2012/_wp12262.pdf)

## Other formats

- [Markdown version](/-/media/websites/imf/imported-full-text-pdf/external/pubs/ft/wp/2012/_wp12262.pdf.md)
- [Structured JSON version](/-/media/websites/imf/imported-full-text-pdf/external/pubs/ft/wp/2012/_wp12262.pdf.json)

---

### Objectives and scope
- Measure how the sensitivity ("beta") of Asian equity markets to systemic economies varies across time and countries using an extended CAPM framework.
- Two complementary approaches:
  - Approach I (cross-sectional): two-stage factor model to estimate country-level betas (constant over the sample) and explain cross-country differences via bilateral trade, FDI, banking linkages, and macroeconomic policy measures.
  - Approach II (time-series/panel): time-varying beta specification for a panel of 42 economies (13 Asian + other OECD and emerging markets) to study low-frequency trends and high-frequency fluctuations and to quantify determinants of beta dynamics.
- Sample coverage:
  - Monthly panel for Approach II: 1990–2011 (42 economies, excess returns over 3-month US T-bill).
  - First-stage monthly data in Approach I: starts in 1990 (ends March 2012); second-stage annual panel: 1992–2011 (full sample) and subsamples 1992–01, 2002–11, 2002–07, 2008–11.

### Methodology — key model features
- Approach I first-stage factor model (extended CAPM) controls for five systemic countries (France, Germany, Japan, United Kingdom, United States; C=5) and five global factors (G=5): global commodity price fluctuations, global growth (PPP-weighted industrial production of the five systemic economies), global equity market residual, international interest rates, and VIX.
- Approach I second-stage explains estimated betas (panel of 60 betas) as functions of bilateral TRADE, FDI, BANK (all scaled by country i GDP), and policy variables (reserves scaled by GDP or M2, fiscal metrics in percent of GDP, exchange rate regime, capital account openness).
- Approach II directly models time-varying beta:
  - Excess return equation with interaction terms between US equity return and observable country-specific factors X_it and global factors Z_t.
  - General-to-specific model selection at 10 percent confidence level to identify parsimonious determinants.
- Common country-specific linkage variables: TRADE, BANK, FDI, trade openness (exports+imports to GDP).
- Common global factors used: VIX and yield curve slope (10y minus 2y).

### Main empirical findings — high-level conclusions
- Three principal conclusions:
  - Regional financial markets in Asia have become more sensitive to systemic economies over time, reflecting increasing financial and trade integration.
  - Asian financial sensitivities exhibit strong cyclical fluctuations tied to international capital market conditions; betas peaked at historically high levels during 2008–11.
  - Macroeconomic policies—lower government debt and higher international reserves (up to a limit)—are associated with lower financial betas, but cannot fully insulate Asian markets from major global shocks.

### Quantitative results — selected coefficients and statistics
- Representative estimation (Approach II, Pure CAPM, GS 3 column, Table 3):
  - Coefficient on US equity return (ݎ_US): 0.616*** (0.000)
  - Trade: 0.001*** (0.000)
  - Banking: -0.018*** (0.000)
  - FDI: 0.021** (0.018)
  - Fiscal: 0.006* (0.076)
  - Reserves: -0.003*** (0.006)
  - Debt/GDP: 0.002*** (0.000)
  - Reserve/Debt: -0.000** (0.025)
  - GDP: -0.013** (0.017)
  - VIX: 0.006*** (0.000)
  - R-squared: 0.27; Observations: 8082
- VIX and yield slope evidence:
  - VIX significant across specifications (examples: 0.007*** (0.000) in Table 4 GS 3; 0.007*** (0.000) in Table 5 GS 3).
  - Yield slope often significant with negative sign (example Table 4 GS 3: -0.073** (0.020)).
- ASEAN average beta (Approach II panel averages over full sample): ~0.75 (a 10 percent increase in the U.S. stock market associated with a 7½ percent increase across ASEAN stock markets, all else equal).
- Time evolution and crisis spikes:
  - Betas display a steady upward trend over two decades with strong cyclical spikes at episodes including the NASDAQ crash/2001 and the Lehman bankruptcy/2008 and European financial turbulence.
  - For all Asian economies, financial betas reached unprecedented levels during 2008–11, surpassing levels prevailing during the Asian crisis in 1997.
- Decomposition and drivers:
  - The 2008–11 pickup in betas is driven largely by the global financial shock (VIX and global factors); country-specific cyclical factors played a minor role during 2008–11 but were more important in 1997–2001.
  - Cross-country (Approach I second-stage) determinants (full sample illustrative coefficients):
    - Trade: 0.0116 (Full Sample)
    - FDI: 0.6240 (Full Sample)
    - Bank: 0.0060 (Full Sample)
    - Reserves (scaled by M2): -0.0033 (Full Sample; negative contribution)
    - Observations: 60; R-squared: 12.4% (Full Sample)
- Reserves nonlinearities (Table 5):
  - Linear reserves coefficient: -0.003* (0.076)
  - log(Reserves): -0.044* (0.082)
  - Nonlinear least squares specification: parameter -8.069*** (0.000); residual sigma 0.005*** (0.005)
  - Implication: diminishing marginal insurance benefits of reserves in lowering beta as reserves grow.

### Policy-relevant implications and recommendations (empirical)
- Macroeconomic policy space matters:
  - Lower government debt-to-GDP and higher international reserves correlate with lower financial betas.
  - Quantified example: a 10 percentage-point increase in the reserves-to-GDP ratio is associated with financial betas lower by 0.03–0.04.
- Policy strengths and limits:
  - Policies can reduce sensitivity to spillovers in both tranquil and turbulent periods, but cannot fully insulate markets from major global financial shocks (global factors explain majority of betas’ spike in 2008–11).
  - The marginal benefit of reserves diminishes beyond a threshold; nonlinear models indicate decreasing returns to additional reserves in reducing beta.
- Bilateral linkages:
  - Greater bilateral trade, FDI, and banking exposures to systemic economies are associated with higher period-average betas (Approach I).
  - Banking linkages show mixed roles: Approach II finds negative cyclical effects, while Approach I finds positive correlation with period-average beta.
- Policy mix:
  - Given the limited insulation offered by macroeconomic policies alone, an appropriate response to large global shocks may require a comprehensive mix including individual and collective financial measures.

### Robustness and caveats
- Results robust to alternative specifications of systemic economies (dropping France or the UK) and alternative combinations of global factors; estimated factor loadings are sensitive to the sample period.
- Data limitations:
  - BANK and short-term external debt data are limited for some countries; restricted models (dropping BANK or debt variables) increase observations and serve as robustness checks.
  - Exchange rate and monetary policy space are proxied indirectly; consistent cross-country monetary policy metrics are difficult to design.
- Methodological notes:
  - Approach II’s general-to-specific selection used a 10 percent acceptance threshold; specifications reported include Pure CAPM, Non-zero Expected Risk Premium, and Heterogeneous Expected Premium (fixed effects).
  - Fixed-effects and clustering adjustments applied; p-values reported in tables.

### Appendix — Estimated factor loadings (selected entries)
- Panel A. Factor model with five systemic economies (Multi5 variants, shown as comma-separated loadings per systemic country):
  - USA (Multi5): 0.25, 0.09, 0.53, 0.34, 0.40, -0.01, 0.44, 0.27, 0.27, 0.45, 0.33, 0.07
  - JPN (Multi5): 0.13, 0.21, 0.08, 0.42, 0.22, 0.29, 0.23, 0.07, 0.12, 0.24, 0.06, 0.02
  - DEU (Multi5): 0.08, 0.00, 0.26, 0.04, 0.30, 0.65, 0.37, 0.55, 0.45, 0.57, 0.01, 0.06
  - GBR (Multi5): 0.32, 0.32, 0.45, 0.76, 0.34, 0.06, 0.15, 0.30, 0.31, 0.75, 0.17, 0.05
  - FRA (Multi5): -0.03, 0.05, -0.14, -0.30, -0.20, 0.06, -0.11, -0.40, -0.25, -0.75, -0.08, 0.48
- Panel B. Factor model with three systemic economies (Multi3 variants):
  - USA (Multi3): 0.39, 0.25, 0.70, 0.62, 0.51, 0.03, 0.48, 0.32, 0.35, 0.63, 0.39, 0.20
  - JPN (Multi3): 0.15, 0.24, 0.11, 0.45, 0.23, 0.30, 0.23, 0.06, 0.12, 0.23, 0.06, 0.06
  - DEU (Multi3): 0.15, 0.12, 0.29, 0.05, 0.26, 0.71, 0.33, 0.37, 0.37, 0.28, 0.01, 0.39

*Source: _wp12262 - References .............................................................................................................*

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

### _wp12262 - References .............................................................................................................

### Figures
- 1. Equity Returns: Selected Systemic Economies and Asia ......................................................3
- 2. Asian Financial Betas and Global Financial Shocks ...........................................................17
- 3. Financial Betas across Asian Economies .............................................................................18
- 4. Contributions to Differences in Betas between 2008–2011 and 2002–2007 .......................19
- 5. Contributions to Differences in Betas between 1997–2001 and 2002–2007 .......................19
- 6. The Decomposition of Financial Betas across Asia over Time ...........................................23

### Tables
- 1. Pure and CAPM ...................................................................................................................14
- 1. Non-Zero Expected Risk Premium  .....................................................................................15
- 2. Heterogeneous Expected Premium (Fixed Effects) .............................................................16
- 4. Determinants of Asian Financial Betas ................................................................................22
- 5. Nonlinear Effects of Reserves .............................................................................................23

*Source: _wp12262 - References .............................................................................................................*

### Appendix Table: Estimated Factor Loadings (Betas) Over the Full Sample  ..........................25

### Appendix Table: Estimated Factor Loadings (Betas) Over the Full Sample  ..........................25

### Objectives and scope
- Measure how the sensitivity ("beta") of Asian equity markets to systemic economies varies across time and countries using an extended CAPM framework.
- Two complementary approaches:
  - Approach I (cross-sectional): two-stage factor model to estimate country-level betas (constant over the sample) and explain cross-country differences via bilateral trade, FDI, banking linkages, and macroeconomic policy measures.
  - Approach II (time-series/panel): time-varying beta specification for a panel of 42 economies (13 Asian + other OECD and emerging markets) to study low-frequency trends and high-frequency fluctuations and to quantify determinants of beta dynamics.
- Sample coverage:
  - Monthly panel for Approach II: 1990–2011 (42 economies, excess returns over 3-month US T-bill).
  - First-stage monthly data in Approach I: starts in 1990 (ends March 2012); second-stage annual panel: 1992–2011 (full sample) and subsamples 1992–01, 2002–11, 2002–07, 2008–11.

### Methodology — key model features
- Approach I first-stage factor model (extended CAPM) controls for five systemic countries (France, Germany, Japan, United Kingdom, United States; C=5) and five global factors (G=5): global commodity price fluctuations, global growth (PPP-weighted industrial production of the five systemic economies), global equity market residual, international interest rates, and VIX.
- Approach I second-stage explains estimated betas (panel of 60 betas) as functions of bilateral TRADE, FDI, BANK (all scaled by country i GDP), and policy variables (reserves scaled by GDP or M2, fiscal metrics in percent of GDP, exchange rate regime, capital account openness).
- Approach II directly models time-varying beta:
  - Excess return equation with interaction terms between US equity return and observable country-specific factors X_it and global factors Z_t.
  - General-to-specific model selection at 10 percent confidence level to identify parsimonious determinants.
- Common country-specific linkage variables: TRADE, BANK, FDI, trade openness (exports+imports to GDP).
- Common global factors used: VIX and yield curve slope (10y minus 2y).

### Main empirical findings — high-level conclusions
- Three principal conclusions:
  - Regional financial markets in Asia have become more sensitive to systemic economies over time, reflecting increasing financial and trade integration.
  - Asian financial sensitivities exhibit strong cyclical fluctuations tied to international capital market conditions; betas peaked at historically high levels during 2008–11.
  - Macroeconomic policies—lower government debt and higher international reserves (up to a limit)—are associated with lower financial betas, but cannot fully insulate Asian markets from major global shocks.

### Quantitative results — selected coefficients and statistics (preserving reported values)
- Representative estimation (Approach II, Pure CAPM, GS 3 column, Table 3):
  - Coefficient on US equity return (ݎ_US): 0.616*** (0.000)
  - Trade: 0.001*** (0.000)
  - Banking: -0.018*** (0.000)
  - FDI: 0.021** (0.018)
  - Fiscal: 0.006* (0.076)
  - Reserves: -0.003*** (0.006)
  - Debt/GDP: 0.002*** (0.000)
  - Reserve/Debt: -0.000** (0.025)
  - GDP: -0.013** (0.017)
  - VIX: 0.006*** (0.000)
  - R-squared: 0.27; Observations: 8082
- Non-zero expected risk premium and heterogeneous premium specifications (Tables 4–5) corroborate the significance of VIX and yield slope:
  - VIX significant across specifications (examples: 0.007*** (0.000) in Table 4 GS 3; 0.007*** (0.000) in Table 5 GS 3).
  - Yield slope often significant with negative sign (example Table 4 GS 3: -0.073** (0.020)).
- ASEAN average beta (Approach II panel averages over full sample): ~0.75 (interpreted as: a 10 percent increase in the U.S. stock market associated with a 7½ percent increase across ASEAN stock markets, all else equal).
- Time evolution and crisis spikes:
  - Betas display a steady upward trend over two decades with strong cyclical spikes at episodes including the NASDAQ crash/2001 and the Lehman bankruptcy/2008 and European financial turbulence.
  - For all Asian economies, financial betas reached unprecedented levels during 2008–11, surpassing levels prevailing during the Asian crisis in 1997.
- Decomposition and drivers:
  - The 2008–11 pickup in betas is driven largely by the global financial shock (VIX and global factors); country-specific cyclical factors played a minor role during 2008–11 but were more important in 1997–2001.
  - Cross-country (Approach I second-stage) determinants (Table 4) — full sample estimated coefficients shown for illustrative determinants:
    - Trade: 0.0116 (Full Sample)
    - FDI: 0.6240 (Full Sample)
    - Bank: 0.0060 (Full Sample)
    - Reserves (scaled by M2): -0.0033 (Full Sample; negative contribution)
    - Observations: 60; R-squared: 12.4% (Full Sample)
- Reserves nonlinearities (Table 5):
  - Linear reserves coefficient: -0.003* (0.076)
  - log(Reserves): -0.044* (0.082)
  - Nonlinear least squares specification: parameter -8.069*** (0.000); residual sigma 0.005*** (0.005)
  - Implication: diminishing marginal insurance benefits of reserves in lowering beta as reserves grow.

### Policy-relevant implications and recommendations (empirical)
- Macroeconomic policy space matters:
  - Lower government debt-to-GDP and higher international reserves correlate with lower financial betas.
  - Quantified example: a 10 percentage-point increase in the reserves-to-GDP ratio is associated with financial betas lower by 0.03–0.04 (reported in the text and corroborated by panel regressions).
- Policy strengths and limits:
  - Policies can reduce sensitivity to spillovers in both tranquil and turbulent periods, but cannot fully insulate markets from major global financial shocks (global factors explain majority of betas’ spike in 2008–11).
  - The marginal benefit of reserves diminishes beyond a threshold; nonlinear models indicate decreasing returns to additional reserves in reducing beta.
- Bilateral linkages:
  - Greater bilateral trade, FDI, and banking exposures to systemic economies are associated with higher period-average betas (Approach I).
  - Banking linkages show mixed roles: Approach II finds negative cyclical effects (banking inflows may substitute for capital markets and help lower beta at high frequency), while Approach I finds positive correlation with period-average beta reflecting secular exposure.
- Policy mix:
  - Given the limited insulation offered by macroeconomic policies alone, an appropriate response to large global shocks may require a comprehensive mix including individual and collective financial measures (as adopted in Asia post-crisis).

### Robustness and caveats
- Results robust to alternative specifications of systemic economies (dropping France or the UK) and alternative combinations of global factors; however, estimated factor loadings are sensitive to the sample period.
- Data limitations:
  - BANK and short-term external debt data are limited for some countries; restricted models (dropping BANK or debt variables) increase observations and serve as robustness checks.
  - Exchange rate and monetary policy space are proxied indirectly; consistent cross-country monetary policy metrics are difficult to design.
- Methodological notes:
  - Approach II’s general-to-specific selection used a 10 percent acceptance threshold; specifications reported include Pure CAPM, Non-zero Expected Risk Premium, and Heterogeneous Expected Premium (fixed effects).
  - Fixed-effects and clustering adjustments applied; p-values reported in tables.

*Source: https://www.imf.org/-/media/websites/imf/imported-full-text-pdf/external/pubs/ft/wp/2012/_wp12262.pdf*

### Appendix Table. Estimated Factor Loadings (Betas) Over the Full Sample

### Appendix Table. Estimated Factor Loadings (Betas) Over the Full Sample

### Panel A. Factor model with five systemic economies
- Multi5
  - USA: 0.25, 0.09, 0.53, 0.34, 0.40, -0.01, 0.44, 0.27, 0.27, 0.45, 0.33, 0.07
  - JPN: 0.13, 0.21, 0.08, 0.42, 0.22, 0.29, 0.23, 0.07, 0.12, 0.24, 0.06, 0.02
  - DEU: 0.08, 0.00, 0.26, 0.04, 0.30, 0.65, 0.37, 0.55, 0.45, 0.57, 0.01, 0.06
  - GBR: 0.32, 0.32, 0.45, 0.76, 0.34, 0.06, 0.15, 0.30, 0.31, 0.75, 0.17, 0.05
  - FRA: -0.03, 0.05, -0.14, -0.30, -0.20, 0.06, -0.11, -0.40, -0.25, -0.75, -0.08, 0.48

- Multi5_rw
  - USA: 0.25, 0.09, 0.53, 0.34, 0.40, -0.01, 0.44, 0.27, 0.27, 0.45, 0.31, 0.07
  - JPN: 0.13, 0.21, 0.08, 0.42, 0.22, 0.29, 0.23, 0.07, 0.12, 0.24, 0.07, 0.02
  - DEU: 0.08, 0.00, 0.26, 0.04, 0.30, 0.65, 0.37, 0.55, 0.45, 0.57, 0.02, 0.06
  - GBR: 0.32, 0.32, 0.45, 0.76, 0.34, 0.06, 0.15, 0.30, 0.31, 0.75, 0.17, 0.05
  - FRA: -0.03, 0.05, -0.14, -0.30, -0.20, 0.06, -0.11, -0.40, -0.25, -0.75, -0.08, 0.48

- Multi5_IC
  - USA: 0.24, 0.12, 0.50, 0.37, 0.38, -0.02, 0.40, 0.29, 0.26, 0.41, 0.21, -0.08
  - JPN: 0.13, 0.20, 0.08, 0.40, 0.21, 0.26, 0.19, 0.06, 0.09, 0.20, 0.04, -0.01
  - DEU: 0.08, -0.01, 0.26, 0.02, 0.30, 0.64, 0.36, 0.54, 0.44, 0.56, 0.03, 0.08
  - GBR: 0.32, 0.31, 0.45, 0.75, 0.32, 0.02, 0.12, 0.29, 0.28, 0.71, 0.19, 0.03
  - FRA: -0.04, 0.06, -0.13, -0.28, -0.18, 0.11, -0.06, -0.38, -0.21, -0.69, -0.11, 0.51

- Multi5_ICY
  - USA: 0.24, 0.12, 0.50, 0.35, 0.37, -0.02, 0.39, 0.27, 0.24, 0.39, 0.20, -0.09
  - JPN: 0.13, 0.21, 0.08, 0.40, 0.21, 0.26, 0.19, 0.05, 0.09, 0.19, 0.04, -0.02
  - DEU: 0.08, -0.01, 0.26, 0.03, 0.30, 0.64, 0.36, 0.54, 0.44, 0.56, 0.03, 0.08
  - GBR: 0.32, 0.31, 0.45, 0.76, 0.33, 0.02, 0.13, 0.30, 0.29, 0.73, 0.20, 0.04
  - FRA: -0.04, 0.06, -0.13, -0.27, -0.18, 0.11, -0.06, -0.37, -0.21, -0.69, -0.10, 0.52

- Multi5_ICYV
  - USA: 0.24, 0.12, 0.50, 0.35, 0.37, -0.02, 0.40, 0.27, 0.25, 0.39, 0.20, -0.09
  - JPN: 0.13, 0.19, 0.09, 0.41, 0.21, 0.27, 0.16, 0.04, 0.07, 0.20, 0.03, -0.02
  - DEU: 0.07, -0.04, 0.27, 0.05, 0.31, 0.67, 0.31, 0.52, 0.41, 0.57, 0.02, 0.08
  - GBR: 0.32, 0.31, 0.45, 0.76, 0.33, 0.02, 0.13, 0.30, 0.29, 0.73, 0.21, 0.04
  - FRA: -0.03, 0.07, -0.14, -0.28, -0.18, 0.10, -0.03, -0.37, -0.19, -0.69, -0.10, 0.52

- Multi5_rwICYV
  - USA: 0.25, 0.13, 0.52, 0.36, 0.38, -0.02, 0.40, 0.29, 0.25, 0.39, 0.20, -0.09
  - JPN: 0.13, 0.19, 0.09, 0.41, 0.21, 0.27, 0.16, 0.04, 0.07, 0.20, 0.04, -0.02
  - DEU: 0.07, -0.04, 0.27, 0.04, 0.31, 0.67, 0.31, 0.52, 0.40, 0.57, 0.02, 0.08
  - GBR: 0.33, 0.31, 0.45, 0.76, 0.33, 0.02, 0.13, 0.30, 0.30, 0.73, 0.21, 0.04
  - FRA: -0.03, 0.07, -0.14, -0.28, -0.18, 0.10, -0.03, -0.37, -0.19, -0.69, -0.10, 0.52

### Panel B. Factor model with three systemic economies
- Multi3
  - USA: 0.39, 0.25, 0.70, 0.62, 0.51, 0.03, 0.48, 0.32, 0.35, 0.63, 0.39, 0.20
  - JPN: 0.15, 0.24, 0.11, 0.45, 0.23, 0.30, 0.23, 0.06, 0.12, 0.23, 0.06, 0.06
  - DEU: 0.15, 0.12, 0.29, 0.05, 0.26, 0.71, 0.33, 0.37, 0.37, 0.28, 0.01, 0.39

- Multi3_rw
  - USA: 0.39, 0.25, 0.70, 0.62, 0.51, 0.03, 0.48, 0.32, 0.35, 0.63, 0.37, 0.20
  - JPN: 0.15, 0.24, 0.11, 0.45, 0.23, 0.30, 0.23, 0.06, 0.12, 0.23, 0.08, 0.06
  - DEU: 0.15, 0.12, 0.29, 0.05, 0.26, 0.71, 0.33, 0.37, 0.37, 0.28, 0.01, 0.39

- Multi3_IC
  - USA: 0.38, 0.27, 0.68, 0.65, 0.49, 0.01, 0.44, 0.34, 0.34, 0.58, 0.27, 0.05
  - JPN: 0.15, 0.23, 0.10, 0.43, 0.21, 0.27, 0.20, 0.04, 0.09, 0.19, 0.04, 0.04
  - DEU: 0.15, 0.12, 0.30, 0.05, 0.27, 0.72, 0.35, 0.37, 0.38, 0.30, 0.02, 0.43

- Multi3_ICY
  - USA: 0.38, 0.27, 0.68, 0.64, 0.49, 0.01, 0.44, 0.33, 0.33, 0.57, 0.27, 0.04
  - JPN: 0.15, 0.23, 0.10, 0.43, 0.21, 0.27, 0.19, 0.04, 0.09, 0.18, 0.04, 0.04
  - DEU: 0.15, 0.11, 0.30, 0.06, 0.28, 0.72, 0.36, 0.38, 0.39, 0.31, 0.02, 0.43

- Multi3_ICYV
  - USA: 0.38, 0.28, 0.68, 0.64, 0.48, 0.01, 0.45, 0.33, 0.34, 0.57, 0.27, 0.04
  - JPN: 0.15, 0.22, 0.11, 0.44, 0.22, 0.29, 0.16, 0.03, 0.07, 0.19, 0.04, 0.04
  - DEU: 0.15, 0.09, 0.31, 0.07, 0.28, 0.74, 0.32, 0.36, 0.36, 0.32, 0.01, 0.44

- Multi3_rwICYV
  - USA: 0.39, 0.29, 0.69, 0.65, 0.49, 0.01, 0.45, 0.34, 0.35, 0.58, 0.27, 0.05
  - JPN: 0.15, 0.22, 0.11, 0.44, 0.22, 0.29, 0.16, 0.03, 0.07, 0.19, 0.04, 0.04
  - DEU: 0.14, 0.09, 0.30, 0.07, 0.28, 0.74, 0.32, 0.36, 0.36, 0.31, 0.01, 0.43

*Source: authors’ calculations.*

---


_Source: https://www.imf.org/-/media/websites/imf/imported-full-text-pdf/external/pubs/ft/wp/2012/_wp12262.pdf_