## _wp1512

## Source details

**Canonical URL:** [_wp1512](https://www.imf.org/-/media/websites/imf/imported-full-text-pdf/external/pubs/ft/wp/2015/_wp1512.pdf)

## Other formats

- [Markdown version](/-/media/websites/imf/imported-full-text-pdf/external/pubs/ft/wp/2015/_wp1512.pdf.md)
- [Structured JSON version](/-/media/websites/imf/imported-full-text-pdf/external/pubs/ft/wp/2015/_wp1512.pdf.json)

---

### I. Purpose and scope
- Proposes a new approach for analyzing dollarization that abstracts from the value-effect of exchange rate movements.
- Notes limitation of the standard nominal-dollarization measure: foreign currency deposits (FCDs) are converted to local currency using the exchange rate and deflated by either broad money or total deposits (all in local currency terms), which allows exchange rate movements to change the dollarization ratio even with an unchanged stock of FCD.
- Argues importance of removing exchange-rate bias to better capture economic agents’ demand for FCD.
- Focuses on low-income countries (LICs) during the Global Financial Crisis, using monthly data spanning 2006–09.

### II. Measurement: real deposit dollarization index
- Deposit dollarization defined as the share of foreign currency deposits in total domestic deposits in the banking system.
- Introduces “real” deposit dollarization to remove effect of exchange rate changes on the share of FCD in total deposits.
- Real deposit dollarization index construction:
  - Convert both foreign currency deposits and bank deposits to dollars.
  - Multiply both (back to domestic currency) by a fixed base-year nominal exchange rate.
- Rationale: prevent exchange-rate-driven changes in the dollarization ratio that do not reflect behavioral changes.
- Formal adjustment example and formula cited:
  - adj FCD_t = (FCD_t / NER_t * NER_{t==2000})
  - Real deposit dollarization = adj FCD_t / (adj FCD_t + LCD_t)
- Rationale for deposits denominator:
  - Broad money-based estimates conflate preferences for deposits versus currency and suffer from measurement error due to lack of accurate data on foreign currency in circulation; using total deposits isolates deposit-holding preference.

### III. Trends and descriptive findings (2006–09 monthly panel)
- LICs:
  - Nominal dollarization surged during the Global Financial Crisis.
  - Real dollarization maintained a downward trend during the same period.
- Emerging Markets (EMs):
  - Dollarization generally surged during the crisis and was higher in nominal terms.
- Additional descriptive insights:
  - Exchange rate movements play an important role; domestic currency depreciations raise the dollarization ratio even before agents change behavior.
  - Beyond the variance of inflation and depreciation, the level of inflation and size of depreciation also affect dollarization.
- Regional and market observations:
  - Sub-Saharan Africa (SSA) and Middle East and Central Asian LICs had the lowest deposit dollarization ratios (real and nominal) and experienced the largest declines during the financial crisis.
  - Country examples highlight large nominal increases partly explained by exchange rate movements (Ghana, Zambia examples described).

### IV. Literature, theoretical framework, and model implications
- Five broad causal factor categories summarized:
  - Price movements (inflation and depreciation) affecting store of value.
  - Relative rates of return (interest rate spreads and exchange rate expectations).
  - Institutions and policy credibility (sovereign risk, fiscal/monetary credibility).
  - Macroprudential tools (capital controls, reserve requirements, prudential rules).
  - Portfolio-based limits (minimum variance portfolio/portfolio balance models explain limited cross-country variation).
- Portfolio-balance theoretical model:
  - Wealth composition: W_t = M_t + B_t + s_t F_t.
  - Drivers of foreign-currency demand: expected movements in domestic currency value, sovereign risk (ρ_t), real/technology shocks.
  - Comparative statics/dynamics:
    - With constant exchange rate expectations, expansionary monetary policy and exchange rate depreciations increase demand for foreign assets and dollarization.
    - Positive real shocks that appreciate the domestic currency reduce dollarization.
    - Higher sovereign risk produces flight-to-quality and increases foreign asset demand.
  - Equilibrium dynamics require a Marshall-Lerner-type condition for saddle-point stability (text describes condition: s_t p_t n_t^-1 > 1).

### V. Empirical approach and estimation strategy
- Empirical specification (dynamic linear model, equation (9)):
  - ln y_it = α_i + β ln RR_it + γ ln S_it + δ ln VIX_it + η' X_it + u_it
    - y_it: real deposit dollarization index.
    - RR_it: expected real returns (or indicators such as inflation affecting real returns) on domestic assets relative to dollar-denominated assets.
    - S_it: exchange rate (US dollar per domestic currency; an increase denotes domestic currency appreciation).
    - VIX_it: measure of global risk aversion/uncertainty.
    - X_it: controls (exchange rate regime, prudential requirements, quality of institutions).
- Estimators and robustness:
  - Baseline: fixed effects estimator with autoregressive disturbances (AR(1)).
  - Endogeneity addressed with Arellano and Bond (1991) system GMM estimator using lagged levels as instruments.
  - Short-term dynamics examined with panel VAR (PVAR) to derive impulse-response functions and variance decompositions.
- Data and focus:
  - Monthly series covering 2006m1-2009m12; main empirical focus on LICs (country list in Appendix Table A1).
  - Variable descriptions and summary statistics available in Appendix Tables A2 and A3.

### VI. Key empirical results and dynamics
- Main drivers for LICs in sample period:
  - Relative prices, exchange rate movements, global volatility (VIX), and institutional factors.
- Specific econometric findings (Box 1, Table 1; dependent variable = log. of real dollarization indicator):
  - Fixed Effects estimates (probability values in parentheses):
    - Log. relative prices: 0.1638** (0.0320)
    - Log. exchange rates: -0.2226*** (0.0000)
    - Kauffman entrepreneurial activity indicator (KIEA): 0.1547*** (0.0050)
    - International Country Risk Guide (ICRG): 0.0106*** (0.0000)
    - Exchange rate related measures: -0.0205 (0.7220)
    - Log. global risk aversion indicator (VIX): -0.0108** (0.0110)
    - Crisis time dummy: 0.0642* (0.0570)
    - Exchange rate regime: -0.0725 (0.4760)
    - Constant: 0.2319*** (0.0000)
    - Number of observations: 944
    - R-squared: 0.4749
    - F-ratio: 100.7450
    - Prob > F: 0.0000
  - System GMM estimates (with AR(1) disturbances (t-2)):
    - Log. relative prices: 0.1405* (0.0680)
    - Log. exchange rates: -0.0946*** (0.0020)
    - Kauffman entrepreneurial activity indicator (KIEA): 0.2576*** (0.0000)
    - International Country Risk Guide (ICRG): 0.0181*** (0.0000)
    - Exchange rate related measures: 0.8230 (0.1280)
    - Log. global risk aversion indicator (VIX): -0.0221 (0.4570)
    - Crisis time dummy: 0.0638 (0.5550)
    - Exchange rate regime: -0.0562 (0.6550)
    - Number of observations: 989
    - ρ (AR(1)): 0.8354
    - AR(1) test in first differences (z): -0.6200 (Prob >z 0.5320)
    - Arellano-Bond AR(2) test in first differences (z): 0.3300 (Prob >z 0.7440)
    - Sargan Test of Overidentification (χ2): 352.9 (Prob > (χ2) 1.0000)
    - F-ratio: 1.49
    - Prob > F: 0.0000
- Interpretation of coefficients and robustness:
  - Exchange rate appreciation/depreciation:
    - "A one-percent appreciation in the exchange rate is associated with a 0.09–0.2 percent decrease in the real deposit dollarization ratio." Results described as strongly significant and robust across specifications.
  - Global risk aversion (VIX):
    - "A ten percent increase in the VIX index induces a statistically significant 0. 1 percent decline in real dollarization." Economic impact described as low; results sensitive to estimation strategy (insignificant under system GMM but significant in PVAR).
  - Institutions (ICRG) and entrepreneurial activity (KIEA) are positively associated with real dollarization in both estimators.
  - Exchange-rate-related prudential measures show mixed/insignificant estimated effects once macroeconomic and institutional factors are controlled for.

### VII. Dynamic responses and variance decomposition (PVAR)
- Panel VAR specification and identification:
  - Estimated in first differences using linear GMM; Choleski decomposition used for shock identification.
  - Variables include log. VIX, log. Rel P., log. Exrate, log. Rdol (real dollarization).
- Impulse-response findings:
  - One standard deviation shocks to VIX, relative prices (RR), and exchange rates (S) have strong, significant, but short-lived impacts on real dollarization, typically lasting up to about 2 months.
  - A one standard deviation innovation in VIX increases domestic prices (and relative prices) over the following five months, domestic currency depreciates, and real dollarization surges.
  - Similar surge effects for shocks to relative prices and exchange rates.
  - Alternative variable ordering in Choleski yields broadly similar results though transmission details differ.
- Variance decompositions:
  - Shocks explain a non-trivial percentage of short-term variation in relative prices, exchange rates, and real dollarization (Appendix Tables A4 and A5 report percent of variation at k-periods after a one standard deviation shock).

### VIII. Policy relevance and implications
- Measurement contribution:
  - Adopts a measure of deposit dollarization that adjusts FCDs for exchange rate fluctuations—real dollarization—to capture behavior-driven changes and remove ex post measurement effects.
  - Using adjusted measure, the paper finds a continued decline in real dollarization ratio in LICs, contrasting with the impression of a surge when using unadjusted nominal measures.
- Policy-relevant conclusions and recommendations:
  - Primary policy focus should be on macroeconomic stability to contain inflationary pressures and to anchor exchange rate expectations.
  - Prudent debt management to avoid debt distress and to taper sovereign risk will help better integrate LICs into international financial markets and dampen drivers of dollarization.
  - While there could be scope for macroprudential measures in dampening dollarization, estimated effects of exchange-rate-related prudential measures are statistically insignificant once macroeconomic and institutional factors are controlled for; thus policy efforts should prioritize macroeconomic stability and sovereign risk management.
- Additional policy considerations:
  - High dollarization weakens monetary policy transmission because monetary instruments mainly impact domestic liquidity; high dollarization reduces central banks’ capacity to control liquidity that could fuel consumer price inflation.
  - High dollarization limits central banks’ capacity to act as lender of last resort (LOLR) in foreign currency; for FCDs, limited international reserves are the only buffer against liquidity crises, making foreign currency deposit runs more serious.
  - The paper investigates the extent to which policy interventions (including capital controls and other prudential measures) can limit or damp dollarization and the channels through which global financial stress affects dollarization.

*Source: _wp1512 (excerpt: Introduction, measurement, trends, theoretical model, empirical approach, results, Box 1).*

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

### _wp1512 - References

### I. Introduction: purpose and scope
- Proposes a new approach for analyzing dollarization that abstracts from the value-effect of exchange rate movements.
- Notes limitation of the standard nominal-dollarization measure: foreign currency deposits (FCDs) are converted to local currency using the exchange rate and deflated by either broad money or total deposits (all in local currency terms), which allows exchange rate movements to change the dollarization ratio even with an unchanged stock of FCD.
- Argues importance of removing exchange-rate bias to better capture economic agents’ demand for FCD.
- Focuses on low-income countries (LICs) during the Global Financial Crisis, using monthly data spanning 2006–09.

### II. Measurement: real deposit dollarization index
- Deposit dollarization defined as the share of foreign currency deposits in total domestic deposits in the banking system.
- Introduces “real” deposit dollarization to remove effect of exchange rate changes on the share of FCD in total deposits.
- Real deposit dollarization index construction:
  - Convert both foreign currency deposits and bank deposits to dollars.
  - Multiply both (back to domestic currency) by a fixed base-year nominal exchange rate.
- Rationale: prevent exchange-rate-driven changes in the dollarization ratio that do not reflect behavioral changes.

### III. Trends and descriptive findings (2006–09 monthly panel)
- For LICs:
  - Dollarization in nominal terms surged during the Global Financial Crisis.
  - Real dollarization maintained a downward trend during the same period.
- For Emerging Markets (EMs):
  - Dollarization generally surged during the crisis and was higher in nominal terms.
- Interpretation: exchange rate movements play an important role; domestic currency depreciations raise the dollarization ratio even before agents change behavior.
- Additional descriptive insight:
  - Beyond the variance of inflation and depreciation, the level of inflation and size of depreciation also affect dollarization.

### IV. Empirical approach and identification
- Estimation methods:
  - Panel system GMM estimator.
  - Fixed effects estimator with autoregressive disturbances (AR(1)).
- Supplementary analysis:
  - Impulse response functions using panel vector autoregression (PVAR).
  - Variance decompositions to quantify the relative importance of shocks.
- Robustness:
  - Results from system GMM and fixed effects (with AR(1) disturbances) are broadly similar, indicating limited endogeneity bias under the latter estimator.
  - Exception: sensitivity of VIX results across estimators.

### V. Key econometric and dynamic results
- Determinants of deposit dollarization among LICs in the sample period:
  - Largely driven by changes in relative prices, exchange rate movements, global volatility (Chicago Board Options Exchange Volatility Index, the VIX), and institutional factors.
- Specific findings:
  - Exchange rate depreciation is associated with a surge in dollarization.
  - A rise in VIX helps contain dollarization (results sensitive: system GMM not statistically significant; fixed effects shows low economic impact).
  - Effect of the exchange rate regime is statistically insignificant but consistent with Calvo-Reinhart-type fear-of-floating effects.
- PVAR impulse response analysis:
  - Shocks to VIX, relative prices, and exchange rate have strong, significant but short-lived impact on real dollarization.
  - Positive, albeit short-term, relationship between the VIX and dollarization suggests that during global financial volatility domestic residents increased dollar holdings due to uncertainty about spillovers—flight-to-quality outweighs home asset preference.

### VI. Policy relevance and implications
- High dollarization implications:
  - Weakens monetary policy transmission because monetary instruments mainly impact domestic liquidity; high dollarization reduces the capacity of central banks to control liquidity that could fuel consumer price inflation.
  - Limits central banks’ capacity to act as lender of last resort (LOLR) in foreign currency; for FCDs, limited international reserves are the only buffer against liquidity crises, making foreign currency deposit runs more serious.
- Investigates the extent to which policy interventions (including capital controls and other prudential measures) can limit or damp dollarization; channels through which global financial stress affects dollarization are examined.

*Source: _wp1512 - References (excerpt: pages with Introduction and overview; monthly panel data 2006–09).*

### Section IV reviews the literature on dollarization and outlines the contributions of this paper

### _wp1512 - Section IV reviews the literature on dollarization and outlines the contributions of this paper

### Measurement of deposit dollarization
- Dollarization takes forms: transacting, storing financial assets/liabilities, and/or indexing prices in foreign currency.
- Payment dollarization is difficult to measure because monetary authorities cannot perfectly monitor foreign currency circulating in economies.
- Common literature measure: nominal deposit dollarization ratio = ratio of FCDs to broad money or total deposits.
- Issue with nominal FCD measures: FCDs are compiled in local currency terms after converting multiple foreign currencies; exchange rate movements can produce large swings in the nominal ratio even when foreign-currency stocks (in foreign currency terms) are unchanged.
  - Numerical example: 70 units of FCD and 30 units of LCD. If currency depreciates by 10 percent, FCD in local terms goes from 70 to 77; total becomes 107 and FCD share = 77/107 = 72 percent, although foreign exchange holdings in foreign currency were unchanged.
- Proposed measure: a “real” deposit dollarization indicator that abstracts from exchange rate movements by using a constant-exchange-rate (base-year) adjustment:
  - adj FCD_t = (FCD_t / NER_t * NER_{t==2000}), where NER is the nominal exchange rate (local currency per dollar).
  - Real deposit dollarization = adj FCD_t / (adj FCD_t + LCD_t)
- Rationale for using deposits (denominator) instead of broad money:
  - Broad money-based estimates conflate preferences for deposits versus currency and suffer from measurement error due to lack of accurate data on foreign currency in circulation.
  - Using total deposits (only deposits captured in broad money and expected to affect aggregate demand) isolates preference for holding deposits in foreign currency versus domestic deposits.

### Recent trends in deposit dollarization (empirical patterns)
- Comparison periods and data: monthly series covering 2006m1-2009m12; figures and analyses use IMF International Financial Statistics (IFS) and author estimates.
- Distinct patterns for nominal vs “real” dollarization around the global financial crisis:
  - Nominal dollarization surged during the global financial crisis.
  - Real dollarization shows:
    - Continued decline in LICs.
    - A smaller increase in EMs compared with nominal measures.
    - Lower volatility (smaller swings) than nominal dollarization.
- Exchange rate depreciations during the crisis explain much of the nominal surge; many countries’ exchange rates remained more depreciated and volatile than in the pre-crisis period.
  - Country examples: large increases in nominal dollarization ratios in Ghana and Zambia partly explained by large exchange rate movements (figures use $ per Kwacha and $ per Cedi rebased to 100 in 2006m10).
- Regional and institutional heterogeneity:
  - Sub-Saharan Africa (SSA) and Middle East and Central Asian LICs had the lowest deposit dollarization ratios (real and nominal) and experienced the largest declines during the financial crisis.
- Relationship with global risk aversion (VIX):
  - Scatter plots indicate a negative correlation between dollarization growth and the VIX for the LIC sample (2006m1-2009m12).
    - When global volatility (VIX) increases, deposit dollarization in LICs tends to decline.
    - No clear relationship between dollarization and VIX in EMs.
  - Plausible explanations offered:
    - Heightened home-asset preference during global uncertainty.
    - Limited financial integration and limited direct exposure to the global financial crisis for many LICs.
    - Global shortage of U.S. dollars and associated outflows (channels: curtailing of private capital flows, capital flight, portfolio inflow reversals, drop in foreign trading of local debt, collapse in trade financing and stock markets).
      - EMTA evidence cited: between Q2 and Q3 of 2008, local debt trade by foreign investors declined by 71 percent in LICs compared to 22 percent for EMs; LICs’ volume of trade financing dropped by 18 percent in the last quarter of 2008; Merrill Lynch Africa Lions index declined by almost 70 percent during March–December 2008.
- Prudential measures and dollarization:
  - Changes in prudential measures during the global financial crisis had significant effects on dollarization in LICs.
    - Both LICs that introduced exchange-rate-related measures and those that changed other prudential measures saw declines in dollarization ratios; declines were larger (in real terms) for countries implementing exchange-rate-related measures.
  - Common policy actions during 2005–2011 included controls on lending locally in foreign currency, raised reserve requirements, and differential treatment of deposit accounts in foreign exchange (data from IMF AREAER).
  - Most countries that introduced lending/FX controls removed them shortly after the crisis onset.
- Inflation and depreciation effects on nominal dollarization variance:
  - Countries with high pre-crisis inflation or depreciation experienced increases in the variance of nominal dollarization during the crisis.
  - Countries with lower pre-crisis inflation tended to see reductions in the variance of dollarization.
  - High starting levels of depreciation or inflation are associated with larger increases in the variance of dollarization.

### Literature review: causal drivers of (short-run) dollarization
Five broad categories of causal factors in the literature:
- Price movements affecting store of value:
  - Inflation and/or exchange rate depreciation reduce the real value of financial assets; expectations of these lead agents to hold assets in foreign currency.
  - Ize and Levy-Yeyati (2003, 2005) emphasize relative volatility of inflation and exchange rate.
- Relative rates of return:
  - Spreads between domestic and foreign interest rates influence currency-holding decisions.
  - These rate differentials can be interrelated with exchange rate expectations (consistent with uncovered interest parity).
- Institutions and policy credibility:
  - Concerns about financial system collapse, exchange rate regime collapse, or debt default increase dollarization.
  - Expectations about monetary policy conduct in a regime collapse affect holdings regardless of collapse probability.
  - Institutional factors (political, legal, cultural) and fiscal discipline influence dollarization.
  - If agents expect government bailouts that ignore currency risk, this can increase foreign-currency borrowing and dollarization.
- Macroprudential tools:
  - Measures such as capital controls, reserve requirements, and prudential rules can reduce incentives for banks to borrow externally or in foreign currency.
- Other references and model limitations:
  - Portfolio-based explanations (minimum variance portfolio and portfolio balance models) explain only a limited share of cross-country variation in dollarization.

### Theoretical model (portfolio balance framework)
- Framework basis: portfolio balance models (Kouri 1976; Calvo and Rodríguez 1977; Branson and Henderson 1985).
- Wealth composition: domestic money (M), domestic bonds (B), foreign bonds denominated in foreign currency (F). Demand for foreign bonds implies demand for foreign currency.
- Key drivers of deposit dollarization in model:
  - Expected movements in domestic currency value (expectations/confidence; fiscal/monetary policy).
  - Perceptions of sovereign risk (ρ_t).
  - Real/technology shocks.
- Rational expectations model summarized in equations (1)–(5) (variables in levels):
  - (1) W_t = M_t + B_t + s_t F_t
  - (2)–(4) first-order conditions for money, bonds, foreign currency demand (sign and shape conditions specified in text).
  - (5) current account / accumulation of foreign assets: n_t = g(s_t, Z_t) with specified comparative statics (real exchange depreciation increases net savings/current account balance; real shocks do likewise).
- Key comparative-static and dynamic implications:
  - With constant exchange rate expectations (E_t^[] Δs_t = 0):
    - Expansionary monetary policy and exchange rate depreciations increase demand for foreign assets and dollarization.
    - Positive real/technology shocks that appreciate the domestic currency reduce demand for foreign assets and reduce dollarization.
    - Increases in sovereign risk (ρ_t) produce flight-to-quality, increasing foreign asset demand.
- Dynamics and equilibrium:
  - Differential equations for shares of money and foreign currency derived (equation (6)); expected change in exchange rate as a function of asset shares (equation (7)).
  - Saddle-point equilibrium requires Marshall-Lerner-type condition: s_t p_t n_t^-1 > 1 (expressed in text as Marshall-Lerner condition given ts_p_s_n_n be greater than one).
  - Graphical implications: exchange rate depreciation increases foreign currency demand (movement along F = 0 schedule); positive real shocks shift foreign currency demand schedule down (reducing demand); unanticipated expansionary monetary policy that shifts expected depreciation schedule up requires offsetting depreciation or increased foreign asset holdings to keep expectations constant; increased sovereign risk raises foreign currency demand as safe haven.
- Policy environment matters: monetary policy, prudential measures, exchange rate regime, and institutional quality are included in empirical estimations.

### Empirical approach and estimation strategy
- Empirical specification: dynamic linear model (equation (9)):
  - ln y_it = α_i + β ln RR_it + γ ln S_it + δ ln VIX_it + η' X_it + u_it
    - y_it: real deposit dollarization index (as defined above).
    - RR_it: expected real returns (or indicators such as inflation affecting real returns) on domestic assets relative to dollar-denominated assets.
    - S_it: exchange rate (US dollar per domestic currency; an increase denotes domestic currency appreciation).
    - VIX_it: measure of global risk aversion/uncertainty.
    - X_it: controls (exchange rate regime, prudential requirements, quality of institutions).
    - η: country-specific unobservable fixed effects; v_it: serially uncorrelated errors.
- Estimators and robustness:
  - Baseline: fixed effects estimator with autoregressive disturbances.
    - Fixed effects preferred if normality conditions hold.
  - Endogeneity concerns: dependent variable may be correlated with error term; explanatory variables may be endogenous.
    - Addressed with Arellano and Bond (1991) system GMM estimator using lagged levels as instruments.
  - Short-term dynamics and responses: panel VAR (PVAR) estimations to derive impulse-response functions and variance decompositions.
- Data focus and appendices:
  - Explanatory variable descriptions grouped in five categories (Box 1 in source).
  - Variable descriptions and summary statistics for LICs available in Appendix Tables A2 and A3.
  - Empirical analysis focuses on LICs (country list in Appendix Table A1).

*Source: _wp1512 - Section IV reviews the literature on dollarization and outlines the contributions of this paper*

### Box 1. Determinants of Surge in Deposit Dollarization

### Box 1. Determinants of Surge in Deposit Dollarization

### Determinants (categorical overview)
- Macroeconomic variables
  - Returns/Price differential or relative returns/prices
  - Exchange rate: depreciation (level and volatility), exchange market pressure
- Institutions
  - Bureaucracy quality (e.g., International Country Risk Guide (ICRG))
  - Entrepreneurial quality (e.g., Kauffman index of entrepreneurial activity (KIEA))
- Prudential measures
  - Reserve requirements
  - Net open positions and other exchange related measures
  - Other prudential requirements
- Flight to quality
  - Global risk aversion (VIX)
- Controls
  - Level of dollarization
  - Exchange rate regime

### Key empirical findings (panel results and Table 1)
- Exchange rate effects
  - "A one-percent appreciation in the exchange rate is associated with a 0.09–0.2 percent decrease in the real deposit dollarization ratio." These results are described as strongly significant and robust across specifications.
  - Countries with more flexible exchange rate regimes tended to experience lower dollarization, though the elasticity of exchange rate regime is not statistically significant (possibly because regime effects are captured by exchange rate changes).
- Global risk aversion (VIX)
  - "A ten percent increase in the VIX index induces a statistically significant 0. 1 percent decline in real dollarization." The economic impact is described as low; results are sensitive to estimation strategy (insignificant under system GMM but significant in the PVAR).
- Institutions and sovereign risk
  - Higher sovereign risk classification/perception (ICRG) is associated with higher real deposit dollarization.
  - Higher entrepreneurial activity (KIEA) is associated with higher real deposit dollarization.
- Prudential measures
  - Strengthening prudential/exchange-rate-related measures during an upsurge did not moderate dollarization and may worsen it; estimated elasticities are statistically insignificant and point to a mixed impact once other variables are controlled for.
- Inflation and depreciation
  - Beyond variance of inflation and depreciation, the level of inflation and size of depreciation also matter for dollarization (supported by earlier work cited).

### Exact coefficients and test statistics (Table 1: Dependent variable = log. of real dollarization indicator)
- Fixed Effects column (probability values in parentheses)
  - Log. relative prices: 0.1638** (0.0320)
  - Log. exchange rates: -0.2226*** (0.0000)
  - Kauffman entrepreneurial activity indicator (KIEA): 0.1547*** (0.0050)
  - International Country Risk Guide (ICRG): 0.0106*** (0.0000)
  - Exchange rate related measures: -0.0205 (0.7220)
  - Log. global risk aversion indicator (VIX): -0.0108** (0.0110)
  - Crisis time dummy: 0.0642* (0.0570)
  - Exchange rate regime: -0.0725 (0.4760)
  - Constant: 0.2319*** (0.0000)
  - Number of observations: 944
  - R-squared: 0.4749
  - F-ratio: 100.7450
  - Prob > F: 0.0000
- System GMM column (with AR(1) disturbances (t-2))
  - Log. relative prices: 0.1405* (0.0680)
  - Log. exchange rates: -0.0946*** (0.0020)
  - Kauffman entrepreneurial activity indicator (KIEA): 0.2576*** (0.0000)
  - International Country Risk Guide (ICRG): 0.0181*** (0.0000)
  - Exchange rate related measures: 0.8230 (0.1280)
  - Log. global risk aversion indicator (VIX): -0.0221 (0.4570)
  - Crisis time dummy: 0.0638 (0.5550)
  - Exchange rate regime: -0.0562 (0.6550)
  - Number of observations: 989
  - ρ (AR(1)): 0.8354
  - AR(1) test in first differences (z): -0.6200 (Prob >z 0.5320)
  - Arellano-Bond AR(2) test in first differences (z): 0.3300 (Prob >z 0.7440)
  - Sargan Test of Overidentification (χ2): 352.9 (Prob > (χ2) 1.0000)
  - F-ratio: 1.49
  - Prob > F: 0.0000

Notes: All variables are expressed in logarithms. Figures in parenthesis are probability values; ***, **, and * indicate statistical significance at 1 percent, 5 percent, and 10 percent respectively.

### Dynamic responses (panel VAR / impulse responses)
- Model specification and identification
  - Panel VAR estimated in first differences using linear GMM; simple Choleski decomposition used to identify shocks.
  - Variables examined include log. VIX, log. Rel P., log. Exrate, log. Rdol (real dollarization).
- Main dynamic findings
  - One standard deviation shocks to global uncertainty (VIX), relative prices (RR), and exchange rates (S) have strong and significant impacts on real dollarization surge.
  - Estimated responses to a shock in VIX, RR, or S tend to be large but short-lived, lasting up to about 2 months.
  - A one standard deviation innovation in VIX increases domestic prices (and relative prices) significantly over the following five months, the domestic currency depreciates, and real dollarization surges (Appendix Figure 7, column 1).
  - Similar surge effects are obtained for shocks to relative prices (column 2) and shocks to exchange rates (column 3).
  - Alternative ordering (exchange rate preceding relative prices in Choleski) yields similar broad results though transmission mechanisms differ (Appendix Figure 8).
  - The positive short-term relationship between VIX and dollarization may reflect domestic residents increasing foreign currency holdings due to uncertainty about spillovers from global financial volatility; results may also reflect limited adverse impact of the global financial crisis on many LICs.
- Variance decompositions (selected summary)
  - Shocks explain a non-trivial percentage of short-term variation in relative prices, exchange rates, and real dollarization (Appendix Tables A4 and A5 and related text). Results reported as percent of variation at k-periods after a one standard deviation shock (k values reported in Appendix tables).

### Conclusions and policy implications
- Measurement contribution
  - The paper adopts a measure of deposit dollarization that adjusts foreign exchange deposits for exchange rate fluctuations to capture behavior-driven changes (real dollarization), removing ex post measurement effects of exchange rate movements.
  - Using the adjusted measure, the paper finds a continued decline in real dollarization ratio in LICs (contrasting with the impression of a surge when using unadjusted measures).
- Policy-relevant conclusions
  - Primary policy focus should be on macroeconomic stability to contain inflationary pressures and to anchor exchange rate expectations.
  - Prudent debt management to avoid debt distress and to taper sovereign risk will help better integrate LICs into international financial markets and dampen the impact and amplitude of drivers of dollarization.
  - While there could be scope for macroprudential measures in dampening dollarization, the estimated effects of exchange-rate-related prudential measures are statistically insignificant once macroeconomic and institutional factors are controlled for; thus policy efforts should prioritize macroeconomic stability and sovereign risk management.

*Source: IMF World Economic Outlook (WEO) database; authors' estimations and text from Box 1, "Determinants of Surge in Deposit Dollarization."*

---


_Source: https://www.imf.org/-/media/websites/imf/imported-full-text-pdf/external/pubs/ft/wp/2015/_wp1512.pdf_