## Monetary Policy Transmission in Euroized Countries: Evidence from Emerging Europe — Working Paper No. WP/2026/177

## Source details

**Canonical URL:** [Monetary Policy Transmission in Euroized Countries: Evidence from Emerging Europe — Working Paper No. WP/2026/177](https://www.imf.org/-/media/files/publications/wp/2025/english/wpiea2025177-print-pdf.pdf)

## Other formats

- [Markdown version](/-/media/files/publications/wp/2025/english/wpiea2025177-print-pdf.pdf.md)
- [Structured JSON version](/-/media/files/publications/wp/2025/english/wpiea2025177-print-pdf.pdf.json)

---

### Introduction: motivation, scope, and key mechanisms
- Motivation and scope
  - Renewed interest in effects of monetary policy on output and prices following surging inflation in the post-pandemic period.
  - Focus: how financial euroization (households and firms borrowing or saving in euro rather than local currency) affects monetary policy transmission in Central Eastern and South-Eastern European (CESEE) countries that are not part of the euro area.
  - Sample: Albania, Czech Republic, Hungary, North Macedonia, Norway, Poland, Romania, Serbia, and Sweden.
  - Sample period: 2007q1 to 2023q1 (North Macedonia begins 2012q2; Albania begins 2008q1).
  - Norway and Sweden serve as non-euroized benchmarks (small open European economies with independent monetary policy; their currencies are not reserve or safe haven currencies).
- How euroization can impair monetary policy transmission
  - Channel 1: Interest, credit, cash flow, and balance sheet channels are undermined when a significant share of loans and savings are in euro because interest rates on euro loans and deposits are influenced by the ECB’s policy.
  - Channel 2: Euroization often motivates exchange rate stabilization, limiting the exchange rate’s role as a transmission channel.
  - Indirect effect: foreign exchange interventions associated with euroization can cement a monetary policy framework where central banks pursue de facto inflation and exchange rate objectives, creating potential policy trade-offs.
- Financial structure and exchange rate regimes (overview)
  - Euroization measured by two metrics: share of FX loans to total loans and share of FX deposits in total deposits.
  - Observed pattern: Czech Republic and Poland relatively low euroization; Romania, Hungary, North Macedonia, Albania, and Serbia relatively higher euroization.
  - General de-euroization over the past decade, but euroization remains high in Serbia, Albania, and North Macedonia.
  - Exchange rate regimes (AREAER assessment):
    - Czech Republic and Poland: freely floating exchange rates.
    - Albania and Hungary: floating arrangements.
    - North Macedonia, Romania, Serbia: stabilized arrangements.
    - Norway and Sweden: pure inflation targeters with freely floating exchange rates.
  - Financial sector size differences:
    - Sweden and Norway: domestic credit-to-GDP ratios exceed 100 percent.
    - Euroized countries typically: domestic credit-to-GDP ratio of less than 40 percent.
    - Credit-to-GDP has not improved over past decades for highly euroized economies.

### Empirical framework and identification
- Estimation approach
  - Method: Bayesian Estimation of the Structural Vector Autoregressive Model (Bayesian SVAR) applied to each country separately.
  - Model: xt = [yt, pt, rt, et], where yt = real GDP, pt = inflation (headline HICP), rt = policy rate, et = nominal effective exchange rate (domestic currency per euro; increase = depreciation).
  - Variables seasonally adjusted and logged except interest rates; lag length p = 2.
- Identification
  - Sign restrictions identify three shocks (baseline): demand, supply, and monetary policy shock.
  - Baseline sign restrictions (contemporaneous and first-quarter responses):
    - Demand shock: GDP +, HICP +, INT +.
    - Supply shock: GDP +, HICP -, INT (blank).
    - Monetary policy shock (contractionary): GDP -, HICP -, INT +.
  - Robustness: four-shock model adds an exchange rate shock with an extended set of sign restrictions (see main text for full pattern).
- Scaling and interpretation
  - Responses interpreted as the impact of a 100-basis point (bps) monetary policy shock.
  - Normalization: responses of GDP, HICP, and exchange rates are normalized by the standard deviation of the identified contemporary monetary policy shocks within each country; impulse responses to a 100-bps shock obtained via linear scaling.
- Caveats
  - Sign-restriction SVAR yields a set of admissible decompositions and can produce wide ranges; results sensitive to priors, implementation, and assumption of structural stability over time (may not hold across crises such as COVID).

### 4. Results — aggregate demand and prices (key quantitative findings)
- Definition: euroization level defined as the average share of FX loans in total loans from 2007m1 to 2023m3.
- First-quarter responses to a 100-bps monetary policy shock (baseline model)
  - GDP first-quarter fall:
    - Around 7 percent fall in benchmark countries, Norway and Sweden.
    - 4 percent fall in Czech Republic, Poland, and Hungary.
    - 3 percent fall in Romania and Serbia.
    - Large falls in GDP in North Macedonia and Albania despite very high euroization levels.
  - HICP first-quarter fall (quarter-on-quarter percentage change relative to period 0 steady state):
    - 2.5 percent in Norway and Sweden.
    - 2 percent in Czech Republic and Poland.
    - 1.5 percent in Hungary, Romania, and Serbia.
    - Albania and North Macedonia: decline in inflation in between Poland and Romania despite higher euroization levels.
- Comparison and interpretive context
  - Literature benchmarks: Ramey (2016) reports a trough effect range for a 100-bps interest rate increase between -0.7 and -5 percent for output and prices; Conti et al. (2017) find a 100-bps shock reduces GDP and HICP by around 2-3 percent.
  - Note: standard deviation of monetary shocks across countries ranges from 5-bps to 25-bps; the 100-bps reference shock can produce large magnitudes.
- Robustness
  - Second-quarter responses differ less by euroization level and GDP falls are generally smaller; no clear HICP pattern by euroization.
  - Exchange rate shock-identified model yields analogous first-quarter results: first-quarter effects on GDP and inflation are inversely correlated with euroization, with Albania and North Macedonia as outliers.

### 4.2 Exchange rate channel — stylized results
- First-quarter exchange rate responses:
  - Countries with lower euroization levels show larger exchange rate responses to monetary policy shocks.
  - Countries with higher euroization show muted exchange rate responses.
- Interpretation:
  - Muted exchange rate responsiveness in more euroized economies helps explain attenuated GDP and price responses.
  - The muted response is linked to policy choices (exchange rate stabilization) often associated with euroization rather than euroization mechanically alone.

### 4.3 Credit channel — stylized facts and drivers
- Hypothesis: households and firms switch to FX loans during tightening episodes, so overall credit growth could slow less in euroized economies.
- Evidence during the post-pandemic tightening cycle:
  - In some countries, the share of FX loans to total outstanding NFC loans increased significantly.
    - Example: In Czech Republic, peak increase in the stock of NFC FX loans since the beginning of the monetary policy tightening in 2021 amounted to 20 percentage points.
  - Cumulative real growth of NFC loans:
    - In all countries except Albania and North Macedonia, NFC loans in FX grew more strongly (or declined less) than loans in domestic currency.
    - Romania: strong growth of FX loans plus a relatively large initial share of FX loans implied total NFC credit grew during the post-pandemic tightening cycle despite contraction of domestic currency NFC loans.
- Drivers and cross-country heterogeneity:
  - Variation in differences between domestic and ECB policy rates determines differences between domestic and foreign currency loan rates.
  - Differences between domestic and ECB policy rates in the post-pandemic period:
    - Lowest in Albania, followed by North Macedonia and Serbia — these are the most euroized economies in the sample.
  - Resulting mean differences between rates of new domestic and foreign currency loans to NFCs during the post-pandemic tightening cycle are relatively low in Albania and North Macedonia.

### 4.4 Interest rate channel — pass-through evidence
- Measurement approach
  - Interest rate pass-through measured as sensitivity of bank interest rates on loans and deposits to changes in policy rates.
  - Bank interest rate betas computed as ratio of cumulative increase in bank interest rates on new domestic currency loans to cumulative increase in the policy rate, using the post-pandemic hiking cycle truncated from month of first rate hike until three months after the last hike (or first rate cut or end of sample).
- Main findings on pass-through
  - Pass-through to new loans and mortgages in domestic currency:
    - Pass-through generally higher for more euroized countries in the sample, with exceptions.
    - North Macedonia: high pass-through to mortgages, low pass-through to domestic currency NFC loans.
    - Albania: low pass-through to mortgage rates but relatively high for NFC loans.
  - Pass-through to existing (outstanding) mortgage rates:
    - No clear pattern; pass-through correlated with share of flexible rate mortgages.
  - Pass-through to euro-denominated loans:
    - Pass-through of the domestic policy rate to NFC loans and mortgages in euro is significantly lower (these euro rates do not necessarily respond to domestic policy rate changes).
    - Using the ECB policy rate to calculate pass-through: pass-through to euro NFC loans and mortgages is similar to or above pass-through in the euro area.
- Key implication
  - Domestic policy rate pass-through to domestic currency retail rates is effective in euroized economies; rates of foreign currency loans mostly follow the ECB policy rate.
  - Co-movement between domestic policy rate and rates of loans in euro is small, which plausibly drags down aggregate monetary policy transmission in euroized countries.
  - Albania is an exception where co-movement between domestic policy rate and rates of euro NFC loans and mortgages is larger than that of the ECB policy rate, possibly due to relatively low spread between domestic and ECB policy rates.
- Methodological note
  - The term ‘pass-through’ to euro loan rates may reflect co-movement between domestic and ECB policy rates during a particular episode rather than mechanical pass-through.

### Conclusions and policy implications (summary)
- Aggregate comparison and main finding
  - Aggregate effects on output and prices are correlated with the level of financial euroization.
    - In Norway and Sweden, first-quarter fall in output and inflation is roughly twice as high as in Romania and Serbia (both high euroization).
    - Other countries with lower levels of financial euroization lie between these extremes.
  - Findings are robust across models and apply to second-quarter effects.
- Channel-specific conclusions
  - Exchange rate channel: exchange rates in euroized countries are often stabilized; response to monetary policy shocks is muted, impairing transmission through this channel.
  - Interest rate channel: pass-through to domestic currency retail rates is effective in euroized economies; rates of foreign currency loans mostly follow the ECB policy rate; effectiveness of domestic policy depends on degree of policy synchronization with the ECB.
  - Credit channel: domestic monetary tightening can induce substitution toward foreign currency-denominated credit in highly euroized economies, undermining the credit channel.
- Policy implications and recommended strategies
  - Successful efforts to reduce the level of euroization would increase the effectiveness of monetary policy.
  - Strategies should be designed by understanding underlying drivers of financial euroization:
    - Deposit euroization and loan euroization require different strategies.
    - Deposit euroization: establish a strong track record of prudent macroeconomic policies to reduce perception of the euro as the only reliable hedge.
    - Loan euroization: driven in part by interest-rate differentials; allowing more exchange rate flexibility and regulatory measures could make FX loans less attractive.
  - Greater exchange rate flexibility could itself enhance monetary policy transmission and support more effective inflation targeting.
- Suggested avenues for future research
  - Investigate why Albania and North Macedonia are often outliers (Albania’s synchronization with ECB policy rates and relatively flexible exchange rate regime; North Macedonia’s monetary policy framework).
  - Examine whether exchange rate stabilization triggered by euroization has led to suboptimal monetary policy from an inflation-targeting perspective.
  - Study determinants of euroization (including monetary policy) using microeconomic data or structural models to understand financial institutions' decision-making.
- Methodological caveats reiterated
  - Sign restrictions identification is widely used and sufficient for cross-country comparison here, but identification strategies could be augmented; data limitations restrict choices for some countries.

*IMF Working Paper — Monetary Policy Transmission in Euroized Countries: Evidence from Emerging Europe*

### 1. Introduction ........................................................................................................

### 1. Introduction

### Motivation and scope
- Renewed interest in the effects of monetary policy on output and prices following surging inflation in the post-pandemic period.
- Focus: how financial euroization (households and firms borrowing or saving in euro rather than local currency) affects monetary policy transmission in Central Eastern and South-Eastern European (CESEE) countries that are not part of the euro area.
- Paper re-examines effects of euroization on monetary policy transmission in Europe using seven partially euroized CESEE countries with independent currencies and two non-euroized regional comparator countries (Norway and Sweden).

### How euroization can impair monetary policy transmission
- Channel 1: Interest, credit, cash flow, and balance sheet channels are undermined when a significant share of loans and savings are in euro because interest rates on euro loans and deposits are influenced by the ECB’s policy.
- Channel 2: Euroization often motivates exchange rate stabilization, limiting the exchange rate’s role as a transmission channel.
- Indirect effect: foreign exchange interventions associated with euroization can cement a monetary policy framework where central banks pursue de facto inflation and exchange rate objectives, creating potential policy trade-offs (e.g., when inflation shocks have asymmetric effects relative to the euro area).

### Literature context
- Mixed evidence on dollarization/euroization and monetary transmission:
  - Some papers find reduction in dollarization is associated with stronger monetary policy transmission (e.g., Acosta-Ormaechea and Coble, 2011).
  - Bordon and Weber (2010) find de-dollarization can improve transmission, especially with stronger frameworks.
  - Brandao-Marques et al. (2020) do not find significant effects of financial dollarization on transmission strength.
  - Baliño et al. (1999) find no significant difference in meeting inflation targets between dollarized and non-dollarized countries.
  - Ize and Yeyati (2003) find stronger credibility helps reduce dollarization.

### Contributions of this paper
- Framework quantifying how levels of euroization correlate with the effectiveness of domestic monetary policy in influencing aggregate macro variables.
  - General finding: monetary policy shocks have smaller impacts on inflation and output in economies with higher euroization compared to those with lower euroization (exceptions: Albania and North Macedonia).
  - Example comparison: effects are twice as high in Norway and Sweden compared to Serbia (Serbia has the highest euroization level in the sample).
- Disaggregation of transmission channels: exchange rate, credit, and interest rate channels.
  - Monetary policy has at best limited effects on the exchange rate in euroized countries (consistent with managed or stabilized exchange rate policies).
  - Substitution toward foreign-currency loans during the post-pandemic tightening period in some countries dampened the impact of domestic monetary policy tightening by softening the decline in overall credit growth.
  - Little correlation found between domestic policy rates and rates on foreign-currency loans.
  - Policy rate pass-through to domestic currency interest rates is not weaker in countries with higher degrees of euroization.

### Sample and data (overview)
- Countries analyzed: Albania, Czech Republic, Hungary, North Macedonia, Norway, Poland, Romania, Serbia, and Sweden.
- Sample period: 2007q1 to 2023q1 (North Macedonia begins 2012q2; Albania begins 2008q1).
- Norway and Sweden used as non-euroized benchmarks (small open European economies with independent monetary policy; their currencies are not reserve or safe haven currencies).
- Euroization measured by two metrics: share of foreign exchange (FX) loans to total loans and share of FX deposits in total deposits.
  - Observed pattern: Czech Republic and Poland relatively low euroization; Romania, Hungary, North Macedonia, Albania, and Serbia relatively higher euroization.
  - General de-euroization over the past decade, but euroization remains high in Serbia, Albania, and North Macedonia.
- Exchange rate regimes (as assessed by IMF AREAER):
  - Czech Republic and Poland: freely floating exchange rates.
  - Albania and Hungary: floating arrangements.
  - North Macedonia, Romania, Serbia: stabilized arrangements.
  - Norway and Sweden: pure inflation targeters with freely floating exchange rates.
- Financial sector size differences:
  - Sweden and Norway: domestic credit-to-GDP ratios exceed 100 percent.
  - Euroized countries typically: domestic credit-to-GDP ratio of less than 40 percent.
  - Credit-to-GDP has not improved over past decades for highly euroized economies.

### Estimation approach and identification (technical)
- Method: Bayesian Estimation of the Structural Vector Autoregressive Model (Bayesian SVAR) applied to each country separately.
- Model specification:
  - Endogenous variables vector xt = [yt, pt, rt, et], where:
    - yt = real GDP,
    - pt = inflation (headline HICP),
    - rt = policy rate (country-specific definitions in Annex Table 2),
    - et = nominal effective exchange rate (domestic currency per euro; increase = depreciation).
  - Variables are seasonally adjusted and enter in logs, except interest rates.
  - Lag length p = 2 selected using FPE, AIC, BIC, and HQIC.
- Identification: sign restrictions to identify structural shocks.
  - Baseline model: three shocks — demand, supply, and monetary policy shock.
  - Sign restrictions (baseline; contemporaneous and first-quarter responses):
    - Demand shock: GDP +, HICP +, INT +.
    - Supply shock: GDP +, HICP -, INT (blank in table for supply shock under INT—text explains output and inflation move in opposite directions).
    - Monetary policy shock (contractionary, i.e., unexpected increase in interest rate): GDP -, HICP -, INT +.
  - Robustness: four-shock model adding an exchange rate shock with additional restrictions (following Conti et al., 2017).
    - Four-shock sign pattern summarized in Table 3.2:
      - Demand shock: GDP +, HICP +, INT +, NEER -.
      - Supply shock: GDP +, HICP -, INT (blank), NEER -.
      - Monetary policy shock: GDP -, HICP -, INT +, NEER +.
      - Exchange rate shock: GDP +, HICP +, INT +, NEER -.
- Interpretation scaling:
  - Responses interpreted as the impact of a 100-basis point (bps) monetary policy shock on macro variables across countries.
  - Normalization: responses of GDP, HICP, and exchange rates are normalized by the standard deviation of the identified contemporary monetary policy shocks within each country.
  - Impulse responses to a 100-bps shock are obtained via linear scaling of SVAR-estimated responses to a one standard deviation shock.
- Caveats of the sign-restriction SVAR approach:
  - Yields set rather than point identification; multiple decompositions can satisfy restrictions.
  - Restrictions can be weak over finite IRF periods, producing a wide range of admissible responses.
  - Results can be sensitive to choice of priors and implementation method.
  - Assumes structural stability over time; may not hold across crises such as the COVID period.

*IMF Working Paper — Monetary Policy Transmission in Euroized Countries: Evidence from Emerging Europe*

### 4. Results

### 4. Results

### 4.1 Impact on aggregate demand and prices
- Method and scope:
  - Report effects of monetary policy shocks on prices and output by summarizing impulse response functions across countries.
  - Euroization level defined as the average share of FX loans in total loans from 2007m1 to 2023m3.
  - Focus on responses over a relatively short horizon because responses are not significant for longer time horizons (partly due to sign restrictions identification).
- First-quarter responses to a 100-bps monetary policy shock (baseline model):
  - GDP:
    - Around 7 percent fall in benchmark countries, Norway and Sweden.
    - 4 percent fall in Czech Republic, Poland, and Hungary (relatively low euroization levels).
    - 3 percent fall in Romania and Serbia (relatively high euroization levels).
    - Large falls in GDP in North Macedonia and Albania despite very high euroization levels.
  - HICP index (first-quarter fall, quarter-on-quarter percentage change relative to period 0 steady state):
    - 2.5 percent in Norway and Sweden.
    - 2 percent in Czech Republic and Poland.
    - 1.5 percent in Hungary, Romania, and Serbia (higher euroization levels).
    - Albania and North Macedonia: decline in inflation in between Poland and Romania despite higher euroization levels.
- Comparison with literature and interpretation:
  - Ramey (2016) reports a trough effect range for a 100-bps interest rate increase between -0.7 and -5 percent both for output and prices.
  - Conti et al. (2017) find a 100-bps monetary policy shock reduces GDP and HICP by around 2-3 percent.
  - Large magnitudes here could be attributed to the size of the shock (100-bps), noting the standard deviation of monetary shocks across countries ranges from 5-bps to 25-bps.
- Robustness:
  - Second-quarter responses differ less by euroization level and the fall in GDP is generally smaller; no clear pattern for HICP across euroization levels.
  - Exchange rate shock-identified model yields analogous first-quarter results: first-quarter effects on GDP and inflation are inversely correlated with euroization, with Albania and North Macedonia again showing larger effects than euroization levels would suggest.
- Data availability note:
  - Country-by-country impulse response plots and longer-horizon responses provided in Appendix and Annex III.

### 4.2 Monetary policy transmission through the exchange rate
- First-quarter exchange rate responses to monetary policy shocks:
  - Countries with lower euroization levels show larger exchange rate responses to monetary policy shocks.
  - Countries with higher euroization show muted exchange rate responses.
- Interpretation:
  - Reduced exchange rate responsiveness in more euroized economies can help explain attenuated GDP and price responses.
  - The muted response is not a direct mechanical effect of euroization alone; euroization often prompts countries to stabilize exchange rates to the euro.

### 4.3 Monetary policy transmission through the credit channel
- Hypothesis:
  - Transmission can be impaired if firms and households switch to FX loans during tightening episodes as the interest rate differential to other central banks widens, so overall credit growth could slow less in euroized economies.
- Stylized facts for loans to non-financial corporations (NFCs) during the post-pandemic tightening cycle:
  - In some sample countries, the share of FX loans to total outstanding NFC loans increased significantly.
    - Example: In Czech Republic, peak increase in the stock of NFC FX loans since the beginning of the monetary policy tightening in 2021 amounted to 20 percentage points.
  - Cumulative real growth of NFC loans (post-pandemic tightening cycle):
    - In all countries except Albania and North Macedonia, NFC loans in FX grew more strongly (or declined less) than loans in domestic currency.
    - Romania: strong growth of FX loans plus a relatively large initial share of FX loans implied total NFC credit grew during the post-pandemic tightening cycle despite contraction of domestic currency NFC loans.
- Drivers and cross-country heterogeneity:
  - Variation in differences between domestic and ECB policy rates determines differences between domestic and foreign currency loan rates.
  - Difference between domestic and ECB policy rates in the post-pandemic period:
    - Lowest in Albania, followed by North Macedonia and Serbia — these are the most euroized economies in the sample.
    - North Macedonia and Serbia have the most stabilized exchange rates, which could explain high correlation between domestic and ECB policy rates.
  - Resulting mean differences between rates of new domestic and foreign currency loans to NFCs during the post-pandemic tightening cycle are relatively low in Albania and North Macedonia.

### 4.4 Monetary policy transmission through the interest rate channel
- Measurement:
  - Interest rate pass-through measured as sensitivity of bank interest rates on loans and deposits to changes in policy rates.
  - Bank interest rate betas computed as ratio of cumulative increase in bank interest rates on new domestic currency loans to cumulative increase in the policy rate, using the post-pandemic hiking cycle truncated from month of first rate hike until three months after the last hike (or first rate cut or end of sample).
- Pass-through to new loans and mortgages in domestic currency:
  - Pass-through generally higher for more euroized countries in the sample, with exceptions.
    - North Macedonia: high pass-through to mortgages, low pass-through to domestic currency NFC loans (possibly due to significant month-to-month volatility and low volumes).
    - Albania: low pass-through to mortgage rates but relatively high for NFC loans.
- Pass-through to existing (outstanding) mortgage rates:
  - No clear pattern; pass-through correlated with share of flexible rate mortgages.
- Pass-through comparing domestic currency vs euro-denominated loans:
  - Pass-through to domestic currency NFC loans and mortgages (using domestic policy rate) is generally higher in euroized countries than in the euro area (with exceptions like North Macedonia).
  - Pass-through of the domestic policy rate to NFC loans and mortgages in euro is significantly lower (these euro rates do not necessarily respond to domestic policy rate changes).
  - If using the ECB policy rate to calculate pass-through, pass-through to euro NFC loans and mortgages is similar to or above pass-through in the euro area.
- Key implication:
  - Domestic policy rate pass-through to domestic currency loans and mortgages, and ECB rate pass-through to euro-denominated loans and mortgages are both high in euroized economies.
  - Co-movement between domestic policy rate and rates of loans in euro is small, which plausibly drags down aggregate monetary policy transmission in euroized countries.
  - Albania is an exception where co-movement between domestic policy rate and rates of euro NFC loans and mortgages is larger than that of the ECB policy rate — possibly due to relatively low spread between domestic and ECB policy rates.
- Methodological note:
  - The term ‘pass-through’ to euro loan rates may reflect co-movement between domestic and ECB policy rates during a particular episode rather than mechanical pass-through.

### 5. Conclusions and policy implications (from Results section)
- Main comparisons and findings:
  - Aggregate effects on output and prices are correlated with the level of financial euroization.
    - In Norway and Sweden, first-quarter fall in output and inflation is roughly twice as high as in Romania and Serbia (both high euroization).
    - Other countries with lower levels of financial euroization lie between these extremes.
  - Findings robust across models and apply to second-quarter effects.
- Channel-specific conclusions:
  - Exchange rate channel: exchange rates in euroized countries are often stabilized; response to monetary policy shocks is muted, impairing transmission through this channel.
  - Interest rate channel: pass-through to domestic currency retail rates is effective in euroized economies, while rates of foreign currency loans mostly follow the ECB policy rate; thus impact of domestic policy changes depends on degree of policy synchronization with the ECB.
  - Credit channel: domestic monetary tightening can induce substitution toward foreign currency-denominated credit in highly euroized economies, undermining the credit channel.
- Policy implications and recommendations:
  - Successful efforts to reduce the level of euroization would increase the effectiveness of monetary policy.
  - Strategies to reduce euroization should be designed by understanding underlying drivers of financial euroization:
    - Deposit euroization and loan euroization require different strategies (Geng et al., 2018).
    - Deposit euroization can be addressed by establishing a strong track record of prudent macroeconomic policies that reduce perception of the euro as the only reliable hedge.
    - Loan euroization driven in part by interest-rate differentials; allowing more exchange rate flexibility and regulatory measures could make FX loans less attractive.
  - Greater exchange rate flexibility could itself enhance monetary policy transmission and support more effective inflation targeting.
- Suggested avenues for future research:
  - Investigate why Albania and North Macedonia are often outliers: Albania’s strong synchronization with ECB policy rates and relatively flexible exchange rate regime; North Macedonia’s peculiar monetary policy framework.
  - Examine whether exchange rate stabilization triggered by euroization has led to suboptimal monetary policy from an inflation-targeting perspective.
  - Study determinants of euroization (including monetary policy) using microeconomic data or structural models to understand financial institutions' decision-making.
- Methodological caveats:
  - Sign restrictions identification is widely used and sufficient for cross-country comparison here, but identification strategies could be augmented with recent methodological innovations; data limitations restrict choices for some countries.

*IMF Working Paper — Monetary Policy Transmission in Euroized Countries: Evidence from Emerging Europe*

### References

### References

### Bibliography
- Acosta-Ormaechea, S., & Coble, D. (2011). Monetary transmission in dollarized and non-dollarized economies: The cases of Chile, New Zealand, Peru and Uruguay. IMF Working Paper WP/11/87.
- Bekaert, G., Engstrom, E., & Ermolov, A. (2022). Identifying aggregate demand and supply shocks using sign restrictions and higher-order moments. Columbia Business School Research Paper.
- Beyer, R., Chen, R., Li, C., Misch, F., Ozturk, E. O., & Ratnovski, L. (2024). Monetary policy pass-through to interest rates: Stylized facts from 30 European countries, IMF Working Paper WP/24/9.
- Baliño T. J.T., Bennett, A., & Borensztein, E. (1999). Monetary policy in dollarized economies, IMF Occasional Paper No.171.
- Bordon, A. R., & Weber, A. (2010). The transmission mechanism in Armenia: New evidence from a regime switching VAR analysis. IMF Working Paper WP/10/270.
- Brandao-Marques, M. L., Gelos, M. R., Harjes, M. T., Sahay, M. R., & Xue, Y. (2020). Monetary policy transmission in emerging markets and developing economies. IMF Working Paper WP/20/35.
- Breitenlechner, M., Geiger, M., & Sindermann, F. (2019). ZeroSignVAR: A zero and sign restriction algorithm implemented in MATLAB.
- Canova, F., & De Nicolo, G. (2002). Monetary disturbances matter for business fluctuations in the G-7. Journal of Monetary Economics, 49(6), 1131-1159.
- Checo, A, F Grigoli and D Sandri (2024), “Monetary policy transmission in emerging markets: Proverbial Concerns, Novel Evidence”, BIS Working Papers 1170, Bank for International Settlements.
- Conti, A. M., Neri, S., & Nobili, A. (2017). Low inflation and monetary policy in the euro area. ECB Working Paper No.2005.
- Deb, P., Julia E., Melih F., Davide F., & Siddharth K. (2023). Monetary policy transmission heterogeneity: Cross-country evidence. IMF Working Paper WP/23/204.
- della Valle, G., Kota, V., Veyrune, R., Cabezon, E., & Guo, S., (2018). Euroization drivers and effective policy response: An application to the case of Albania. IMF Working Paper WP/18/21.
- Geng, M. N., Scutaru, T., & Wiegand, M. J. (2018). Carry trade vs. deposit-driven Euroization. IMF Working Paper WP/18/58.
- Ilzetzki, E., Reinhart, C. M., & Rogoff, K. S. (2019). Exchange arrangements entering the twenty-first century: Which anchor will hold?. The Quarterly Journal of Economics, 134(2), 599-646.
- Ilzetzki, E., Reinhart, C. M., & Rogoff, K. S. (2021). Rethinking exchange rate regimes. In Handbook of international economics (Vol. 5).
- International Monetary Fund (2023). Romania’s macroeconomic policy mix    through the lenses of the integrated policy framework. Annex of the 2023 Article IV report, IMF.
- Ize, A., & Yeyati, E. L. (2003). Financial dollarization. Journal of International Economics, 59(2), 323-347.
- Kokenyne, A., Ley, J., & Veyrune, R. (2010). Dedollarization. IMF Working Paper WP/10/188.
- Kolasi, G., Shijaku, H., & Shtylla, D. (2010). Monetary transmission mechanism in Albania. Bank of Albania.
- Moder, I. (2023). The transmission of euro area monetary policy to financially euroized countries. Economics & Politics, 35, 718–751.
- Rubio-Ramirez, J. F., Waggoner, D. F., & Zha, T. (2010). Structural vector autoregressions: Theory of identification and algorithms for inference. The Review of Economic Studies, 77(2), 665-696.
- Ramey, Valerie A., Macroeconomic shocks and their propagation (February 2016). NBER Working Paper No. w21978.
- Stojanović, D., & Stojanović, D. (2017). Monetary policy transmission mechanisms in Serbia: evidence from the fully-fledged inflation targeting regime. Post-Communist Economies, 29(1), 117-137.
- Scheiber, T. and Wörz, J. (2018). How are reduced interest rate differentials affecting euroization in Southeastern Europe? Evidence from the OeNB Euro Survey, Focus on European Economic Integration, Oesterreichische Nationalbank (Austrian Central Bank), issue Q1-18, 48-60.
- Uhlig, H. (1994). What macroeconomists should know about unit roots: a Bayesian perspective. Econometric Theory, 10(3-4), 645-671.
- Uhlig, H. (2005). What are the effects of monetary policy on output? Results from an agnostic identification procedure. Journal of Monetary Economics, 52(2), 381-419.
- Uhlig, H. (2017). Shocks, sign restrictions, and identification. Advances in Economics and Econometrics, 2, 95.
- Vargas, M. & Sanchez, J. (2023). Taming financial dollarization: determinants and effective policies–The case of Uruguay. IMF Working Paper WP/23/244.

### Annex I — Data Appendix: Exchange rate regimes (AREAER labels)
- ALB: Floating (independently floating in 1999-2007; floating in 2008-2022)
- CZE: Free floating (managed floating with no predetermined path for the exchange rate in 1999-2000; independently floating in 2001; managed floating with no pre-determined path for the exchange rate in 2002-2006; independently floating in 2007; free floating in 2008-2012; other managed arrangement in 2013; stabilized arrangement in 2014-2016; floating in 2017-2018; free floating in 2019-2022)
- HUN: Floating (crawling band in 1999; pegged exchange rate within horizontal bands in 2000-2006; independently floating in 2007; floating in 2008-2022)
- MKD: Stabilized arrangement (conventional pegged arrangement in 1999-2007; stabilized arrangement in 2008-2022)
- NOR: Free floating (managed floating with no pre-determined path for the exchange rate in 1999-2000; independently floating in 2001-2007; free floating in 2008-2022)
- POL: Free-floating (independently floating in 1999-2007; free-floating in 2008-2022)
- ROU: Crawl-like arrangement (managed floating with no predetermined path for the exchange rate in 1999-2000; crawling band in 2001-2003; managed floating with no predetermined path for the exchange rate in 2004-2007; floating in 2008-2017; stabilized arrangement in 2018; crawl-like arrangement in 2019-2022)
- SRB: Stabilized arrangement (managed floating with no predetermined path for the exchange rate in 2000-2004; exchange arrangement with no separate legal tender in 2005; managed floating with no predetermined path for the exchange rate in 2005-2007; floating in 2008-2015; stabilized arrangement in 2016; crawl-like arrangement in 2017; stabilized arrangement in 2018-2022)
- SWE: Free floating (independently floating in 1999-2007; free floating in 2008-2022)
- Source: IMF The Annual Report on Exchange Arrangements and Exchange Restrictions (AREAER)

### Annex I — Data Appendix: Policy Rates Definition (Annex Table 2)
- ALB: One-week repurchase agreement, percent, quarterly, average
- CZE: repo rate, 2 week, percent per annum, quarterly
- HUN: base rate, percent, quarterly, not seasonally adjusted, average
- MKD: central bank bill rate, percent, quarterly, not seasonally adjusted, average (MKD has the policy rate data available only after 2012m4.)
- NOR: Sight deposit rate, percent per annum, quarterly, average
- POL: Narodowy Bank Polski Base Rates, reference rate, 7-day market operations, percent per annum, quarterly, average
- ROU: monetary policy rate, percent, quarterly
- SRB: 1-week repo rate, percent per annum, quarterly, average
- SWE: Policy rate, percent, quarterly, average

### Annex II — Methodological Details (identification algorithm and estimation)
- The identification algorithm and package follow Breitenlechner, Geiger, and Sindermann (2019).
- Reduced form VAR parameters are estimated using a Bayesian approach.
- Following Uhlig (1994), the reduced form is estimated using an uninformative Normal-Inverse-Wishart prior, yielding a posterior distribution characterized by a Normal-Wishart density.
- Orthogonal impulse responses are obtained by multiplying the reduced form responses with a lower triangular Cholesky factor, which is the product of the variance-covariance matrix and a random orthogonal matrix.
- The orthogonal impulse responses are checked against the sign restrictions defined in the previous section over the contemporary and the first quarter; qualifying orthogonal responses are saved.
- If restrictions are not satisfied, a new draw of the random orthogonal matrix is checked.
- When the maximum of the matrix transformation set in the options is reached and no qualifying matrix could fulfill the restrictions, the algorithm proceeds with another model draw with the Bayesian estimation.

### Annex III — Impulse response plots: Notes (baseline model)
- Each panel shows the impulse response of a row variable (GDP, inflation (HICP), interest rate (INT), or exchange rate (EX)) to a column shock (demand, supply, monetary, or exchange rate).
- For example, the top-left plot shows the response of GDP to a shock in demand.
- Solid lines indicate point-wise medians, dashed lines show the closest-to-median responses.
- The light and dark grey areas represent 90% and two third of the identified posterior distribution, respectively.
- Country-specific impulse response figures are provided for: Albania; Czech Republic; Hungary; North Macedonia; Norway; Poland; Romania; Serbia; Sweden.

*Monetary Policy Transmission in Euroized Countries: Evidence from Emerging Europe — Working Paper No. WP/2026/177*

---


_Source: https://www.imf.org/-/media/files/publications/wp/2025/english/wpiea2025177-print-pdf.pdf_