## Effect of Exchange Rate Movements on Inflation in Sub-Saharan Africa — Chapter 1 & Results

## Source details

**Canonical URL:** [Effect of Exchange Rate Movements on Inflation in Sub-Saharan Africa — Chapter 1 & Results](https://www.imf.org/-/media/files/publications/wp/2024/english/wpiea2024059-print-pdf.pdf)

## Other formats

- [Markdown version](/-/media/files/publications/wp/2024/english/wpiea2024059-print-pdf.pdf.md)
- [Structured JSON version](/-/media/files/publications/wp/2024/english/wpiea2024059-print-pdf.pdf.json)

---

### Purpose and scope
- Focus: estimating the exchange rate passthrough to inflation in sub-Saharan Africa (SSA).
- Objectives:
  - Measure how large and persistent the exchange rate passthrough to inflation is for the region.
  - Explore heterogeneity across exchange rate regimes and country characteristics (resource endowment, competition in domestic market, monetary policy effectiveness).
  - Estimate passthrough for changes in the exchange rate against the US dollar and the nominal effective exchange rate (NEER).
  - Capture dynamic and state-dependent cumulative response functions using monthly data and a proxy for monthly output gap.
  - Test for asymmetries and nonlinearities and document the evolution of passthrough over time.
  - Discuss policy implications for the region.

### Contributions and main findings
- Coverage and methods:
  - Nearly all SSA countries included.
  - Analyze both bilateral exchange rate against the US dollar and NEER.
  - Use local projection methods (Jordà, 2005) with monthly data to generate dynamic, state-dependent cumulative responses and examine speed of passthrough.
  - Study heterogeneity by exchange rate regime, natural resource endowment, market competition, and monetary policy effectiveness.
  - Examine nonlinearities: persistent vs temporary depreciations, effect of large depreciations.
- Key quantitative findings:
  - "A one percentage point increase in the rate of depreciation against the US dollar leads, on average, to an increase in inflation of 0.22 percentage point in a year, compared to 0.15 in emerging Asia and 0.18 in Latin America."
  - Larger depreciations are associated with larger and more persistent passthrough.
  - "The exchange rate passthrough to inflation is asymmetric in the region. During episodes of depreciation, the passthrough to consumer prices is estimated to be eight times stronger than during episodes of appreciation."
  - Passthrough in SSA economies is high but lower than previously estimated by Razafimahefa (2012); passthrough in SSA surpasses that observed in other regions on average.
  - Passthrough has been falling in the region over time.

### Empirical strategy and data
- Methodology:
  - Local projection specification (h ∈ {0,1, ..., 24}):
    - P_i,t+h − P_i,t−1 = μ_i^h + Σ_j γ_j^h ΔP_i,t−j + β_h ΔEXR_i,t + Σ_j θ_j^h X_i,t−j + Σ_j ε_i,t^h
  - ΔEXR is the year-over-year log change in the nominal exchange rate; results reported for both bilateral against the US dollar and NEER.
  - μ_i^h are country fixed effects; β_h traces the impact h months after the shock.
  - Number of lags (l) chosen equal to twelve.
  - Standard errors: Driscoll-Kraay standard errors.
- Controls and external variables:
  - External drivers: annual percent change in international oil, fertilizer, and food prices; Baltic Dry Index for global shipping costs.
  - Country-specific drivers: cyclical component of night lights (proxy for local demand); EM-DAT climate-related events.
- Data sources and sample:
  - Monthly data spanning 2013m1-2022m7.
  - Exchange rates from IMF’s International Financial Statistics.
  - CPI, international oil, fertilizer, food prices, and Baltic Dry Index from Haver Analytics.
  - Night lights and climate events from Hu and Yao (2019) updates and EM-DAT.

### Baseline results and magnitudes
- NEER vs bilateral:
  - The exchange rate passthrough to inflation 12 months after the initial shock is around 0.25 when using the NEER and 0.22 when using bilateral exchange rates.
  - The passthrough reaches 0.37 and 0.31, respectively, after two years.
  - On average in the region, a one percentage point (ppt) increase in the rate of depreciation of the nominal exchange rate against the dollar (NEER) leads to:
    - an increase in inflation by 0.22 (0.25) ppt after one year, and
    - an increase in prices by 0.31 (0.37) percent after two years.
- Interpretation and caveats:
  - Results align with literature showing higher passthrough in EMDEs versus AEs.
  - Price controls in SSA may limit transmission, so estimates could be interpreted as lower bounds.
  - Night lights data availability begins 2013m1; outlier country-years (hyperinflation, excessive exchange rate volatility) excluded.

### Heterogeneity: exchange rate regime, monetary policy, competition, and resources
- Exchange rate regime (IMF de facto classification, 2022):
  - Bilateral exchange rate against the US dollar — twelve months after shock:
    - Managed float regimes: a one ppt increase in rate of depreciation → increase in inflation of 0.20 ppt.
    - Floating regimes: a one ppt increase → increase in inflation of 0.15 ppt.
    - Pegged countries: increase less than 0.08 ppt.
  - NEER-based results — twelve months after shock:
    - Pegged regimes: cumulative response of inflation = 0.19.
    - Managed float regimes: cumulative response = 0.16.
    - Float regimes: cumulative response = 0.18.
  - NEER-based results — two years:
    - Pegged regimes: cumulative passthrough > 0.35.
    - Managed float regimes: 0.23.
    - Float regimes: 0.29.
  - Explanation: trade invoicing and currency of the peg affect which exchange rate measure drives passthrough (example: ~72 percent of trade in WAEMU invoiced in euros).
- Monetary policy effectiveness:
  - Proxy: median of inflation rate in the 2000s to split "high" vs "low" past inflation.
  - Exchange rate passthrough is significantly higher in countries with historically high inflation; difference starker for bilateral USD shocks.
  - First-round passthrough is considered complete regardless of monetary policy effectiveness; weaker credibility amplifies second-round effects.
  - Inflation targeting: little evidence of a difference in SSA between inflation targeting and non-targeting regimes (possible nascent stage of inflation targeting in SSA).
- Level of competition:
  - BTI “Organization of the Market and Competition” subcomponent used (range 0–10); countries split by median.
  - Passthrough is higher when there is less competition; low competition implies market dominance and faster pass-through of cost shocks.
- Natural resource status:
  - Dummy = 1 if net oil exporter.
  - Some evidence of a higher and more persistent passthrough in oil-exporting (and resource-intensive) countries.

### Nonlinearity and asymmetries
- Two nonlinear sources investigated: magnitude of depreciation and persistence of depreciation.
- Magnitude:
  - Threshold: high depreciation when monthly year-on-year percent change in the exchange rate larger than 10.
  - Finding: modest depreciations → inflation increase mostly in first year; high depreciations (>10 percent) → disproportionately larger first-year inflation and remain high into second year.
  - Bilateral USD passthrough during first year is faster and higher during high depreciation episodes.
- Persistence:
  - Persistent episode defined as depreciation >10 percent for more than three consecutive months.
  - Finding: persistent episodes have disproportionately larger impacts on inflation.
- Asymmetry (depreciations vs appreciations):
  - Dummy = 1 if depreciation, 0 otherwise.
  - Finding: During episodes of depreciation, passthrough is estimated to be about eight times stronger—in absolute terms—than during periods of appreciation.
  - Implication: prices do not come down quickly when currencies strengthen after depreciations.
  - Competitiveness channel: more than 70 percent of SSA countries are in the bottom half globally for domestic and foreign competition; mark-ups/profitability higher in SSA firms compared to other emerging markets.

### Declining passthrough over time and determinants
- Rolling 10-year median passthrough (NEER):
  - 2012: 0.25
  - 2022: 0.16
- Rolling 10-year median passthrough (bilateral USD):
  - 2012: 0.16
  - 2022: 0.10
- NEER median by regime:
  - Conventional pegs: 2012 = 0.35; 2022 = 0.13.
  - Managed float: 2013 = 0.25; 2022 = 0.19.
  - Floating regimes: median fell to almost half of its 2012 level (exact medians not restated).
- Regression determinants (annual country-level ERPT, 2012–2022; dynamic panel with IV):
  - ERPT increases with:
    - average inflation (examples: 0.0103*; 0.0129* reported in Table 1 columns).
    - average depreciation (examples: 0.0180***, 0.0175***, 0.0390***, 0.0152**).
    - inflation volatility (examples: 0.0206**, 0.0188*, 0.0356**, 0.0227**).
  - ERPT decreases with:
    - market competition (examples: -0.0140*, -0.0457**, -0.0338**, -0.0319**).
    - central bank independence: Romelli (2022) index between 0 and 1; an increase of 0.1 associated with a drop in the estimated passthrough of 0.1–0.15.
  - Other variables (import share, energy subsidies) show mixed significance across specifications.
  - Lagged ERPT persistence: ERPT(-1) coefficients include 0.755***, 0.718***, 1.078***, 0.791***.
  - Adjusted R-squared examples: 0.402, 0.470, 0.519, 0.528.
  - Statistical significance notation preserved: * p < 0.1, ** p < 0.05, *** p < 0.01.

### Robustness and endogeneity checks
- Endogeneity concern: exchange rate movements endogenous to domestic fundamentals and monetary shocks.
- Instruments:
  - Bilateral USD exchange rate instrumented by US monetary policy shocks (Nakamura and Steinsson 2018).
  - NEER instrumented by IMF global financial conditions index.
- Finding: IV estimates are consistent with baseline results.

### Policy considerations and conclusions
- Exchange rates matter for inflation dynamics in SSA and help explain "fear of floating".
- Large depreciations:
  - Associated with high passthrough and increased risk of inflation de-anchoring.
  - May require central banks to maintain a tight monetary policy stance for a sustained period.
- Monetary policy effectiveness:
  - Passthrough is more muted where inflation has been kept under check historically.
  - Improving monetary policy frameworks and central bank credibility can reduce passthrough.
- Foreign exchange intervention:
  - May be justified where monetary transmission is weak or FX markets shallow, provided sufficient reserves and shocks are temporary or related to financial risk-off episodes.
  - Interventions should weigh costs, including on credibility and FX market functioning.
- Monitoring:
  - Central banks need to monitor both NEER and bilateral USD exchange rates, as they can move in different directions and have distinct inflation implications.
- Structural policies:
  - Promoting competition and economic diversification (especially away from oil dependence) can help lower exchange rate passthrough.

*Source: IMF Working Paper — "Effect of Exchange Rate Movements on Inflation in Sub-Saharan Africa", Chapter 1 (Introduction and Literature Review) and selected Results, Policy Considerations and Conclusion excerpts.*

### 1. Introduction and Literature Review ..................................................................................

### 1. Introduction and Literature Review

### Purpose and scope
- Focus: estimating the exchange rate passthrough to inflation in sub-Saharan Africa (SSA).
- Objectives:
  - Measure how large and persistent the exchange rate passthrough to inflation is for the region.
  - Explore heterogeneity across exchange rate regimes and country characteristics (resource endowment, competition in domestic market, monetary policy effectiveness).
  - Estimate passthrough for changes in the exchange rate against the US dollar and the nominal effective exchange rate (NEER).
  - Capture dynamic and state-dependent cumulative response functions using monthly data and a proxy for monthly output gap.
  - Test for asymmetries and nonlinearities and document the evolution of passthrough over time.
  - Discuss policy implications for the region.

### Literature context
- Cites seminal and recent literature on exchange rate passthrough: Dornbusch (1987), Taylor (2000), Campa and Goldberg (2005), Goldberg and Campa (2010), Gopinath (2015), Gopinath and others (2020).
- Prior SSA-focused work:
  - Razafimahefa (2012): finds incomplete passthrough in SSA; passthrough more pronounced after depreciations and devaluations than appreciations, with the latter having almost no impact due to downward price rigidity; lower passthrough in flexible regimes vs fixed regimes; documented decline in passthrough in SSA since mid-1990s.
  - Country studies (e.g., South Africa: Jooste and Jhavery, 2014; Kabundi and Mbelu, 2018; Kabundi and Mlachila, 2019) find: (1) exchange rate → import prices complete and rapid (first-stage), but partial second-stage passthrough to final consumers; (2) passthrough decreased substantially after adoption of inflation targeting.
  - Mixed-sample studies (2 to 12 SSA countries) conclude: (1) passthrough in developing economies declined from late 1990s to late 2000s and stabilized thereafter, but remains higher than in advanced economies; (2) flexible regimes, low inflation, and monetary policy credibility lower passthrough; (3) macro and geographic characteristics (income, distance to partners, tariffs, wages, competitiveness) matter.

### Contributions and main findings
- Coverage and methods:
  - Nearly all SSA countries included (broader than most prior studies).
  - Analyze both bilateral exchange rate against the US dollar and NEER.
  - Motivations: dominant currency pricing literature and IMF (2023a) evidence that 84 percent of exports from and 67 percent of imports into the median SSA country in 2022 were priced in US dollars.
  - Use local projection methods (Jordà, 2005) with monthly data to generate dynamic, state-dependent cumulative responses and examine speed of passthrough.
  - Study heterogeneity by exchange rate regime, natural resource endowment, market competition, and monetary policy effectiveness.
  - Examine nonlinearities: persistent vs temporary depreciations, effect of large depreciations.
- Key quantitative findings (preserve wording and figures exactly):
  - "A one percentage point increase in the rate of depreciation against the US dollar leads, on average, to an increase in inflation of 0.22 percentage point in a year, compared to 0.15 in emerging Asia and 0.18 in Latin America."
  - Larger depreciations are associated with larger and more persistent passthrough.
  - "The exchange rate passthrough to inflation is asymmetric in the region. During episodes of depreciation, the passthrough to consumer prices is estimated to be eight times stronger than during episodes of appreciation."
  - Passthrough in SSA economies is high but lower than previously estimated by Razafimahefa (2012); passthrough in SSA surpasses that observed in other regions on average (consistent with Ozkan and Erden 2015; Ha and others 2019; Aisen and others 2021).
  - Passthrough has been falling in the region over time.
- Heterogeneity insights:
  - Bilateral exchange rate against the US dollar: countries with non-pegged regimes display a higher passthrough compared to countries with a conventional peg regime.
  - NEER-based estimates: cumulative response of inflation is higher in countries with a conventional peg regime compared to countries with non-pegged regimes.
  - Explanation: while trade is mainly priced in US dollars for most SSA countries, in pegged-regime countries trade is mostly invoiced in the currency of the peg, limiting passthrough from bilateral USD rates.
  - Monetary policy effectiveness reduces passthrough: passthrough effects are weaker in countries with more effective monetary policy; central bank independence is an important factor explaining the declining trend of passthrough over the last decade in SSA.
  - Degree of market competition, inflation volatility, and exchange rate volatility are additional important factors explaining passthrough trends.

### Empirical strategy and data
- Methodology:
  - Use local projection methods (Jordà, 2005) to estimate dynamic cumulative responses of domestic prices to exchange rate movements and to allow for nonlinear/state-dependent effects.
  - Baseline specification follows Gopinath (2015), Caselli and Roitman (2019), Carrière-Swallow and others (2023), and Obstfeld and Zhou (2022).
  - Estimation equation structure (symbols preserved as in source):
    - P_i,t+h − P_i,t−1 = μ_i^h + Σ_j γ_j^h ΔP_i,t−j + β_h ΔEXR_i,t + Σ_j θ_j^h X_i,t−j + Σ_j ε_i,t^h
    - Where h ∈ {0,1, ..., 24}.
  - Key estimation details:
    - ΔEXR is the year-over-year log change in the nominal exchange rate; results reported for both bilateral against the US dollar and NEER.
    - μ_i^h are country fixed effects; β_h traces the impact of exchange rate movements on prices h months after the shock.
    - Number of lags (l) chosen equal to twelve to control for additive seasonal effects.
    - Standard errors: Driscoll-Kraay standard errors used to control for serial correlation and cross-sectional dependence.
- Controls and external variables:
  - External drivers: annual percent change in international oil, fertilizer, and food prices; global shipping costs proxied by the Baltic Dry Index.
  - Country-specific drivers: cyclical component of the night lights indicator of economic activity (proxy for local demand conditions); climate-related events variable from EM-DAT to capture local supply shocks (droughts, floods, insect infestations).
- Data sources and sample:
  - Monthly data from multiple sources spanning 2013m1-2022m7.
  - Exchange rates from IMF’s International Financial Statistics.
  - CPI, international oil, fertilizer, food prices, and Baltic Dry Index from Haver Analytics.
  - Night lights and climate-related events derived using updated data from Hu and Yao (2019) and the international disasters database (EM-DAT).

*Source: IMF Working Paper — "Effect of Exchange Rate Movements on Inflation in Sub-Saharan Africa", Chapter 1 (Introduction and Literature Review).*

### 3. Results

### 3. Results

### 3.1. Baseline Results
- For a given horizon h, the h-month-ahead cumulative percent change in the CPI level attributable to a 1 percent change in the rate of depreciation is shown (solid lines) with 95 percent confidence bands (shaded areas).
- Results based on the NEER and the bilateral exchange rate vis-à-vis the US dollar are statistically similar, except for the first two months where results based on the latter measure are not statistically different from zero.
- Exchange rate passthrough statistics:
  - The exchange rate passthrough to inflation 12 months after the initial shock is around 0.25 when using the NEER and 0.22 when using bilateral exchange rates.
  - The passthrough reaches 0.37 and 0.31, respectively, after two years.
  - On average in the region, a one percentage point (ppt) increase in the rate of depreciation of the nominal exchange rate against the dollar (NEER) leads to:
    - an increase in inflation by 0.22 (0.25) ppt after one year, and
    - an increase in prices by 0.31 (0.37) percent after two years.
- Interpretation and caveats:
  - These results align with empirical literature showing higher exchange rate passthrough to inflation in emerging and developing economies (EMDEs) versus advanced economies (AEs).
  - Potential explanatory factors include less anchored inflation expectations in EMDEs and lower competition (greater pricing power) in domestic markets in EMDEs.
  - Given the pervasiveness of price controls in SSA—limiting transmission of exchange rate movements to prices—the estimates could be interpreted as lower bounds.
  - Night lights indicator data is only available starting from 2013m1; country-years considered outliers (e.g., hyperinflation episodes and excessive exchange rate volatility) are excluded.

### 3.2. Heterogeneity Analyses: Key Characteristics Affecting the Passthrough
- Overview:
  - Average passthrough estimates mask variation across countries driven by characteristics such as exchange rate regime, monetary policy effectiveness, level of competition in domestic product markets, and natural resource status.

A. Exchange Rate Regime
- Classification:
  - Countries grouped into three categories using IMF de facto classification (2022): (i) Pegged regimes (conventional pegs), (ii) Managed float (crawling pegs, crawl-like arrangements and other managed arrangements), (iii) Floating regimes (flexible exchange rate regimes).
- Bilateral exchange rate against the US dollar (Figure 3, left panel) — twelve months after shock:
  - Managed float regimes: a one ppt increase in rate of depreciation → increase in inflation of 0.20 ppt.
  - Floating regimes: a one ppt increase → increase in inflation of 0.15 ppt.
  - Pegged countries: increase less than 0.08 ppt.
- NEER-based results (Figure 3, right panel) — twelve months after shock:
  - Pegged regimes: cumulative response of inflation = 0.19.
  - Managed float regimes: cumulative response = 0.16.
  - Float regimes: cumulative response = 0.18.
- NEER-based results — two-year horizon:
  - Pegged regimes: cumulative passthrough to consumer prices after two years is more than 0.35.
  - Managed float regimes: 0.23.
  - Float regimes: 0.29.
- Interpretation:
  - Lower bilateral-US-dollar passthrough in pegged regimes may reflect a smaller portion of imports invoiced in US dollars; trade in pegged regimes is often invoiced in the currency of the peg.
  - Example: around 72 percent of trade in the West African Economic and Monetary Union is invoiced in euros.

B. Monetary Policy Effectiveness
- Method:
  - Effectiveness proxied by central bank’s track record at delivering price stability: use the median of inflation rate in the 2000s to split sample into “high” and “low” past inflation bins.
  - State-dependent estimation (equation (2)) includes dummy = 1 when monetary policy considered “effective”.
- Findings:
  - Exchange rate passthrough is significantly higher in SSA countries with historically high inflation compared to countries with a good track record of low inflation.
  - The difference is starker when considering shocks on the bilateral exchange rate to the US dollar.
- Mechanisms:
  - Direct (first-round) channels: imported consumption and cost channels—exchange rate changes affect prices of imported goods and imported intermediate inputs, which alter domestic CPI via final good prices and production cost pressures.
  - Indirect (second-round) channels: exchange rate depreciation affects asset prices and inflation expectations, leading to broader and more entrenched inflation—more pronounced where monetary policy credibility is weak.
- Note:
  - There is consensus in the literature that the first-round passthrough is complete, regardless of monetary policy effectiveness.

- Inflation targeting as a policy-framework proxy:
  - Empirical literature often finds inflation targeting lowers passthrough in other contexts.
  - SSA results (Figure 5): little evidence of a difference in exchange rate passthrough estimates between inflation targeting and non-targeting regimes.
  - Possible explanation: nascent stage of inflation targeting regimes in SSA (exceptions include South Africa).

C. Level of Competition in the Domestic Products Market
- Measurement:
  - Use Bertelsmann-Stiftung Transformation Index (BTI) “Organization of the Market and Competition” subcomponent (range 0–10).
  - Countries split into “high competition” if average BTI score over 2006-2022 is above the sample median, and “low competition” if below the median.
  - Robustness check: Global Competition Index yields similar short-run results.
- Findings:
  - Exchange rate passthrough is higher when there is less competition in the domestic products market.
  - Interpretation: low competition often implies market dominance by a few, less-productive firms with mark-up power that pass exchange rate-induced cost shocks rapidly to consumer prices.

D. Natural Resource Status
- Hypothesis:
  - Natural resource dependence (e.g., net oil exporters) can affect passthrough via exchange rate overvaluation during resource booms, greater reliance on imported goods, and weaker institutions.
- Estimation:
  - State-dependent version of equation (1) with dummy = 1 if country is a net oil exporter.
- Findings:
  - Some evidence of a higher and more persistent passthrough in oil-exporting countries relative to other country groups.
  - Robustness: using a dummy for resource-intensive countries rather than just net oil exporters leaves results virtually unchanged.

*IMF Working Paper — “Effect of Exchange Rate Movements on Inflation in Sub-Saharan Africa”*

### Annex Figure 4). These results may reflect the fact that non-resource-intensive countries generally have more

### Annex Figure 4). These results may reflect the fact that non-resource-intensive countries generally have more

### Nonlinearity and asymmetries in exchange rate passthrough
- Literature and SSA evidence indicate asymmetric price responses: depreciations generally transmit to inflation more strongly than appreciations (references in text: Pollard and Coughlin 2004; Bussière 2013; Frankel and others 2012; Caselli and Roitman 2019; Razafimahefa 2012).
- Two sources of nonlinearity investigated:
  - Magnitude of depreciations (high vs. modest).
  - Persistence of depreciation episodes (persistent/long-lived vs. temporary/shorter-lived).
- Firms/importers transmit movements to prices more when movements are perceived as long-lasting.

### A. Magnitude of depreciations
- Historical context: except for some countries (Angola, DRC, Zambia, Ghana and Mozambique), depreciations above 20 percent are infrequent across SSA (see Annex Figure 1).
- Thresholds used in estimation:
  - High depreciation episode: monthly year-on-year percent change in the exchange rate larger than 10.
  - Dummy constructed: (high_depr_i,t).
- Empirical findings (Figure 8, NEER and bilateral against USD):
  - When depreciations are modest, the increase in inflation occurs mostly in the first year after the shock.
  - At higher levels of depreciation (>10 percent), the percentage increase in inflation is disproportionately larger in the first year and remains high into the second year after the shock.
  - For bilateral exchange rate against the US dollar, passthrough during the first year is faster and higher during high depreciation episodes.
- Policy implication: Large depreciations could be associated with considerable risks of inflation de-anchoring in SSA, especially where inflation is already high and central bank credibility is weak.

### B. Persistence of depreciation shocks
- Definition: persistent episode defined as exchange rate depreciates by more than 10 percent for more than three consecutive months.
- Dummy constructed for persistent episodes substituted into equation (3).
- Finding (Figure 9): Persistent episodes tend to have disproportionately larger impacts on inflation—importers and firms more rapidly transmit perceived long-lasting depreciations into prices.

### C. Depreciations vs. appreciations (asymmetry)
- Dummy constructed for sign of exchange rate shock: 1 if depreciation, 0 otherwise.
- Empirical finding (Figure 10): Exchange rate passthrough to inflation is asymmetric in SSA:
  - During episodes of depreciation, passthrough is estimated to be about eight times stronger—in absolute terms—than during periods of appreciation.
  - Implication: Prices may not come down quickly when local currencies strengthen after periods of depreciation.
- Competitiveness channel:
  - Lower competition (higher market power) leads firms to pass on depreciations more and not fully pass back appreciations.
  - Evidence: more than 70 percent of SSA countries fall in the bottom half globally in terms of domestic and foreign competition (Cherif and others 2020).
  - Firm-level findings: profitability and mark-ups generally higher in SSA compared to other emerging market economies.

### 3.4 Declining exchange rate passthrough over time
- Rolling 10-year country-by-country estimates (starting 2002) of ERPT show a general decline across the region.
- Median passthrough (NEER):
  - 2012: 0.25
  - 2022: 0.16
- Median passthrough (bilateral USD):
  - 2012: 0.16
  - 2022: 0.10
- Heterogeneities by exchange rate regime (NEER, median values):
  - Conventional pegs: median passthrough dropped from 0.35 in 2012 to 0.13 in 2022.
  - Floating regimes: median passthrough fell to almost half of its 2012 level (exact medians not restated beyond “almost half” in text).
  - Managed float: median decreased from 0.25 in 2013 to 0.19 in 2022.
- The pace of decline is irregular and heterogeneous across country groups.

### Determinants of ERPT (regression analysis and key quantitative findings)
- Regression framework: annual country-level passthrough estimates (2012–2022) regressed on determinants with a lagged dependent variable and panel IV using Anderson and Hsiao (1981) approach; equation (4) specified:
  - ERPT_it = α_i + γ_t + δ ERPT_it−1 + β_j X_it + ε_it
- Potential determinants (X): average inflation, inflation volatility, average exchange rate fluctuations, exchange rate volatility, central bank credibility (Romelli 2022 index), level of competition (BTI score), energy subsidies (% of GDP), import content of consumption (imports/total final consumption).
- Main empirical relationships (summary of Table 1 and text):
  - ERPT increases with:
    - average inflation (coefficients positive and often statistically significant).
    - average depreciation (size of exchange rate depreciation) — e.g., average depreciation coefficients reported as 0.0180***, 0.0175***, 0.0390***, 0.0152** (see Table 1 columns).
    - inflation volatility (coefficients positive and frequently significant; examples in Table 1: 0.0206**, 0.0188*, 0.0356**, 0.0227**).
  - ERPT decreases with:
    - improvements in domestic market competition: market competition coefficients negative and statistically significant (examples in Table 1: -0.0140*, -0.0457**, -0.0338**, -0.0319**).
    - central bank independence: described quantitatively in text — an increase of 0.1 in the central bank independence index is associated with a drop in the estimated passthrough of 0.1–0.15.
      - Romelli (2022) index range: between 0 and 1 (note in text: an increase of 0.1 can be interpreted as an increase of one unit).
  - Import share and energy subsidies show mixed significance across specifications in Table 1.
  - Lagged ERPT exhibits persistence: ERPT(-1) coefficients reported (Table 1) as 0.755***, 0.718***, 1.078***, 0.791*** in different specifications.
- Robustness: Instrumental variable (IV) and OLS specifications reported; IV approach instruments ERPT(-1) with ERPT(-2). Results generally confirm baseline patterns.

### 4. Robustness checks (endogeneity and instruments)
- Concern: exchange rate movements may be endogenous to domestic fundamentals including inflation; expansionary monetary shocks can cause both depreciation and higher inflation.
- Identification strategy:
  - Instrument bilateral exchange rate to the US dollar by US monetary policy shocks (Nakamura and Steinsson 2018).
  - Instrument NEER by IMF global financial conditions index for NEER estimations.
- Finding (Figure 13): IV estimates are consistent with baseline results.

### Key quantitative thresholds, definitions, and values preserved exactly
- Historical infrequency of depreciations above 20 percent across most SSA countries (exceptions listed: Angola, DRC, Zambia, Ghana and Mozambique).
- High depreciation threshold used in analysis: monthly year-on-year percent change in the exchange rate larger than 10.
- Persistent depreciation episode defined as depreciation >10 percent for more than three consecutive months.
- Median passthrough (NEER): 2012 = 0.25; 2022 = 0.16.
- Median passthrough (bilateral USD): 2012 = 0.16; 2022 = 0.10.
- Pegged countries median passthrough (NEER): 2012 = 0.35; 2022 = 0.13.
- Managed float median passthrough (NEER): 2013 = 0.25; 2022 = 0.19.
- Central bank independence index: Romelli (2022) takes values between 0 and 1; an increase of 0.1 associated with a drop in passthrough of 0.1–0.15.
- Selected coefficients from Table 1 (Nominal Effective Exchange Rate specifications, examples):
  - Market competition: -0.0140* (OLS), -0.0457** (OLS IV specification).
  - Average inflation: 0.0103* (OLS), 0.0129* (other OLS specification).
  - Inflation volatility: 0.0206** (OLS), 0.0188* (other OLS).
  - Average depreciation: 0.0180***, 0.0175***, 0.0390***, 0.0152** (various columns).
  - ERPT(-1): 0.755***, 0.718***, 1.078***, 0.791*** (various columns).
  - Adjusted R-squared examples: 0.402, 0.470, 0.519, 0.528 (selected columns).
- Statistical significance notation preserved as in source: * p < 0.1, ** p < 0.05, *** p < 0.01.

*Source: IMF Working Paper text excerpt (figures and tables referenced in text).*

### 5. Policy Considerations and Conclusion

### 5. Policy Considerations and Conclusion

### Role of exchange rates in inflation dynamics
- Exchange rates play an important role in inflation dynamics in SSA, which helps to understand the fear of floating among policy makers.
- Large exchange rate depreciations are associated with high passthrough and can increase the risk of inflation de-anchoring.
- Macroeconomic policies that can create significant exchange rate pressures such as monetization of fiscal deficits or fiscal policies that fuel external imbalances can have significant inflationary costs through the exchange rate channel.

### Monetary policy effectiveness and implications
- Exchange rate passthrough to inflation is more muted in countries where inflation has been kept under check historically.
- Improving monetary policy frameworks, and a proactive role of central banks in combating inflation can help reduce the exchange rate passthrough.
- Inflationary impulses from large depreciations can be persistent, and hence may require central banks maintaining a tight monetary policy stance for a sustained period.
- In countries where monetary policy framework and transmissions are weak or foreign exchange markets are shallow, significant risk of inflation de-anchoring from large exchange rate depreciations can justify foreign exchange interventions by the central banks provided they have sufficient reserve and the shocks are temporary in nature or due to financial risk-off episodes (see IMF, 2022).
- Such foreign exchange interventions should carefully consider the costs involved including on credibility of the monetary policy framework and functioning of the foreign exchange market. IMF (2023b) provides guidance on how foreign exchange intervention should be used.

### Asymmetries and monitoring exchange rate measures
- Estimated asymmetries in the exchange rate passthrough imply that prospects of inflation coming down when the exchange rate is appreciating is more muted than the possibility of inflation increasing when the exchange rate is depreciating.
- Central banks may need to take a proactive role in anchoring inflation during episodes of significant currency volatility.
- Changes in the nominal exchange rate against the US dollar—the key currency of trade invoicing—are an important source of inflation in addition to changes in the NEER, particularly in countries with non-pegged exchange rate regimes.
- NEER and bilateral exchange rate against the US dollar can move in different directions (as in many SSA countries in 2022); central banks need to monitor movements of both types of exchange rates to gauge inflation risks.

### Structural factors and policy recommendations
- Higher levels of competition in domestic and more diversified economies, particularly countries that do not export oil, tend to have lower exchange rate passthrough.
- Structural policies that promote competition in domestic markets and greater diversification may help lower the exchange rate passthrough in the region.

*Source: 5. Policy Considerations and Conclusion, Effect of Exchange Rate Movements on Inflation in Sub-Saharan Africa, Working Paper No. WP/2024/059*

---


_Source: https://www.imf.org/-/media/files/publications/wp/2024/english/wpiea2024059-print-pdf.pdf_