## wpiea2023039-print-pdf

## Source details

**Canonical URL:** [wpiea2023039-print-pdf](https://www.imf.org/-/media/files/publications/wp/2023/english/wpiea2023039-print-pdf.pdf)

## Other formats

- [Markdown version](/-/media/files/publications/wp/2023/english/wpiea2023039-print-pdf.pdf.md)
- [Structured JSON version](/-/media/files/publications/wp/2023/english/wpiea2023039-print-pdf.pdf.json)

---

### Context and definition of supply chain disruptions
- Supply chain disruptions: a business’ inability to receive, produce, ship, and sell their products.
- Two key dimensions emphasized: disruptions in manufacturing production and higher transportation costs.
- Global Supply Chain Pressure Index (GSCPI) combines cross-border transportation cost indicators and global manufacturing data to capture these dimensions.
- GSCPI signals:
  - Disruptions in 2021 and 2022 were three to four standard deviations above the historical average.
  - During the pandemic, global supply chain pressures peaked at almost 4.5 standard deviations in late-2022.
- Shipping-cost developments in 2021: the Baltic Dry Index increased fourfold and the Harpex Index increased sevenfold from pre-pandemic levels.

### Motivation and policy challenge
- Monetary policy concerns:
  - First-round direct effects on prices of tradable manufactured goods.
  - Second-round effects that could entrench higher inflation expectations and spill into non-tradable goods and services.
- Specific risks for sub-Saharan Africa:
  - High shares of food and tradable goods in consumption basket and relatively weak monetary credibility raise the risk of unanchored inflation expectations.
- Standard advice for commodity supply shocks may not apply because global supply chain disruptions directly affect tradable manufactured goods and core consumption.

### Data, sample, and measurement
- Monthly panel of 29 sub-Saharan African countries for the period 2000-22.
- CPI disaggregation:
  - Tradable inflation weights: food (36%), transport (11%), clothing (7%), furnishing (5%), alcohol (4%).
  - Non-tradable inflation weights: housing (13%), restaurant (6%), communication (5%), health (3%), education (3%), recreation (3%).
- Supply chain pressure measure: GSCPI from the Federal Reserve Bank of New York (integrates Baltic Dry Index, Harpex Index, air transport costs, and manufacturing PMI components).
- Summary sample statistics (monthly observations, N=4246 unless noted):
  - Headline inflation (yoy % chg): Mean 5.80, Std. Dev. 6.86, Min -6.24, Max 58.42.
  - Tradable core inflation (yoy % chg): Mean 5.53, Std. Dev. 7.34, Min -8.21, Max 60.07.
  - Non-tradable inflation (yoy % chg): Mean 5.07, Std. Dev. 6.90, Min -8.54, Max 60.31.
  - Food inflation (yoy % chg): Mean 6.73, Std. Dev. 8.92, Min -15.45, Max 76.78.
  - GSPCI: Mean 0.36, Std. Dev. 1.22, Min -1.52, Max 4.35.
  - Global food price (yoy % chg): Mean 3.70, Std. Dev. 12.54, Min -32.23, Max 33.70.
  - Global oil price (yoy % chg): Mean 4.70, Std. Dev. 38.10, Min -97.91, Max 98.87.
  - Real GDP growth: Mean 3.75, Std. Dev. 3.87, Min -14.90, Max 14.60.
  - World output gap: Mean -1.00, Std. Dev. 1.25, Min -3.92, Max 1.96.
  - Output gap: Mean -0.02, Std. Dev. 1.38, Min -7.74, Max 6.60.
  - Global policy rate: Mean 1.07, Std. Dev. 1.43, Min 0.05, Max 6.54.
  - Monetary policy rates (N=3164): Mean 7.25, Std. Dev. 3.71, Min 0.25, Max 33.50.
  - VIX Uncertainty Index: Mean 19.74, Std. Dev. 8.13, Min 9.51, Max 59.89.
  - Nominal exchange rates (yoy % chg) (N=4186): Mean 4.31, Std. Dev. 14.01, Min -36.89, Max 109.60.

### Empirical approach and identification
- Methodology:
  - Local projections (Jordà (2005)) to compute impulse response functions (IRFs) to a one standard deviation shock.
  - Decomposition of realized inflation into three shocks: global supply chain disruptions (GSCPI), global food prices, and global oil prices (following Benigno et al (2022)).
  - IRFs estimated for horizons h = 0,...,15 months with lags j = 12; OLS with heteroskedasticity-robust standard errors clustered at country level; 90 percent confidence intervals.
- Tradable core defined to include alcohol, clothing, household furnishings, and transportation; non-tradable core includes health, communications, recreation, education, restaurants, housing.

### Key empirical results (local projections)
- One-standard deviation shock to the GSCPI:
  - Headline inflation:
    - Increases headline inflation by 0.41 percentage points five months after the initial shock.
    - Effect reverts to zero after 14 months.
    - Table 2 All controls coefficients for headline inflation (selected): h=5: 0.41 (std. err. 0.15); h=6: 0.36 (0.12); h=7: 0.32 (0.10); h=8: 0.30 (0.13).
  - Food inflation:
    - Increases food inflation by 0.54 percentage points five months after the shock; coefficient not statistically significant in the specification with controls.
    - No controls coefficients for food (selected): h=5: 0.62 (0.22) **; h=6: 0.66 (0.19) **; h=7: 0.76 (0.18) ***; h=8: 0.87 (0.17) ***.
  - Tradable core inflation:
    - Increases tradable inflation by 0.49 percentage points after five months with a 10 percent significance level.
    - No controls coefficients for tradable (selected): peak around h=8: 0.58 (0.12) ***.
  - Non-tradable core inflation:
    - Smaller and delayed effect: flat first three months, gradual increase from month 4, peak at month 7 with smaller and generally not statistically significant coefficients (No controls h=7: 0.26 (0.11) *).
- Comparative shocks:
  - Global oil price shock: positive, sizeable, and persistent impacts across headline, food, tradable, and non-tradable inflation; headline impact materializes in first three months and reverts seven months after shock.
  - Global food price shock: effect on headline peaks after eight months; food inflation rises quickly until month five and persists until month 12.
- Overall interpretation:
  - Global supply chain pressures have a sizable, positive, persistent, and statistically significant impact on headline, food, and tradable core inflation; second-round effects on non-tradables are smaller and delayed.

### Quantification of contributions during 2020-22
- Supply chain pressures pushed up headline, food, and tradable core inflation by about 2 percentage points during the 2020-22 period.
- Disruptions accounted for:
  - 45 percent of headline inflation during 2020-22.
  - 55 percent of tradable core inflation during 2020-22.
- Decomposition results (average median inflation rate contributions, 2020-22):
  - Supply chain pressures: 45 percent of headline; Food inflation: 40 percent contribution from GSCPI; Tradable inflation: 55 percent; Non-tradable inflation: 40 percent.
- Role of oil and food prices over 2020-22:
  - Oil prices helped keep inflation in check from 2020 up to Q4 2021, then contributed considerably to the rise in headline inflation thereafter.
  - Food prices contributed to lower inflation from 2020 to early 2021, positively to headline inflation from early 2021 until the beginning of 2022, and negatively to headline inflation in the first half of 2022.

### Robustness checks and policy interpretation
- Additional controls tested: monetary policy rates (global and country-specific), VIX index, nominal effective exchange rate (NEER).
- Key robustness findings:
  - Including monetary policy rates: the impact of GSCPI on headline inflation disappears, suggesting monetary authorities respond to inflation caused by supply chain disruptions.
  - Including VIX or NEER: headline inflation response follows the baseline pattern.
- Policy interpretation:
  - Supply chain disruptions transmit to domestic inflation mainly through tradable and food prices with limited immediate pass-through to non-tradable inflation.
  - Monetary policy responses appear to offset inflationary effects from global supply chain pressures when accounting for monetary policy rates.

### Structural model and monetary policy implications
- Model: New Keynesian small open economy with tradable and non-tradable sectors, price rigidities, and a supply curve linking tradable inflation to global supply chain disruptions.
- Key mappings and equations:
  - Headline inflation: π_t ≡ (1−ω) π_t^N + ω π_t^T.
  - Supply curve for tradable inflation: π_t^T = λ π_{t−1}^T + ∑_{τ=0}^μ τ ⋅ ΔC_{t−1−τ}^T.
  - Open-economy Phillips curve for non-tradable inflation: π_t^N = α E(π_t^N) + β x_{t−1} + ε ω π_{t−1}^T.
  - IS equation: x_t = ζ x_{t−1} − η (R_t − E(π_t) − r̄).
  - Global supply chain disruption: ΔC_t = ν ΔC_{t−1} + ι_t.
  - Central bank objective: ∑_t ξ x_t^2 + (1−ξ) (π_t − π̄)^2.
  - Inflation expectations: E(π_t^N) = γ π̄ + (1−γ) π_{t−1}^N, with γ = 0 (adaptive, low credibility) or γ = 0.5 (anchored, high credibility).
- Calibration (selected parameters):
  - ω — Share of tradable goods in consumption basket: 0.4 or 0.7.
  - λ — Persistence of tradable inflation: 0.5.
  - α — Persistence of non-tradable inflation: 0.99.
  - β — Slope of core Phillips curve for non-tradable inflation: 0.2 (annual).
  - ε — Second-round effects of global shipping costs on non-tradable inflation: 0.5.
  - ζ — Persistence of output gap: 0.5.
  - η — Elasticity of output gap to real policy rate: 1 (annual).
  - ν — Persistence of shock to global shipping costs: 0 or 0.5.
  - ξ — Weight of output gap in the central bank's objective function: 0.5.
- Policy insights:
  - A trade-off emerges between stabilizing inflation and closing the output gap in response to supply chain disruptions.
  - Central banks with low credibility may need large policy-rate movements to stabilize inflation and output.
  - Pre-emptive tightening (raising policy rates as supply chain disruptions increase) lowers second-round effects on non-tradable inflation and inflation expectations, improving joint stabilization of inflation and output compared with delayed tightening.
  - Tighter policy cannot prevent direct tradable-price effects but can reduce second-round effects; forward-looking tightening is especially effective in low-income countries where tradable goods are a larger share of the consumption basket and second-round effects are likelier.

*Source: wpiea2023039-print-pdf - IMF Working Paper "Global Supply Chain Disruptions: Challenges for Inflation and Monetary Policy in SSA" (selected chapters and appendix).*

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

### 1. Introduction

### Context and definition of supply chain disruptions
- Supply chain disruptions: a business’ inability to receive, produce, ship, and sell their products.
- Two key dimensions emphasized: disruptions in manufacturing production and higher transportation costs.
- The Global Supply Chain Pressure Index (GSCPI) combines cross-border transportation cost indicators and global manufacturing data to capture these dimensions.
- The GSCPI suggests disruptions in 2021 and 2022 were three to four standard deviations above the historical average.
- During the pandemic, global supply chain pressures peaked at almost 4.5 standard deviations in late-2022.
- Shipping-cost developments in 2021: the Baltic Dry Index increased fourfold and the Harpex Index increased sevenfold from pre-pandemic levels.

### Motivation and policy challenge
- Supply chain volatility and the prospect of sustained disruptions pose significant challenges for monetary policymakers.
- Concerns:
  - First-round direct effects on prices of tradable manufactured goods.
  - Second-round effects that could entrench higher inflation expectations and spill into non-tradable goods and services.
  - High shares of food and tradable goods in sub-Saharan Africa’s consumption basket and relatively weak monetary credibility raise the risk of unanchored inflation expectations.
- Standard policy advice for commodity supply shocks (accommodate first-round but not second-round effects) may not apply directly because global supply chain disruptions directly affect tradable manufactured goods and core consumption.

### Empirical approach
- Sample and frequency:
  - Panel of 29 countries in sub-Saharan Africa at a monthly frequency.
- Data and identification:
  - Detailed CPI component dataset allowing separation of headline, food, tradable core, and non-tradable core inflation.
  - Tradable core defined to include alcohol, clothing, household furnishings, and transportation; non-tradable core includes health, communications, recreation, education, restaurants, housing.
- Methodology:
  - Impulse response functions (IRFs) computed to a one standard deviation shock using local projections following Jordà (2005).
  - GSCPI used as proxy for supply chain pressures.
  - Decomposition of realized inflation into three shocks: global supply chain disruptions, global food prices, and global oil prices (following Benigno et al (2022)).

### Key empirical findings (local projections)
- A one-standard deviation shock to the GSCPI:
  - Increases headline inflation by 0.41 percentage points five months after the initial shock.
  - Headline inflation effect reverts to zero after 14 months.
  - Increases both food inflation and tradable core inflation by about half a percentage point five months later.
  - Generates indirect second-round effects: non-tradable core inflation increases by 0.2 percentage points on average seven months after the shock (smaller and delayed relative to tradable inflation).
- Magnitude and persistence: global supply chain pressures have a sizable, positive, persistent, and statistically significant impact on all measures of domestic inflation in the sample.

### Quantification of contributions during 2020-22
- Supply chain pressures pushed up headline, food, and tradable core inflation by about 2 percentage points during the 2020-22 period.
- Disruptions accounted for:
  - 45 percent of headline inflation during 2020-22.
  - 55 percent of tradable core inflation during 2020-22.

### Modeling and monetary policy implications
- Model used:
  - New Keynesian small open economy model combining:
    - A supply curve linking tradable inflation to the direct impact of global supply chain disruptions.
    - An expectations-augmented Phillips curve linking non-tradable inflation to the output gap and second-round effects from global supply chain disruptions.
  - Calibrated to empirical findings to study policy responses.
- Policy insights:
  - A trade-off emerges between stabilizing inflation and closing the output gap in response to supply chain disruptions.
  - Central banks with low credibility may need large policy-rate movements to stabilize inflation and output.
  - Pre-emptive tightening (raising policy rates as supply chain disruptions increase) lowers second-round effects on non-tradable inflation and inflation expectations, improving the joint stabilization of inflation and output compared with delayed tightening.
  - Tighter policy cannot prevent the direct tradable-price effects but can reduce second-round effects; forward-looking tightening is especially effective in low-income countries where tradable goods are a larger share of the consumption basket and second-round effects are likelier.

### Relation to existing literature and contribution
- Links to three strands:
  - Impact of shipping costs and supply chain disruptions on inflation (e.g., Benigno et al (2022); LaBelle and Santacreu (2022); Finck and Tillmann (2022); Carrière-Swallow et al (2022)).
  - Global drivers of inflation, notably pass-through from global commodity price shocks to domestic inflation (e.g., Gelos and Ustyugova 2017; IMF 2011).
  - Optimal monetary policy under global supply or commodity shocks (e.g., Habermeier et al (2009); IMF (2011); De Gregorio (2012)).
- This paper’s contributions:
  - Examines first- and second-round effects of global supply chain disruptions on tradable and non-tradable core inflation across many emerging and low-income economies.
  - Shows global supply chain shocks differ from commodity price shocks due to outsized impacts on tradable components of core inflation.
  - Offers the first study of optimal monetary policy specifically in the face of global supply chain shocks.

### Additional contextual facts from the introduction
- IMF survey: sub-Saharan African countries implemented more than 100 policy measures since January 2022 in response to food and fuel price shocks.
- Empirical estimation leverages the GSCPI and detailed CPI component data to separate channels and time profiles of impacts.

*Source: wpiea2023039-print-pdf - 1. Introduction — https://www.imf.org/-/media/files/publications/wp/2023/english/wpiea2023039-print-pdf.pdf*

### 2.2 The inflation outlook in sub-Saharan Africa

### 2.2 The inflation outlook in sub-Saharan Africa

### Overview of recent inflation dynamics
- Median headline inflation: remained around 3 percent throughout 2020 and rose to about 8.5 percent in June 2022.
- Median core inflation: increased starting in 2021 and peaked at 6.25 percent in June 2022.
- Food: represents 36 percent of the region’s median consumption basket and contributed about one-half to headline inflation during the pandemic.
- Sudden sharp increase in tradable inflation in 2021 and 2022 is linked to global supply chain disruptions.
- Inflation of non-tradable goods: remained relatively steady since 2020.

### Data, measurement, and sample
- Monthly panel of 29 sub-Saharan African countries for the period 2000-22.
- Median inflation calculated using a sample of 29 sub-Saharan African economies during 2000-22.
- Supply chain pressure measure: GSCPI (Global Supply Chain Pressure Index) from the Federal Reserve Bank of New York.
  - GSCPI integrates goods manufacturing and international transportation components.
  - Transportation cost inputs: Baltic Dry Index (bulk), Harpex Index (container), and air transport costs (Bureau of Labor and Statistics).
  - PMI components used: delivery time, backlogs and inventories from manufacturing PMIs across seven economies.
- CPI disaggregation: twelve standard components used to construct tradable and non-tradable inflation.
  - Tradable inflation weights: food (36%), transport (11%), clothing (7%), furnishing (5%), alcohol (4%).
  - Non-tradable inflation weights: housing (13%), restaurant (6%), communication (5%), health (3%), education (3%), recreation (3%).
- Summary sample statistics (monthly observations, N=4246 unless noted):
  - Headline inflation (yoy % chg): Mean 5.80, Std. Dev. 6.86, Min -6.24, Max 58.42.
  - Tradable core inflation (yoy % chg): Mean 5.53, Std. Dev. 7.34, Min -8.21, Max 60.07.
  - Non-tradable inflation (yoy % chg): Mean 5.07, Std. Dev. 6.90, Min -8.54, Max 60.31.
  - Food inflation (yoy % chg): Mean 6.73, Std. Dev. 8.92, Min -15.45, Max 76.78.
  - GSPCI: Mean 0.36, Std. Dev. 1.22, Min -1.52, Max 4.35.
  - Global food price (yoy % chg): Mean 3.70, Std. Dev. 12.54, Min -32.23, Max 33.70.
  - Global oil price (yoy % chg): Mean 4.70, Std. Dev. 38.10, Min -97.91, Max 98.87.
  - Real GDP growth: Mean 3.75, Std. Dev. 3.87, Min -14.90, Max 14.60.
  - World output gap: Mean -1.00, Std. Dev. 1.25, Min -3.92, Max 1.96.
  - Output gap: Mean -0.02, Std. Dev. 1.38, Min -7.74, Max 6.60.
  - Global policy rate: Mean 1.07, Std. Dev. 1.43, Min 0.05, Max 6.54.
  - Monetary policy rates (N=3164): Mean 7.25, Std. Dev. 3.71, Min 0.25, Max 33.50.
  - VIX Uncertainty Index: Mean 19.74, Std. Dev. 8.13, Min 9.51, Max 59.89.
  - Nominal exchange rates (yoy % chg) (N=4186): Mean 4.31, Std. Dev. 14.01, Min -36.89, Max 109.60.

### Theoretical framework and empirical strategy
- Small open economy with tradable (T) and non-tradable (N) sectors; headline CPI P_it = (P_it^T)^{e_T} (P_it^N)^{1−e_T}, where e_T is tradable share of consumption.
- Import share m_i and shipping cost shock modeled as increase in τ_t^S > 1.
- Change in price index in response to supply chain shock:
  - dlnP_it = (e_T m_i + (1−e_T) γ_NT) dlnτ_S + ε_it
  - γ_NT is share of input expenditures on tradable sector by the domestic non-tradable sector.
  - ε_it captures residual global and domestic shocks.
- Empirical approach: local projections (Jordà 2005) estimating year-over-year log changes in inflation horizons h = 0,...,H with H = 15 months, lags j = 12, OLS, heteroskedasticity-robust standard errors clustered at country level, 90 percent confidence intervals for IRFs.

### Main empirical results: impact of GSCPI shocks on inflation
- Pandemic peak: global supply chain pressures peaked at almost 4.5 standard deviations in late 2022.
- Headline inflation:
  - One standard deviation increase in GSCPI increases headline inflation by 0.41 percentage points five months after the shock (Table 2, panel A).
  - With no controls, impact statistically significant from month 5 through month 12.
  - With full controls (global oil and food prices, country-specific output gaps, world output gap) the impact is significant at the 10 percent level from month 5 to month 8 while magnitude remains qualitatively unchanged.
  - Table 2 All controls coefficients for headline inflation by horizon h (selected): h=5: 0.41 (std. err. 0.15); h=6: 0.36 (0.12); h=7: 0.32 (0.10); h=8: 0.30 (0.13).
- Food inflation:
  - One standard deviation increase in GSCPI increases food inflation by 0.54 percentage points five months after the shock; this coefficient is not statistically significant (Table 2, panel B).
  - Reflects high sensitivity of food prices to supply chain pressures given food’s 36 percent consumption share.
  - Table 2 No controls coefficients for food by horizon h (selected): h=5: 0.62 (0.22) **; h=6: 0.66 (0.19) **; h=7: 0.76 (0.18) ***; h=8: 0.87 (0.17) ***.
- Tradable core inflation:
  - One standard deviation GSCPI shock increases tradable inflation by 0.49 percentage points after five months with a 10 percent significance level (Table 2, panel C).
  - Effect is weakly significant, elevated through month 6 and reverts after month 7.
  - Sample coefficients (No controls) show peaks around h=8: 0.58 (0.12) ***.
- Non-tradable core inflation:
  - Smaller and delayed effect: flat response first three months, gradual increase from month 4, peak at month 7 after the shock.
  - Coefficients are smaller and not statistically significant.
  - Table 2 No controls coefficients for non-tradable by horizon h (selected): h=7: 0.26 (0.11) *; h=12: 0.21 (0.20).
- Comparative shocks (global oil and global food price shocks):
  - Global oil price shock: positive, sizeable, and persistent impacts across headline, food, tradable, and non-tradable inflation; impact on headline materializes in first three months and reverts seven months after shock.
  - Global food price shock: effect on headline peaks after eight months, similar magnitude to oil shock; food inflation rises quickly until month five and persists until month 12.
  - Both global oil and food shocks have persistent impacts on measures of inflation.

### Robustness checks and interpretation for policy
- Additional controls tested: monetary policy rates (global and country-specific), VIX economic volatility index, nominal effective exchange rate (NEER).
- Key robustness findings:
  - Including monetary policy rates: the impact of GSCPI on headline inflation disappears, suggesting monetary authorities respond to inflation caused by supply chain disruptions.
  - Including VIX: headline inflation response follows the baseline pattern.
  - Including NEER: headline inflation response closely follows the baseline.
- Policy interpretation:
  - Monetary policy responses appear to offset inflationary effects from global supply chain pressures when accounting for monetary policy rates.
  - Supply chain disruptions transmit to domestic inflation mainly through tradable and food prices with limited immediate pass-through to non-tradable inflation.

*Source: IMF Working Paper chapter "2.2 The inflation outlook in sub-Saharan Africa" (wpiea2023039-print-pdf).*

### 5. Contributions of supply chain pressures to the

### 5. Contributions of supply chain pressures to the surge in inflation

### 5.1 Methodology
- Decomposition: year-over-year median headline inflation rate in the region, yt, is decomposed as the sum of contributions from GSCPI shocks, global food price shocks, global oil price shocks, and a residual term δt:
  - Each contribution term is the sum-product of the estimated local-projection coefficients (β̂_h for each horizon h) and the innovations (ε̂_{t−h}) to the respective shock variable (GSCPI, food, oil).
- Two-step implementation:
  1. Retrieve estimated β̂_h coefficients from local projections:
     - Δ_h y_{t+h} = α_t^h + Σ β_{hj} ΔW_{t−j} + Σ θ_{hj} G X_{t−j} + Σ γ_{hj} Δy_{t−j} + ε_{t+h}, for h = 0,...,H;
     - W_t is either GSCPI, oil prices, or food prices; X_G is a set of global controls (global oil prices, global food prices, world output gap).
     - The estimated β_h reflects the response of y_t to one standard deviation in W_t.
  2. Compute residuals (innovations) ε̂_t for GSCPI, global oil prices, and global food prices by regressing each on control variables (lags of the global variables and lags of the median inflation rate) via OLS.
- Implementation details preserved exactly:
  - Controls include lags of global variables and lags of the median inflation rate.
  - Lag length set to p = 12 to account for seasonal variation in monthly series.
- Scope:
  - The approach is applied to headline, food, tradable, and non-tradable median inflation measures.
  - Appendix Figure 2 presents the residuals used as innovations.

### 5.2 Results
- General pattern (2020-22):
  - Supply chain pressure shocks contributed positively to inflation throughout the post-2020 period.
  - Timeline:
    - First months of 2020: supply chain pressures gradually increased to peak towards the end of 2020.
    - First half of 2021: signs of supply chain pressures easing, though still present.
    - From July 2021 onwards: supply chain pressures gradually increase to reach an all-time high in June 2022 (end of sample).
- Quantitative contributions to average median inflation rate during 2020-22 (sub-Saharan Africa):
  - Supply chain pressures contributed to 45 percent of the average median headline inflation rate.
  - Contributions by inflation component:
    - Food inflation: 40 percent.
    - Tradable inflation: 55 percent.
    - Non-tradable inflation: 40 percent.
- Role of oil and food prices:
  - Oil prices:
    - Helped keep inflation in check from 2020 up to the last quarter of 2021.
    - Contributed considerably to the rise in headline inflation thereafter.
  - Food prices:
    - Contributed to lower inflation from 2020 to early 2021.
    - Positively contributed to headline inflation from early 2021 until the beginning of 2022.
    - Contributed negatively to headline inflation in the first half of 2022.
- Visualization:
  - Figure 5 presents contributions of GSCPI shocks to different measures of inflation over 2020-22:
    - Solid black line: the inflation measure.
    - Red bars: contributions of supply chain shocks to inflation.

*Source: IMF Working Paper — "Global Supply Chain Disruptions: Challenges for Inflation and Monetary Policy in SSA" (section 5).*

### Appendix Figure 1. Contribution of global supply shocks to inflation

### Appendix Figure 1. Contribution of global supply shocks to inflation

### Figure description
- Period covered: 2020-22.
- Visualization elements:
  - Solid black line: the considered inflation measure.
  - Red bars: contributions of supply chain shocks to inflation.
  - Blue bars: contributions of food prices to inflation.
  - Green bars: contributions of oil prices to inflation.

### Residuals (Appendix Figure 2)
- Each panel presents the residuals of:
  - global oil prices,
  - global food prices,
  - GSCPI
  from their respective regressions on a set of control variables.

### Data and country coverage (Annex I)
- Country list (Total = 29): Angola; Benin; Burkina Faso; Cameroon; Chad; Cote d'Ivoire; Equatorial Guinea; Ethiopia; Gabon; Ghana; Guinea-Bissau; Kenya; Lesotho; Madagascar not listed; Mauritius; Mozambique; Namibia; Niger; Nigeria; Republic of Congo; Rwanda; Senegal; Seychelles; South Africa; Tanzania; Togo; Uganda; Zambia; Mali.
- Variables and sources:
  - Headline inflation — Haver
  - Tradables inflation — Haver
  - Non-tradables inflation — Haver
  - Food inflation — Haver
  - Output Gap — IMF WEO (Annual frequency)
  - Real GDP — IMF WEO (Annual frequency)
  - GSCPI — Haver (Expressed in standard deviations)
  - Baltic Dry Index — Bloomberg
  - Harpex — Haver (Starts in 2001)
  - World Output Gap — IMF WEO (Annual frequency)
  - World Oil Price — IMF Primary Commodity Prices (Crude Oil (petroleum), Dated Brent)
  - World Food Price Index — IMF Primary Commodity Prices (Commodity Food Price Index, 2016 = 100, includes Bananas, Cereal, Citrus-fruit, Dairy Products, Legumes, Meat, Non-citrus fruit, Seafood, Sugar, Vegetables, and Vegetable Oils)
  - VIX Uncertainty Index — Bloomberg
  - US Effective Federal Funds rate — Bloomberg
  - Monetary policy rate — Bloomberg (Central Banks' deposit rate)

### CPI component weights (Appendix Figure 3)
- Sources: World Bank, and IMF staff calculations.
- Note: Residual component not listed.

### Structural model (Annex II) — summary of key structure and mappings
- Environment:
  - Small open economy Home (H) facing large Foreign (F).
- Production:
  - Two sectors: tradable (T) and non-tradable (N).
  - Tradable production: q_H^T = z_H^T L_H^T with unit cost c_H^T = χ_T (z_H^T)^−1 w_H.
  - Non-tradable production: q_H^N = z_H^N (L_H^N)^{1−γ_H^N} (N_H^N)^{γ_H^N} with unit cost c_H^N = χ_N (z_H^N)^−1 (w_H)^{1−γ_H^N} (P_H^T)^{γ_H^N}.
- Demand and trade:
  - Armington aggregator for tradable composite Q_H^T with elasticity σ.
  - Iceberg trade costs: τ_S to import from F, τ_X to export.
  - Exports: X_H^T = (τ_X)^{1−σ} (c_H^T P_F^T)^{1−σ} E_F^T.
  - Imports: M_H^T = [(τ_S)^{1−σ}] (c_F^T P_H^T)^{1−σ} E_H^T.
- Competitive equilibrium system (variables to solve for): w, Y_T, Y_N, E, P_T, c_T, P_N, c_N with equations including:
  - wL = Y_T + (1−γ_N) Y_N
  - E = wL
  - (P_T)^{1−σ} = (c_T)^{1−σ} + (τ_S c_F^T)^{1−σ}
  - P_N = c_N
  - c_T = χ_T (z_T)^{−1} w
  - c_N = χ_N (z_N)^{−1} (w)^{1−γ_N} (P_T)^{γ_N}
  - Y_T and Y_N expressions (see full model for details)
- First-order approximation and mapping to local projections (normalize w=1, approximate around σ=1):
  - dln P_T = m dln τ_S
  - dln P_N = γ_N dln P_T
  - Headline inflation mapping:
    - dln P = (ω s + (1−ω) γ_N) dln τ_S

### Dynamic model with price rigidities (Model of section 4) — key equations
- Headline inflation definition:
  - π_t ≡ (1−ω) π_t^N + ω π_t^T
- Four-equation system:
  1. Supply curve linking tradable inflation to global supply chains:
     - π_t^T = λ π_{t−1}^T + ∑_{τ=0}^μ τ ⋅ ΔC_{t−1−τ}^T
  2. Open-economy Phillips curve for non-tradable inflation:
     - π_t^N = α E(π_t^N) + β x_{t−1} + ε ω π_{t−1}^T
  3. IS equation:
     - x_t = ζ x_{t−1} − η (R_t − E(π_t) − r̄)
  4. Global supply chain disruption process:
     - ΔC_t = ν ΔC_{t−1} + ι_t
- Central bank:
  - Observes the shock first, then sets the interest rate at the beginning of the period.
  - Objective function: ∑_t ξ x_t^2 + (1−ξ) (π_t − π̄)^2
- Inflation expectations:
  - E(π_t^N) = γ π̄ + (1−γ) π_{t−1}^N
    - γ = 0 – adaptive expectations, low credibility
    - γ = 0.5 – anchored expectations, high credibility

### Calibration (Appendix Table 3) — parameters and values
- ω — Share of tradable goods in consumption basket: 0.4 or 0.7
- λ — Persistence of tradable inflation: 0.5
- μ — Elasticity of tradable inflation to global supply chain disruptions: Local projections (no point value listed)
- α — Persistence of non-tradable inflation: Gali (2015) 0.99
- β — Slope of core Phillips curve for non-tradable inflation: Stock and Watson (2020) 0.2 (annual)
- ε — Second-round effects of global shipping costs on non-tradable inflation: Local projections 0.5
- ζ — Persistence of output gap: 0.5
- η — Elasticity of output gap to real policy rate: Gali (2015) 1 (annual)
- ν — Persistence of shock to global shipping costs: 0 or 0.5
- ξ — Weight of output gap in the central bank's objective function: 0.5

### Stylized economy scenarios (Appendix Figure 4)
- Four economy types illustrated (central bank is backward-looking; central bank raises real policy rate after nine months):
  1. Economy 1: High share of tradables, low credibility and fully adaptive inflation expectations
  2. Economy 2: High share of tradables, high credibility and anchored inflation expectations
  3. Economy 3: Low share of tradables, low credibility and fully adaptive inflation expectations
  4. Economy 4: Low share of tradables, high credibility and anchored inflation expectations

*Source: Appendix Figure 1 and accompanying appendix material from the IMF working paper "Global Supply Chain Disruptions: Challenges for Inflation and Monetary Policy in SSA."*

---


_Source: https://www.imf.org/-/media/files/publications/wp/2023/english/wpiea2023039-print-pdf.pdf_