## wpiea2022169-print-pdf

## Source details

**Canonical URL:** [wpiea2022169-print-pdf](https://www.imf.org/-/media/files/publications/wp/2022/english/wpiea2022169-print-pdf.pdf)

## Other formats

- [Markdown version](/-/media/files/publications/wp/2022/english/wpiea2022169-print-pdf.pdf.md)
- [Structured JSON version](/-/media/files/publications/wp/2022/english/wpiea2022169-print-pdf.pdf.json)

---

### II. STYLIZED FACTS AND MODEL MOTIVATION
- Model calibrated to reflect key stylized facts of the Ghanaian economy for macroeconomic policy analysis and monetary policy making.
- Ghana: developing small open economy with relatively stable macroeconomic environment over the last few decades.
- Monetary policy conducted under the IT framework, formally in place for more than a decade.
- Data sources: Ghana Statistical Office, Bank of Ghana, International Monetary Fund.

- Economic transformation and growth phases:
  - Ghana transitioned from a low-income economy to a lower-middle-income economy in 2011.
  - Growth drivers: robust services and industry sectors.
  - Three phases (last decade):
    - 2010-13: average real GDP growth of 6.6 percent.
    - 2014-16: average real GDP growth of 2.8 percent (down from an average of 6.6 percent in 2010-13); drivers included monetary policy tightness, oil and gas production challenges, power supply constraints, and drop in export prices; accompanied by high fiscal deficits, double-digit inflation, and increased debt burden.
    - 2017-19: GDP growth averaging 7.0 percent; supported by unwinding imbalances, fiscal consolidation, and tight monetary policy.

- COVID-19 impact and resilience:
  - Rapid growth halted by the COVID-19 pandemic starting early-2020.
  - Economy held up comparatively well despite lockdowns; GDP: 0.4 percent in 2020.
  - Government intervention: record fiscal deficit and rising public debt due to the Coronavirus Alleviation Programme.
  - GDP growth: 5.2 percent in 2021, supported by a good cocoa harvest, mining activity, and services.

### Recent macroeconomic indicators (inflation, exchange rate, interest rates)
- Inflation trajectory and targets:
  - Medium-term target: 8±2 percent.
  - Inflation peaked at 19.2 percent in March 2016.
  - 11.4 percent in July 2020 (panic buying linked to a three-week partial lockdown).
  - 7.5 percent in May 2021.
  - 12.6 percent in December 2021.
  - 15.7 percent in February 2022.
  - Food subcomponent pressures dominate; imported inflation contribution increased since H2 2021 due to depreciation and global food and energy price dynamics.

- Exchange rate and FX movements:
  - Exchange rate regime: flexible with intermittent FX operations to minimize volatility impact on inflation.
  - Cedi cumulative annual depreciation against the USD: 15.7 percent in 2015 to depreciation of 4.1 percent in 2021.
  - Year-to-date depreciation of 15.4 percent against the USD in 2022Q1.
  - 2022Q1 depreciation drivers: sovereign downgrades, parliament’s impasse on the electronic transaction levy, and tighter external financing conditions from US monetary tightening.

- Monetary policy and market rates:
  - Policy easing window: MPR lowered by 950 basis points, to 16 percent at end-2019.
  - Policy rate: 14.5 percent at the end-2021.
  - Weighted average interbank rate: 12.7 percent.
  - Policy tightening: policy rate increased by 250 basis points, to 17 percent in March 2022, with additional macroprudential measures.
  - Negative real returns on key domestic Treasury securities due to surge in inflation from H2 2021 vis-à-vis declining market rates.

### QPM — structure and key mechanisms
- Model family: semi-structural New Keynesian (NK) Quarterly Projection Model (QPM); "gap" model with variables as deviations from trends.
- Four core blocks:
  - Aggregate demand block (IS/expectational investment-savings curve).
  - Phillips curves disaggregated into food and non-food.
  - Exchange rate block (modified UIP with persistence and sovereign risk premium).
  - Monetary policy reaction function (forward-looking Taylor-type rule).
- Main transmission channels: interest rate channel and exchange rate channel; expectations play a major role.
- Ghana-specific features:
  - Disaggregation of CPI into food and non-food with two Phillips curves.
  - Inclusion of fiscal policy effects and explicit fiscal impulse.
  - Monetary policy credibility channel (incredibility term affecting expectations).
  - Imperfect exchange rate flexibility and imperfect interest rate pass-through.
  - Monetary aggregates have no explicit role; interest rate is main instrument.

### Inflation block — specification highlights
- Non-food Phillips curve (quarter-on-quarter annualized non-food inflation 휋tnf):
  - Depends on past and expected future non-food inflation, real marginal cost rmc_t^nf, and imported inflation proxy m_t.
  - Expectations include model-consistent expectations augmented by credibility/incredibility term incredd_t.
  - rmc_t^nf = b15·ŷt + (1−b15)·ẑt.
  - m_t = (Δs_t + π_t* − Δz̅_t).
- Food Phillips curve: analogous with different parameterization and greater weight on RER gap in rmc_t^f.
- Headline CPI: p_t = w·p_t^f + (1−w)·p_t^nf + ε_t^p.
- Expectations formation: model-consistent expectations augmented by aggregate central bank credibility; credibility evolves slowly with inertia and depends on past deviations of annual headline inflation (4πt−1) from target (π̅).

### Aggregate demand and fiscal block
- Output gap equation:
  - ŷt = α1·ŷt−1 + α2·Et·ŷt+1 − α3·rmci_t + α4·ŷt* + α5·fimp_t + εt^y.
  - Real monetary conditions index rmci_t = α6·r̂t + (1−α6)(−ẑt).
  - Foreign (USA) output gap ŷt* enters directly.
- Fiscal impulse:
  - Fiscal deficit decomposed into cyclical and structural components.
  - Net fiscal impulse fimp_t identified with shock to structural deficit εt_defstr.
  - def_t = def_cyc,t + def_str,t; def_cyc,t = −f1·ŷt; def_str,t = f2·def_str,t−1 + (1−f2)·def_str,ss + εt_defstr.

### Exchange rate and UIP block
- Modified UIP:
  - s_t = s_t+1^e + (i_t* − i_bt + prem_t)4/⁄ + εt^s.
  - Expectations partly forward- and partly backward-looking: s_t+1^e = c1·Et·s_t+1 + (1−c1)(s_t−1 + 2/4·Δs̅_t).
  - Trend nominal depreciation Δs̅_t = π̅_t − π_t* + Δz̅_t.
  - c1 captures degree of forward-looking behaviour (or central bank smoothing).

### Monetary policy reaction function
- Taylor-type rule with interest rate smoothing:
  - i_t = g1·i_t−1 + (1−g1)[Et·πt+1 + r̅_t + g2(Et·4πt+3 − π̅_t) + g3·ŷt] + εt^i.
  - Interbank rate: i_bt = i_t + sp_t, with spread sp_t = g4·sp_t−1 + εt^sp.
  - Taylor principle: g2 > 1 required for unique solution and to anchor expectations.

### Long-run trends and calibration approach
- External sector and long-run trends modeled as autoregressive processes converging to steady-states: x_t = kx·x_t−1 + (1−kx)·x_ss + εt^x, for x_t ∈ (π_t*, ŷt*, r_t*, Δz̅_t, Δȳ_t, π̅_t, prem̅_t).
- Model calibrated (not estimated) due to small samples, data quality, simultaneity, and rational expectations; calibration matches theoretically-consistent propagation mechanisms.

### Calibrated behavioral parameters (selected)
- Non-food inflation: b11 = 0.7; b12 = 0.4; b13 = 0.1; b14 = 0.5; b15 = 0.7.
- Food inflation: b21 = 0.5; b22 = 0.2; b23 = 0.1; b24 = 0.25; b25 = 0.6.
- Output gap: a1 = 0.4; a2 = 0.3; a3 = 0.1; a4 = 0.1; a5 = 0.5; a6 = 0.8.
- Interest rate and exchange rate: g1 = 0.75; g2 = 1.3; g3 = (value truncated in source excerpt); c1 = 0.6.
- Credibility: b5 = 0.5; b6 = 1.5.
- Additional calibrated items: 푔 4; 0.65; 푤 0.4312; 푓1 0.2; 푓2 0.95.

### IV. MODEL RESULTS — overview and IRFs
- Model tools: impulse response functions (IRFs), historical decomposition, out-of-sample simulations.
- Simulations: one-time single shock exercises from general equilibrium; responses expressed as deviations from equilibrium.

- Aggregate demand shock (Figure 5):
  - Raises overall inflation short-run, driven by non-food inflation surge.
  - Food prices fall due to negative real exchange rate gap dominating higher output gap effect.
  - Central bank raises policy rate, causing appreciation and contributing to fall in food inflation.
  - Domestic currency strengthens in nominal and real terms; positive RIR gap tightens monetary conditions.
  - Positive output gap gradually closes; inflation returns to target but demand-side pressures persist, requiring persistent policy response.

- Non-food price (cost-push) shock (Figure 6):
  - Pick-up in non-food and headline inflation in first quarter.
  - Central bank raises nominal interest rate; induces nominal appreciation and negative RER gap (overvalued currency).
  - RIR gap initially negative due to higher inflation expectations, then turns positive as inflation declines.
  - Output declines short-term due to real appreciation dampening demand via trade channel.
  - Monetary policy reaction key to bringing inflation back; trade-off: higher inflation with negative output gap.

- Exchange rate (UIP) depreciation shock (Figure 7):
  - Nominal and real depreciation; above-target headline inflation via food and non-food imported components.
  - Monetary authority raises short-term rate; RIR gap negative short-run due to higher inflation expectations.
  - Real monetary conditions looser initially, producing a positive output gap.
  - Gradual rate increases strengthen currency and tighten real conditions, returning output and inflation to steady state.

- Adverse food supply shock (Figure 8):
  - Increases food and aggregate prices; non-food inflation declines slightly.
  - Central bank moderately raises nominal rate; initial nominal response lower-than-proportionate producing negative RIR gap initially.
  - Real interest rate rises subsequently; monetary conditions become restrictive due to steady appreciation.
  - Tighter real conditions depress output, trough by fifth quarter, then gradual recovery.
  - Monetary policy reacts less to food shocks than non-food because food shocks are more temporary.

- Monetary policy shock (Figure 9, one-unit positive shock):
  - Contractionary effects: hump-shaped negative output gap.
  - Transmission through RIR gap (higher nominal and real rates) and RER gap (overvalued due to appreciation).
  - Food and non-food inflation decline by a maximum of 0.4 and 0.5 percent in quarterly annualized terms, respectively. Consistent with calibration of higher responsiveness of non-food to policy.

### Equation decompositions and historical narrative
- Kalman filter and smoother estimate unobserved states to replicate historical data; QPM decomposes observed data into structural contributions.
- Key qualitative finding: supply-side shocks have predominant and consistent contributions to inflation dynamics in Ghana.
- Historical episodes highlighted (selected):
  - Late-2007 to early-2009: inflation rise partly due to imported inflation, global commodity prices, weaker cedi, election uncertainty.
  - Mid-2009 to end-2011: disinflation attributed to decline in real marginal costs and lower inflation expectations.
  - 2012-14: acceleration in inflation via rising marginal costs and imported inflation from large real exchange rate depreciation.
  - Early-2015 to H1-2016: spikes in inflation from adverse cost-push shocks associated with energy supply challenges.
  - H2-2016 to 2019Q1: faster disinflation from declines in real marginal cost and inflation inertia and tighter real monetary conditions.
  - 2020Q2 (COVID-19): abrupt spurt in inflation driven by adverse cost-push shocks linked to containment measures; policy support (policy rate cuts and positive fiscal impulse) moderated output contraction.

### Out-of-sample forecasting accuracy (RMSFE ratios relative to random walk, 2017Q1–2021Q4)
- Policy rate: 1Q 0.06, 2Q 0.10, 3Q 0.14, 4Q 0.17, 5Q 0.21, 6Q 0.25, 7Q 0.29, 8Q 0.26
- CPI (yoy): 1Q 0.10, 2Q 0.15, 3Q 0.18, 4Q 0.19, 5Q 0.25, 6Q 0.40, 7Q 0.46, 8Q 0.46
- Output Gap: 1Q 0.21, 2Q 0.33, 3Q 0.30, 4Q 0.36, 5Q 0.29, 6Q 0.40, 7Q 0.34, 8Q 0.41
- GDP Growth (yoy): 1Q 0.32, 2Q 0.57, 3Q 0.55, 4Q 0.78, 5Q 0.54, 6Q 0.72, 7Q 0.52, 8Q 0.67
- Nominal ER depreciation (qoq ann.): 1Q 0.38, 2Q 1.05, 3Q 0.70, 4Q 0.57, 5Q 0.52, 6Q 0.83, 7Q 1.26, 8Q 1.52
- Conclusion: QPM significantly outperforms a random walk for most variables and horizons except nominal exchange rate depreciation at some horizons.

### Monetary policy credibility channel
- Partial/time-varying credibility ("incredibility") manifestations:
  - Excessive inflation persistence, larger responsiveness of inflation to shocks, and asymmetric inflation bias.
- Implementation: incredibility variable adds to inflation persistency; full credibility corresponds to b6 = 0.
- Calibration in BOG QPM: b6 = 1.5, implying significant inflation persistence from lack of credibility.
- Simulation (mixed cost-push shocks, +1 percentage point to food and non-food annual inflation):
  - Full credibility (b6 = 0): sudden inflation jump subsides quickly; modest output decline.
  - Partial credibility (b6 = 1.5): persistently higher inflation via stickier expectations; monetary policy tightens longer; stronger exchange rate appreciation, bigger drop in output gap and real marginal costs; higher output costs—worse inflation-output trade-off.
- Credibility block increases persistence of IRFs for other shocks (e.g., risk premia) and can generate higher inflation with lower output gap relative to full credibility.

### Simulating disinflation scenarios (reducing inflation target by 1 percentage point)
- Purpose: measure output-equivalent cost of disinflation (sacrifice ratio).
- Baseline simulation (agents fully believe the change, starting from full credibility):
  - Sacrifice ratio: 0.3 percent of annual output.
  - Instantaneous exchange rate appreciation of more than 1 percent in the first quarter relative to no-disinflation case.
  - Disinflation quick and entails modest costs in standard setup.

- Counterfactual: exchange rate held unchanged for one year:
  - If exchange rate does not adjust instantly, inflation remains relatively flat and does not follow new target immediately.
  - Central bank must tighten nominal and real interest rates; policy rate increases.
  - Larger output gap reduction required; sacrifice ratio roughly twice as large, about 0.5 percent of annual GDP.

- Alternative: reduced monetary policy credibility (expectations unchanged for one year after announcement):
  - Inflation and policy rate dynamics similar to fixed-exchange-rate-for-one-year case, but output gap dynamics more severe.
  - Central bank sharply appreciates exchange rate on impact; sacrifice ratio equal to 1.3 percent of annual GDP.
  - Empirical benchmark: Erceg and Levin (2003) estimate sacrifice ratio up to 1.7 percent for the Volcker disinflation.

- Policy implications:
  - Good central bank communication and transparency crucial for anchoring expectations.
  - Pre-emptive measures to guard against de-anchoring inflation expectations necessary to avoid large re-anchoring costs.

### Model applications, takeaways, and recommendations
- QPM features:
  - Canonical four-equation semi-structural gap model with Ghana-specific features: heterogenous inflation dynamics, fiscal policy role, excess price persistency via credibility channel.
  - Calibrated model displays theoretically consistent transmission channels; robust out-of-sample fit.

- Uses:
  - Disentangling driving forces of historical business cycle dynamics, including COVID-19 shocks.
  - Assessing importance of monetary policy credibility and anchoring of inflation expectations for stabilization and minimizing cumulative interest rate impulses.
  - Exploring transition dynamics and costs when revealing a new inflation target, including shock-dependent outcomes for policy conduct, interest rate dynamics, and sacrifice ratios.

- Recommended priorities for policymakers:
  - Maintain and strengthen central bank communication and transparency to anchor expectations.
  - Monitor developments carefully during a transition to a different inflation target given shock-dependent risks.
  - Recognize potential for large output costs if expectations are sticky or credibility is weakened when designing disinflation strategies.

*Source: wpiea2022169-print-pdf - Quarterly Projection Model for the Bank of Ghana — Working Paper No. WP/2022/169*

### Section V concludes.

### Section V concludes.

### II. STYLIZED FACTS AND MODEL MOTIVATION
- Model is calibrated to reflect key stylized facts of the Ghanaian economy to ensure relevance for macroeconomic policy analysis and monetary policy making.
- Ghana is a developing small open economy with relatively stable macroeconomic environment over the last few decades.
- Monetary policy formulation is conducted under the IT framework, formally in place for more than a decade.
- Data sources used: Ghana Statistical Office, Bank of Ghana, International Monetary Fund.

### Economic transformation and growth phases
- Ghana transitioned from a low-income economy to a lower-middle-income economy in 2011.
- Growth and stability were underpinned by robust services and industry sectors.
- Three phases of economic performance over the last decade:
  - 2010-13: average real GDP growth of 6.6 percent.
  - 2014-16: sharp economic slowdown with average real GDP growth of 2.8 percent (down from an average of 6.6 percent in 2010-13); drivers included monetary policy tightness, oil and gas production challenges, lingering consequences of the power supply constraints, and a drop in export prices of key commodities. The downturn was accompanied by high fiscal deficits, double-digit inflation, and an increased debt burden.
  - 2017-19: significant improvement with GDP growth averaging 7.0 percent; improvement supported by unwinding large economic imbalances, fiscal consolidation (tightening expenditure controls and plugging revenue leakages), and complementary tight monetary policy to lower inflation and ensure exchange rate stability. Growth was broad-based across industry, services, and agriculture.

### COVID-19 impact and resilience
- The rapid growth was halted by the COVID-19 pandemic starting early-2020.
- The country incurred adverse economic impacts due to the global health crisis.
- Despite worldwide lockdowns and closures of businesses and borders, the Ghanaian economy held up comparatively well (text ends before specific COVID-period GDP growth figure).

*Source: wpiea2022169-print-pdf - Section V concludes.*

### 0.4 percent in 2020. The slowdown reflected weak consumption and net exports dynamics, with

### wpiea2022169-print-pdf - 0.4 percent in 2020. The slowdown reflected weak consumption and net exports dynamics, with

### Recent economic performance (2020–2022)
- GDP: "0.4 percent in 2020."
- Government intervention: record fiscal deficit and rising public debt due to support through the Coronavirus Alleviation Programme.
- GDP growth: "5.2 percent in 2021", supported by a good cocoa harvest, mining activity, and services.
- Inflation trajectory highlights:
  - Medium-term target: "8±2 percent".
  - Inflation peaked at "19.2 percent in March 2016".
  - "11.4 percent in July 2020" (panic buying linked to a three-week partial lockdown).
  - "7.5 percent in May 2021".
  - "12.6 percent in December 2021".
  - "15.7 percent in February 2022".
- Drivers: food subcomponent pressures dominate; imported inflation contribution increased since the second half of 2021 due to continuous depreciation of domestic currency and global food and energy price dynamics.

### Exchange rate, interest rates, and monetary conditions
- Exchange rate regime: flexible; intermittent FX operations to minimize volatility impact on inflation.
- Cedi movements:
  - Cumulative annual depreciation against the USD: "15.7 percent in 2015" to a depreciation of "4.1 percent in 2021".
  - Year-to-date depreciation of "15.4 percent against the USD" in "2022Q1".
  - 2022Q1 depreciation reflected sovereign downgrades, parliament’s impasse on the electronic transaction levy, and tighter external financing conditions from US monetary tightening.
- Monetary policy and market rates:
  - Policy easing window: MPR lowered by "950 basis points, to 16 percent at end-2019".
  - Policy rate: "14.5 percent at the end-2021".
  - Weighted average interbank rate: "12.7 percent".
  - Policy tightening: policy rate increased by "250 basis points, to 17 percent in March 2022", with additional macroprudential measures to tighten liquidity.
- Real returns and monetary stance: surge in inflation from H2 2021 vis-à-vis declining market rates produced negative real returns on key domestic Treasury securities and loosened real monetary conditions.

### QPM (Quarterly Projection Model) — structure and key mechanisms
- Model family: semi-structural New Keynesian (NK) Quarterly Projection Model (QPM); "gap" model with variables expressed as deviations from trends.
- Four core blocks:
  - Aggregate demand block (IS/expectational investment-savings curve).
  - Phillips curves (disaggregated into food and non-food).
  - Exchange rate block (modified UIP with persistence and sovereign risk premium).
  - Monetary policy reaction function (forward-looking Taylor-type rule).
- Two main transmission channels approximated: interest rate channel and exchange rate channel; expectations play a major role.
- Key model features tailored to Ghana:
  - Disaggregation of CPI into food and non-food; two separate Phillips curves.
  - Inclusion of fiscal policy effects and explicit fiscal impulse.
  - Monetary policy credibility channel (incredibility term affecting expectations).
  - Imperfect exchange rate flexibility and imperfect interest rate pass-through.
  - Monetary aggregates have no explicit role; interest rate is main instrument.

### Inflation block — specification highlights
- Non-food Phillips curve (quarter-on-quarter annualized non-food inflation 휋tnf):
  - Depends on past and expected future non-food inflation, real marginal cost (rmc_t^nf), and imported inflation proxy (m_t).
  - Expectations include model-consistent expectations augmented by an credibility/incredibility term (incredd_t).
  - Real marginal cost rmc_t^nf = b15·ŷt + (1−b15)·ẑt.
  - Imported inflation proxy m_t = (Δs_t + π_t* − Δz̅_t).
- Food Phillips curve (quarter-on-quarter annualized food inflation 휋tf):
  - Analogous form with different parameterization and greater relative weight on the RER gap in rmc_t^f.
- Headline CPI: p_t = w·p_t^f + (1−w)·p_t^nf + ε_t^p.
- Expectations formation: model-consistent expectations augmented by an aggregate central bank credibility measure; credibility evolves slowly with inertia and depends on past deviations of annual headline inflation (4πt−1) from target (π̅).

### Aggregate demand and fiscal block
- Aggregate demand (output gap ŷt) specification:
  - ŷt = α1·ŷt−1 + α2·Et·ŷt+1 − α3·rmci_t + α4·ŷt* + α5·fimp_t + εt^y.
  - Real monetary conditions index rmci_t = α6·r̂t + (1−α6)(−ẑt).
  - Foreign (USA) output gap enters directly (ŷt*).
- Fiscal impulse:
  - Fiscal deficit decomposed into cyclical and structural components.
  - Net fiscal impulse (fimp_t) identified with shock to structural deficit εt_defstr.
  - Equations:
    - def_t = def_cyc,t + def_str,t.
    - def_cyc,t = −f1·ŷt.
    - def_str,t = f2·def_str,t−1 + (1−f2)·def_str,ss + εt_defstr.

### Exchange rate and UIP block
- Modified UIP:
  - s_t = s_t+1^e + (i_t* − i_bt + prem_t)4/⁄ + εt^s.
  - Expectations partly forward-looking and partly backward-looking: s_t+1^e = c1·Et·s_t+1 + (1−c1)(s_t−1 + 2/4·Δs̅_t).
  - Trend nominal depreciation Δs̅_t = π̅_t − π_t* + Δz̅_t.
  - c1 captures degree of forward-looking behaviour (or central bank smoothing).

### Monetary policy reaction function
- Taylor-type rule with interest rate smoothing:
  - i_t = g1·i_t−1 + (1−g1)[Et·πt+1 + r̅_t + g2(Et·4πt+3 − π̅_t) + g3·ŷt] + εt^i.
  - Interbank rate: i_bt = i_t + sp_t, with spread sp_t = g4·sp_t−1 + εt^sp.
  - Taylor principle: g2 > 1 required for unique solution and to anchor expectations.

### Long-run trends and calibration approach
- External sector and long-run trends are exogenous and modelled as autoregressive processes converging to steady-states: x_t = kx·x_t−1 + (1−kx)·x_ss + εt^x, where x_t ∈ (π_t*, ŷt*, r_t*, Δz̅_t, Δȳ_t, π̅_t, prem̅_t).
- Calibration rationale: model is calibrated (not estimated) due to small samples, data quality, simultaneity, and rational expectations. Calibration matches theoretically-consistent propagation mechanisms for policy analysis and forecasting.

### Calibrated behavioral parameters (selected)
- Non-food inflation: b11 = 0.7; b12 = 0.4; b13 = 0.1; b14 = 0.5; b15 = 0.7.
- Food inflation: b21 = 0.5; b22 = 0.2; b23 = 0.1; b24 = 0.25; b25 = 0.6.
- Output gap: a1 = 0.4; a2 = 0.3; a3 = 0.1; a4 = 0.1; a5 = 0.5; a6 = 0.8.
- Interest rate and exchange rate: g1 = 0.75; g2 = 1.3; g3 = (value truncated in source excerpt); c1 = 0.6.
- Credibility: b5 = 0.5; b6 = 1.5.
- Note: table in source provides full set of calibrated parameters for behavioral equations.

*Source: wpiea2022169-print-pdf - 0.4 percent in 2020. The slowdown reflected weak consumption and net exports dynamics, with*

### 0.1 Food CPI weight Fiscal sector

### 0.1 Food CPI weight Fiscal sector

### Key model parameters and figures
- 푔 4
- 0.65
- 푤 0.4312
- 푓1 0.2
- 푓2 0.95

### IV. Model results — overview
- The model uses impulse response functions (IRFs), historical decomposition, and out-of-sample simulations to:
  - Ascertain how key macroeconomic variables respond to structural shocks.
  - Decompose actual data into contributions of structural factors and shocks.
  - Appraise forecasting accuracy of the QPM.
  - Assess importance of monetary policy credibility and anchoring of inflation expectations.
- Simulations are one-time single shock exercises starting from the model’s general equilibrium; they express responses as deviations from equilibrium (e.g., inflation deviation from the target).

### A. Impulse Response Functions — main findings
- Aggregate demand shock (Figure 5):
  - Raises overall inflation in the short-run, driven by a surge in non-food inflation.
  - Food prices fall because the negative real exchange rate gap dominates the effect from higher domestic output gap.
  - Central bank raises policy rate, causing appreciation and contributing to the fall in food inflation.
  - Domestic currency strengthens in both nominal and real terms; positive RIR gap tightens monetary conditions.
  - Positive output gap gradually closes and inflation returns to target; demand-side inflation pressures persist due to transmission lags, requiring a more persistent policy response.

- Non-food price shock / cost-push shock (Figure 6):
  - Produces a pick-up in non-food prices and overall inflation in the first quarter.
  - Central bank raises nominal interest rate in response to higher headline inflation expectations, inducing nominal appreciation and a negative real exchange rate gap (overvalued currency).
  - RIR gap initially negative due to higher inflation expectations, then turns positive as inflation and expectations decline.
  - Output declines in the short-term, primarily due to real appreciation dampening aggregate demand via the trade channel.
  - Monetary policy reaction is the key force bringing non-food and headline inflation back to target.
  - Such cost-push shocks create a policy trade-off: higher inflation accompanied by negative output gap; in the BOG QPM the central bank tightens because the inflation objective is assumed relatively more important than output gap stability.

- Exchange rate (UIP) depreciation shock (Figure 7):
  - Leads to nominal and real exchange rate depreciation and above-target headline inflation.
  - Headline inflation rises via both food and non-food price increases, primarily through the imported component of real marginal costs due to positive (undervalued) RER gap.
  - Monetary authority raises the short-term interest rate; RIR gap is negative in the short-run because of initially higher inflation expectations.
  - Real monetary conditions become looser initially, producing a positive output gap.
  - Gradual increases in nominal and real interest rates strengthen the domestic currency, tighten real monetary conditions, and return output and inflation toward steady state.

- Adverse food supply shock (Figure 8):
  - Increases food and aggregate prices; non-food inflation declines slightly.
  - Central bank moderately raises nominal interest rate to dampen second-round effects.
  - Initial nominal response is lower-than-proportionate, producing a negative RIR gap (loose policy stance) in initial quarters; real interest rate rises subsequently as price pressures taper off.
  - Monetary conditions become restrictive due to steady appreciation in both nominal and real terms.
  - Tighter real monetary conditions depress output, with a trough by the fifth quarter after the shock; output recovers thereafter as real conditions ease.
  - Monetary policy reacts relatively less to food price shocks than to non-food inflation shocks because food shocks are more temporary and model disaggregation captures this stylized fact.

- Monetary policy shock (Figure 9, one-unit positive shock):
  - Contractionary effects: hump-shaped negative trajectory of the output gap.
  - Transmission through both components of real monetary conditions: RIR gap tightened via higher nominal and real rates; RER gap becomes overvalued due to nominal appreciation.
  - Food and non-food inflation decline by a maximum of 0.4 and

*Source: IMF Working Papers — Quarterly Projection Model for Bank of Ghana*

### 0.5 percent in quarterly annualized terms, respectively. Consistent with the calibration of higher

### Quarterly Projection Model for Bank of Ghana

### Shock responses: food, non-food, and monetary policy
- Non-food prices register their largest decline three quarters after the shock, which is one quarter later than the food prices.
- The response of headline inflation represents a weighted average of food and non-food reactions.
- The monetary policy shock has a greater impact on non-food inflation than on food inflation, largely due to the assumed higher relative weight of output gap vis-à-vis RER gap in the real marginal cost for non-food items as opposed to food items.
- Reference impulse responses shown: Figure 8 (Food inflation shock) and Figure 9 (Monetary policy shock).

### Equation decompositions and historical narrative
- The Kalman filter (and smoother) estimates unobserved states (gaps, trends, shocks) that replicate historical data; the QPM decomposes observed data into determinant contributions.
- Key qualitative finding: supply-side shocks have predominant and consistent contributions to inflation dynamics in Ghana.
- Evidence and episodes:
  - Late-2007 to early-2009: rise in inflation partly due to imported inflation, soaring global commodity prices, weaker cedi, and election-related uncertainties; raised inflation inertia and expectations.
  - Mid-2009 to end-2011: disinflation attributed to decline in real marginal costs (subdued domestic aggregate demand and moderate real exchange rate appreciation) and lower inflation expectations; favorable cost-push shocks supported disinflation.
  - 2012-14: acceleration in inflation driven by rising marginal costs and imported inflation via large real exchange rate depreciation; loose monetary conditions further reinforced inflation acceleration.
  - Early-2015 to H1-2016: spikes in inflation triggered by adverse cost-push shocks associated with energy supply challenges; these also negatively impacted domestic aggregate demand.
  - H2-2016 to 2019Q1: faster disinflation due to declines in real marginal cost and inflation inertia, aided by subdued aggregate demand pressures and tighter real monetary conditions.
  - 2020Q2 (COVID-19): abrupt spurt in inflation predominantly driven by adverse cost-push shocks linked to containment measures, inducing a sharp contraction in the output gap; policy support (policy rate cuts and positive fiscal impulse) moderated the contraction.
- Figures referenced for decompositions: Figure 10 (Inflation and output gap), Figure 11 (Non-food inflation Phillips curve decomposition), Figure 12 (Real marginal costs), Figure 13 (Real monetary conditions), Figure 14 (Output gap decomposition).

### Out-of-sample forecasting accuracy
- Practice: current and next quarter values are tuned using nowcasting and near-term forecasting methods; medium-term forecasts reflect QPM-based simulations conditional on exogenized external trajectories and near-term outlook.
- Recent performance:
  - QPM tended to underestimate inflation (and hence interest rate trajectories), except late-2018 to mid-2019 when it showed an upward bias.
  - Exchange rate forecasts broadly in line with trend depreciation but occasionally biased upward for medium-term horizons.
  - Output gap forecasts tended to be biased downward in the pre-2020 period; model missed the significant decline in output gap during the COVID-19 pandemic.
  - Recent forecasts point to fast narrowing of the negative output gap consistent with post-pandemic recovery.
  - Policy rate forecasts broadly biased downwards over the sample, in line with inflation underprediction, except pre-2018 when the reverse occurred.
- Formal RMSFE comparison (QPM RMSFE relative to random walk, period 2017Q1 to 2021Q4). A ratio < 1 implies QPM outperforms random walk.
  - Policy rate: 1Q 0.06, 2Q 0.10, 3Q 0.14, 4Q 0.17, 5Q 0.21, 6Q 0.25, 7Q 0.29, 8Q 0.26
  - CPI (yoy): 1Q 0.10, 2Q 0.15, 3Q 0.18, 4Q 0.19, 5Q 0.25, 6Q 0.40, 7Q 0.46, 8Q 0.46
  - Output Gap: 1Q 0.21, 2Q 0.33, 3Q 0.30, 4Q 0.36, 5Q 0.29, 6Q 0.40, 7Q 0.34, 8Q 0.41
  - GDP Growth (yoy): 1Q 0.32, 2Q 0.57, 3Q 0.55, 4Q 0.78, 5Q 0.54, 6Q 0.72, 7Q 0.52, 8Q 0.67
  - Nominal ER depreciation (qoq ann.): 1Q 0.38, 2Q 1.05, 3Q 0.70, 4Q 0.57, 5Q 0.52, 6Q 0.83, 7Q 1.26, 8Q 1.52
- Conclusion: QPM significantly outperforms a random walk for all variables and horizons except for nominal exchange rate depreciation at three horizons.

### Monetary policy credibility channel in the BOG QPM
- Rationale: develop a tractable way to incorporate partial and time-varying monetary policy credibility in a linear QPM to reflect developing-economy realities.
- Manifestations of partial credibility ("incredibility"):
  - (i) Excessive inflation persistence making short-run inflation-output trade-offs harder.
  - (ii) Larger responsiveness of inflation to shocks (e.g., stronger exchange rate pass-through).
  - (iii) Asymmetric inflation bias from low credibility.
- Implementation: the incredibility variable adds another layer to inflation persistency; full credibility corresponds to setting b6 = 0.
- Calibration used in BOG QPM: b6 = 1.5, implying significant inflation persistence due to lack of central bank credibility affecting expectations anchoring.
- Simulation comparing full credibility (b6 = 0) vs BOG partial credibility (b6 = 1.5):
  - Shock: mix of cost-push shocks increasing both food and non-food annual inflation by 1 percentage point each (headline first-round effect also 1 percentage point).
  - Full credibility: sudden jump in inflation that quickly subsides; monetary policy responds initially but only temporarily; output decline relatively modest.
  - Partial/time-varying credibility: inflation shocks generate persistently higher inflation via stickier expectations; monetary policy tightens similarly short-term but remains tight longer; leads to stronger exchange rate, bigger drop in output gap and real marginal costs; inflation persists longer and takes more time to return to target; output costs significantly higher under partial credibility—worse inflation-output trade-off.
- Additional implication: credibility block makes IRFs more persistent for other shocks (e.g., risk premium shocks) and can generate higher inflation with lower output gap relative to full credibility when risk premium reflects lower investment or capital outflows.

### Simulating disinflation scenarios
- Purpose: assess model realism by measuring output-equivalent cost of disinflation.
- Practical context: current inflation target of 8 percent is high relative to other IT central banks; lowering the target is of interest.
- Simulation: 1 percentage point reduction in the inflation target.
  - Baseline assumption: reduction leads to lower inflation expectations.
  - Credibility channel slows dynamics, but inflation expectations decline because the simulation starts from full credibility and agents fully believe the policy change.
  - Real interest rate tightens marginally as monetary policy rate starts declining toward the new steady state.
  - Exchange rate: instantaneous appreciation of more than 1 percent in the first quarter (relative to no-disinflation case).
  - Theoretical note: under rational expectations and full credibility, a permanent reduction in nominal interest rates can induce exchange rate appreciation (citing Schmitt-Grohe and Uribe (2018) and Carvalho et al. (2021)).
- Quantitative outcome: disinflation is quick and entails modest costs in the standard setup, with a sacrifice ratio (cumulative output gaps) of about 0.2-

*IMF Working Paper: Quarterly Projection Model for Bank of Ghana*

### 0.3 percent of  annual  output.  Simulating  a  disinflation  scenario  where  a  target  reduction  is

### Quarterly Projection Model for the Bank of Ghana

### Simulations of reducing inflation target by 1 percentage point
- Scenario: reduction in inflation target by 1 percentage point announced; baseline disinflation simulation compared with alternative setups (Figures 18 and 19).
- Baseline outcome: initial illustrative case implies a sacrifice ratio of 0.3 percent of annual output.
- Announcement one year in advance (not shown) leads to an even lower sacrifice ratio due to expectations channel (anticipating disinflation reduces inflation via expectations).
- Counterfactual with exchange rate held unchanged for one year (Figure 18):
  - Instantaneous exchange rate appreciation is a key component making disinflation quick and less costly.
  - If exchange rate does not adjust instantly, inflation rate remains relatively flat and does not follow the new target immediately.
  - Central bank is forced to tighten the interest rate in nominal terms as well as real terms; policy rate shows an outright increase.
  - More tightening leads to a lower output gap to bring inflation down.
  - Exchange rate appreciation after one year accelerates inflation decline once agents see central bank commitment.
  - Sacrifice ratio in this case is roughly twice as large, of about 0.5 percent of annual GDP.

### Role of monetary policy credibility and inflation expectations
- Alternative scenario with reduced monetary policy credibility: model shocks the credibility process so inflation expectations remain unchanged for one year even after the new target is known (Figure 19).
- Inflation and policy rate dynamics in this scenario are similar to the fixed-exchange-rate-for-one-year case.
- Major difference: output gap dynamics are more severe.
  - Central bank, lacking expectations support, commits to tight policy and sharply appreciates the exchange rate on impact.
  - This generates a significant sacrifice ratio equal to 1.3 percent of annual GDP.
- Empirical context: Erceg and Levin (2003) provide an empirical estimate of as high as 1.7 percent sacrifice ratio for the Volcker disinflation.
- Policy implications:
  - Good central bank communication and transparency are crucial for anchoring inflation expectations.
  - Pre-emptively guarding against de-anchoring inflation expectations is necessary to avoid large costs of re-anchoring them later.

### Model structure, calibration, and applications
- The QPM builds on the canonical four equation semi-structural gap model and includes Ghana-specific features:
  - Heterogeneous dynamics across sub-indices of headline inflation.
  - Importance of fiscal policy for aggregate economic developments.
  - Excess persistency of prices due to past inflation deviations from the target affecting monetary policy credibility.
- The calibrated model displays theoretically consistent transmission channels to structural shocks.
- Out-of-sample simulations confirm the QPM’s data fit is robust.
- The model is useful for:
  - Disentangling driving forces of historical business cycle dynamics, including COVID-19-related shocks.
  - Assessing importance of monetary policy credibility and anchoring of inflation expectations for stabilization and minimizing cumulative interest rate impulses.
  - Exploring transition dynamics and costs when revealing a new inflation target in advance, including shock-dependent outcomes for monetary policy conduct, interest rate dynamics, and sacrifice ratios.

### Key takeaways and recommendations
- Disinflation costs (sacrifice ratios) are highly sensitive to:
  - Exchange rate adjustment dynamics.
  - Degree of credibility and stickiness of inflation expectations.
  - Timing and communication of target changes.
- Recommended priorities for policymakers:
  - Maintain and strengthen central bank communication and transparency to anchor expectations.
  - Monitor economic developments carefully during a transition to a different inflation target due to shock-dependent risks.
  - Consider the potential for large output costs if expectations are sticky or credibility is weakened when designing disinflation strategies.

*Quarterly Projection Model for the Bank of Ghana — Working Paper No. WP/2022/169*

---


_Source: https://www.imf.org/-/media/files/publications/wp/2022/english/wpiea2022169-print-pdf.pdf_