## Algeria Macroeconomic Projection Model (AMPM) — Introduction and Model Summary ## Source details **Canonical URL:** [Algeria Macroeconomic Projection Model (AMPM) — Introduction and Model Summary](https://www.imf.org/-/media/files/publications/wp/2026/english/wpiea2026025-source-pdf.pdf) ## Other formats - [Markdown version](/-/media/files/publications/wp/2026/english/wpiea2026025-source-pdf.pdf.md) - [Structured JSON version](/-/media/files/publications/wp/2026/english/wpiea2026025-source-pdf.pdf.json) --- ### Model purpose, scope, and Algerian-specific features - Documents structure and properties of the Banque d’Algérie (BA)’s Quarterly Projection Model (QPM), dubbed AMPM, developed within BA’s Forecasting and Policy Analysis System (FPAS) with IMF technical assistance. - Algeria-specific features captured: - importance of hydrocarbon exports (oil and gas); - fiscal policy impact and monetary–fiscal interactions; - monetary aggregate targeting framework prioritizing price stability (implicit core inflation target); - managed floating exchange rate arrangement; - quasi-closed capital account. - BA legal and operational context: - 2023 Monetary and Banking Law (MBL) specifies BA’s objective with price stability as the central monetary policy goal (Article 35). - De jure framework: monetary aggregate targeting using base money as operating target; operating instruments include required reserves, open market operations, bilateral absorptions, BA policy rate, deposit and marginal lending facility rates, and exchange rate management. - Historic practice: M2 growth targets often indicative with occasional large deviations; multiple nominal anchors pursued including M2 growth, nominal exchange rate to achieve REER equilibrium, and an implicit inflation target currently defined as 4 percent core inflation. - Structural impediments weakening transmission: - SOBs account for over 85 percent of total assets; subsidized loan rates (taux bonifiés); uneven liquidity distribution; excess liquidity; lack of meaningful competition; complex nexus among SOEs, SOBs, and government. - Example of policy response: - Programme Spécial de Refinancement (PSR): one-year program totaling about 8.3 percent of GDP, launched in 2021 to facilitate public banks’ refinancing with BA using government debt securities in exchange for long-term syndicated loans to SOEs purchased by the Treasury. - FX regime and BA role: - De jure: managed floating; de facto: crawl-like per AREAER. - BA is dominant FX supplier due to mandatory conversion of commodity exporters’ FX inflows; sets a nominal exchange rate reference value with very narrow bid-ask margins; targets gradual REER alignment using an empirical framework including oil prices and the budget balance. ### Sectoral decomposition, prices, and policy rules embedded in AMPM - GDP decomposition: - split into extractive (hydrocarbon, EGDP, denoted e) and non-extractive (NEGDP, denoted ne) sectors; real variables split into gap and trend components. - Prices and inflation: - CPI decomposed into core and non-core (fresh agricultural goods); core used to define medium-term price stability objective with a 4 percent annual core inflation goal. - Non-core items account for 17 percent of the expenditure basket. - Base year for CPI basket is 2001; weights are significantly outdated. - Central bank behavior modeled via two reaction functions: - one for M2 growth (monetary aggregate target); - one for setting the nominal exchange rate reference value. - Money supply determinants include: - expected deviation of core inflation from the implicit target; - cyclical positions of output and real exchange rate; - announced or implicit M2 growth targets; - impact of not fully sterilized official reserves flows. - Fiscal block: - public domestic debt accumulation decomposed into marketable and non-marketable components; - fiscal reaction function for the structural deficit to ensure debt sustainability and provide countercyclical support. ### Model calibration, validation, and operational use - Calibration rationale: - Calibration used instead of estimation given short macro time-series, statistical breaks, and policy regime changes. - Calibration relies on empirical evidence and theoretical benchmarks; Annex C presents calibrated parameter values. - Validation and operationalization: - Impulse responses (IRFs), historical decompositions, and in-sample simulations used for performance checks. - AMPM is intended as central element of BA’s FPAS for specialized sectoral analysis, nowcasts, near-term forecasts, and scenario analysis (e.g., global energy price trajectories). - Policy simulation capability: - Can evaluate transitions (e.g., moving from quantity-based to interest rate-based framework) and simulate multiple-instrument frameworks. ### Numerical and technical conventions (selected equations and specifications) - Real GDP aggregation: - ∆yt = wy_e ∆yt^e + (1−wy_e) ∆yt^ne + ε^y_t - ŷ_t = wy_e ŷ_t^e + (1−wy_e) ŷ_t^ne - NEGDP IS curve determinants (stylized): - ŷ_t^ne = b1_y^ne ŷ_{t−1}^ne − b2_y^ne mci_t + b3_y^ne tb̂_t^rat + b4_y^ne fisc_t + ε^y_ne_t - Monetary conditions index: - mci_t = c1_mci r̂_t + (1−c1_mci)(−ẑ_t) - Core Phillips curve (quarter-on-quarter annualized core inflation π_t^c): - π_t^c = c1_π^c E_t π_{t+1}^c + (1−c1_π^c) π_{t−1}^c + c2_π^c rmc_t^c + ε^π_c_t - Balance of payments identity (ratios to nominal GDP): - ca_t^rat + kfa_t^rat = 0 - Reserve law of motion (quarter scaling shown by 400 factors): - res_t^rat = resfin_t^rat/4 + res_{t−1}^rat (1 + ∆s_t/400) / (1 + ∆ny_t^ne/400) + ε^{res}_t - M2 reaction function core structure (Δm_t includes expectations of 3-quarter-ahead annual core inflation deviation, output gap ŷ_t, interbank rate deviation, reserve changes, persistence Δm_{t−1}, and shock ε_t^Δm). - Exchange rate policy rule (nominal reference and UIP mix): - s_t = c1_s {s_{t−1} + Δs_t^pol /4} + (1−c1_s){s^F_{t+1} − (r^s_t − r^s_{rw} − prem_t)/4} + (1−c1_s){−c3_s t̂b_t^rat + c4_s (debt_t^rat − debt^rat,tar)} + ε_t^s --- ### Inflation: core vs non-core — definitions, stylized facts, and policy implications - Definitions and weights: - Non-core items account for 17 percent of the expenditure basket and include fresh food. - Core prices are used for BA’s medium-term objective: 4 percent annual core inflation goal. - Base year for CPI basket is 2001; weights are outdated. - Historical dynamics: - Annual core inflation averaged 3.5 percent during 2003-15; fluctuated between 0.5 and 6.5 percent in that period. - Sample average core inflation across the full sample is 4.2 percent. - Sample average non-core inflation across the full sample is 7.1 percent. - Recent drivers of higher inflationary pressures: - dinar depreciation raising import prices; - excess liquidity from increased current public spending; - inefficiencies in distribution chains. - Policy implications: - Core inflation is the BA’s operational target (4 percent). - Non-core volatility (fresh food) driven by weather and government price controls, largely outside BA control — justifies CPI decomposition and suggests integrating exogenous outlooks for fresh food in forecasts. ### Key empirical stylized figures (inflation) - Core inflation target: 4 percent (BA definition). - Non-core share of CPI basket: 17 percent. - Sample averages: - core inflation: 4.2 percent; - non-core inflation: 7.1 percent. - Historical averages and episodes summarized (2003–15 average core 3.5 percent; core range 0.5 to 6.5 percent in that period). --- ### Internal balance and key transmission mechanisms (3.1) - Real economy split and gap aggregation parameters (selected): - wy_e assumed constant; measurement error ε^y_t accounts for time variation in sectoral shares. - Extractive sector dynamics: - Energy price composite gap ẑen_t = w_oil ẑoil_t + (1−w_oil) ẑgas_t. - Core and non-core supply relations: - p_t = w_p^c p_t^c + (1−w_p^c) p_t^nc + disc^p_t with w_p^c denoting core weight. - Relative core price trend ∆rp̅_t follows autoregressive process with steady-state ∆rp_ss. - Real marginal cost decomposition for core: - rmc_t^c = c1_rmc^c ŷ_t^ne + c2_rmc^c ẑen_t + (1−c1_rmc^c − c2_rmc^c) ẑ_t - Trade balance behavioral determinants (equation (23)): real energy price gap, foreign demand, NEGDP output gap (import demand), extractive output gap, competitiveness ẑ_t, and structural primary deficit deviation. --- ### Economic Policies: monetary, exchange rate, and fiscal representation (3.3) - Monetary policy representation: - M2 is modeled directly as central bank instrument; forward-looking reaction function for Δm_t incorporates: - three-quarter ahead expected deviation of annual core inflation from target; - output gap ŷ_t; - interbank rate deviation from neutral r^s_t − r^s_t^neutral; - announced/implicit M2 growth target Δ^4 m_t^tar; - changes in official reserves (res_t^rat − res_{t−1}^rat) if not fully sterilized; - persistence Δm_{t−1} and shock ε_t^Δm. - Real money demand gap r̂^m_t depends on real interest rate gap r̂_t, total output gap ŷ_t, government financing needs (debt_t^rat − debt^rat,tar), reserve changes, and shock ε_t^{r^m}. - Exchange rate policy: - AREAER: de jure “managed floating”; de facto often classified as crawl-like. - Exchange rate s_t determined by BA reference depreciation Δs_t^pol (weight c1_s) and modified UIP (1−c1_s), with additional pressure terms from trade balance and debt deviation. - Expectations and long-term target linked via Δs_t^tar = Δz̅_t − π^4_{t,rw} + π^4_{t,tar}. - Policy rule for Δs_t^pol: Δs_t^pol = c1_spol Δs_{t−1}^pol + (1−c1_spol){Δs_t^tar − c2_spol ẑ_t − c3_spol π^4_t^dev} + ε_t^{spol}. - Fiscal block: - Structural primary deficit reaction function anchors public debt accumulation and provides countercyclical support; fiscal impulse fisc_t built as 4-quarter moving average of fimp_t^rat. - Debt decomposition: debt_t^rat = debt_t^{m,rat} + debt_t^{nm,rat}; all public debt is domestic in the model. - New issuances debt1y_t^{j,rat} = share_t^j · def_t^rat + debt1y_{t−4}^{j,rat} exp(Δny_t /4 100) − restruct_t + ε_t^{debt1y_t^{j,rat}}. - Interest rate on 1-year marketable debt: i4_t^m = E_t( (r^s_t + r^s_{t+1} + r^s_{t+2} + r^s_{t+3})/4 + tprem_t^m ). - Term premia respond to debt deviation: tprem_t^j = ρ_tprem tprem_{t−1}^j + (1−ρ_tprem)[tprem_{j,ss} + c1_tprem (debt_t^rat − debt^rat,tar)] + ε_t^{tprem_j}. --- ### Impulse Response Findings (4.1) - Global energy price shock (10 percent increase in global real energy price in one quarter): - extractive sector expands; positive spillovers to non-extractive sector; - current account improves; dinar appreciates (nominal appreciation despite BA adopting a more depreciated nominal exchange rate reference); - fiscal space increases allowing above-trend structural deficits; - headline annual CPI inflation rises close to 0.2 percentage points above the steady state; CPI pass-through low (maximum 0.2 percentage points for a 10 percent energy price shock); - BA tightens M2 and raises interest rates to curb non-extractive sector overheating and to re-anchor core inflation. - Domestic non-extractive sector demand shock (positive, short-lived): - higher domestic prices and imports; current account worsens; nominal depreciation pressure; - BA lowers M2 growth below target and tightens interest rates; sets more depreciated nominal exchange rate reference given real overvaluation and slightly above-target inflation. - Core prices shock (unexpected cost-push in core prices): - headline CPI and real exchange rate overvalue moderately; - BA reduces M2 growth below target, raises interest rates; non-extractive output declines; - fiscal position improves in short run but interest cost increases push debt ratio upward, threatening sustainability. - Fiscal expansion shock (1 percent of GDP increase in structural deficit over four quarters): - non-extractive output rises by about 0.4 percent by end of first year; - headline inflation rises (core increases relatively more); current account worsens and depreciation pressures mount; sovereign risk amplifies depreciation pressures; - BA tightens M2 and increases interbank rates; sets a more appreciated nominal exchange rate reference to limit imported price pass-through; - model suggests fiscal reversal in second year as public debt ratio increases significantly. - Exchange rate depreciation shock (temporary nominal depreciation pressures): - imported cost pressures and inflation; improved current account via competitiveness; positive spillovers to non-extractive industries; - BA reduces M2 growth; fiscal policy reduces structural deficit; nominal exchange rate stabilizes around new level within about two years. --- ### Model-based historical interpretations (2003Q1–2024Q4) (4.3) — key narratives - Real economy: - Extractive sector not dominant in direct GDP share but critical for external and fiscal balances; extractive output trend declined post-global financial crisis. - Non-extractive GDP recovered to about 4 percent on average after 2021; non-extractive output gap is main driver of domestic inflationary pressures. - Pre-2008: neutral output gap on average; 2010–2019 growth slowed to about 2.8 percent, dipping to around 2 percent pre-COVID. - COVID-19 produced severe contraction and negative output gap; post-2020 rebound restored consumption and generated inflationary pressures. - Balance of payments: - Trade balance surplus pre-2013 transitioned into deficits thereafter; trend stabilized close to zero by 2023. - Exports driven by real oil prices and extractive production; imports driven by non-extractive output gap and fiscal deficit. - Fiscal policy and debt: - Debt-to-GDP roughly stable around 7 percent after global financial crisis, rose starting 2015, peaked close to 60 percent, recently stabilized below 50 percent; increase mainly via non-marketable debt. - Interest costs remained relatively low due to non-marketable debt composition. - Monetary policy and exchange rate: - Real interest rate and real exchange rate gaps drive monetary conditions; notable nominal depreciations in 2009 and 2015 (2015: 34 percent dinar depreciation contributed to inflation). - In 2023 dinar appreciated nearly 5 percent supported by hydrocarbon price recovery and current account surpluses. - Low persistence of non-extractive output gap: b1y ne =0.4. - Inflation history and drivers: - Core inflation target: 4 percent; core represents 83 percent of consumption basket. - Historical episodes: 2012 inflation 9 percent; 2015–16 spikes (34 percent depreciation in 2015, inflation 6.4 percent in 2016); 2021–2023 inflation accelerated exceeding 9 percent at peak. - Core inflation sample average 4.2 percent; non-core 7.1 percent. - Major contributors to core inflation variability: domestic demand (non-extractive output gap), RER gap, and global energy prices. --- ### In-sample simulations, forecast performance, and RMSEs (4.4) - Recursive in-sample forecasts (conditional on foreign variables and domestic targets) from 2010Q1 to 2024Q3 with 8-quarter horizons; perfect foresight assumed for conditioned variables. - General performance: - Simulations closely follow observed data; forecast errors relatively small and balanced. - Hydrocarbon sector introduces variability and biases, notably 2016–2020. - Inflation forecasts slightly more volatile than actuals; fresh food subcomponent exhibits larger forecast errors. - Nominal exchange rate forecasts align closely except during 2015 and 2022 extraordinary depreciations. - Simulated M2 growth aligns generally with observed data, reflecting smoothing in policy reaction. - Current account dynamics and government debt forecasts replicate observed patterns without systematic bias. - AMPM RMSE (by horizon) - GDP growth (% yoy): 1-q 1.44; 2-q 1.85; 3-q 2.04; 4-q 2.33; 5-q 2.42; 6-q 2.44; 7-q 2.35; 8-q 2.31 - CPI inflation (% yoy): 1-q 0.92; 2-q 1.42; 3-q 1.76; 4-q 2.23; 5-q 2.21; 6-q 2.20; 7-q 2.15; 8-q 2.12 - Exchange rate (DZD per USD): 1-q 2.01; 2-q 3.16; 3-q 3.94; 4-q 4.51; 5-q 4.87; 6-q 5.06; 7-q 5.11; 8-q 5.13 - M2 growth (% yoy): 1-q 1.20; 2-q 2.11; 3-q 2.98; 4-q 3.74; 5-q 3.87; 6-q 3.91; 7-q 3.87; 8-q 3.79 - Current account balance (% of GDP): 1-q 2.77; 2-q 3.50; 3-q 3.88; 4-q 4.05; 5-q 4.12; 6-q 4.17; 7-q 4.18; 8-q 4.21 - Government debt (% of GDP): 1-q 2.59; 2-q 3.17; 3-q 3.87; 4-q 4.38; 5-q 5.05; 6-q 5.73; 7-q 6.49; 8-q 7.29 - AMPM RMSE relative to RW RMSE (by horizon) - GDP growth (% yoy): 1-q 0.68; 2-q 0.62; 3-q 0.61; 4-q 0.62; 5-q 0.66; 6-q 0.70; 7-q 0.65; 8-q 0.63 - CPI inflation (% yoy): 1-q 0.68; 2-q 0.70; 3-q 0.67; 4-q 0.70; 5-q 0.64; 6-q 0.61; 7-q 0.57; 8-q 0.56 - Exchange rate (DZD per USD): 1-q 0.80; 2-q 0.73; 3-q 0.66; 4-q 0.60; 5-q 0.55; 6-q 0.50; 7-q 0.46; 8-q 0.42 - M2 growth (% yoy): 1-q 0.55; 2-q 0.57; 3-q 0.60; 4-q 0.63; 5-q 0.59; 6-q 0.56; 7-q 0.53; 8-q 0.52 - Current account balance (% of GDP): 1-q 0.93; 2-q 0.77; 3-q 0.69; 4-q 0.63; 5-q 0.57; 6-q 0.52; 7-q 0.48; 8-q 0.44 - Government debt (% of GDP): 1-q 0.82; 2-q 0.78; 3-q 0.74; 4-q 0.75; 5-q 0.74; 6-q 0.75; 7-q 0.76; 8-q 0.76 - Implications: - AMPM outperforms random walk across variables and horizons (ratios below 1). - Useful for medium-term forecasts, especially nominal exchange rate, current account, and debt ratio where relative RMSE improves with horizon. - Planned enhancements: regular out-of-sample evaluations, decompositions of forecast changes, and integrating exogenous fresh food CPI outlook. --- ### Conclusions and policy-relevant messages (5. Conclusion) - AMPM supports BA FPAS modernization and forward-looking policymaking; core contributions: - New-Keynesian semi-structural trend-gap framework with extractive/non-extractive decomposition and core/non-core CPI split. - Fiscal sector representation with marketable and non-marketable debt; balance of payments block driven by oil and gas exports. - Explicit modeling of weak monetary transmission due to shallow financial markets and SOB dominance. - Policy representation: - Monetary policy: forward-looking M2 reaction function; BA uses broad money as main intermediate target while actively managing a nominal exchange rate reference to align REER to fundamentals. - Fiscal policy: structural deficit reaction function anchors public debt accumulation. - Strengths and limitations: - Strengths: robust tool for policy analysis and forecasting with consistent transmission mechanisms. - Limitations: challenges capturing the pandemic downturn and hydrocarbon volatility; inflation forecast volatility. - Suggested improvement: integrate exogenous outlook for fresh food prices to enhance headline inflation accuracy. - Operationalization: - AMPM adaptable for existing and prospective policy regimes; complements internal BA processes and external communications. --- ### Selected calibrated parameter values and steady-state choices (0.4 and calibration summary) - Real economy key parameters: - b2_y_ne: 0.4; b3_y_ne: 0.2; b4_y_ne: 0.1; c1_y_e: 0.3; c2_y_e: 0.6; w_y_e: 0.05; ρ_Δy_e: 0.1; ρ_Δy_ne: 0.8; ρ_Δny_ne: 0.85; Δy̅_ss_e: 0; Δy̅_ss_ne: 0.5; Δny̅_ss_ne: 4.5; Non-extractive nominal GDP trend growth: 8 - Aggregate supply / inflation parameters: - w_p_c: 0.83082; c1_π_c: 0.5; c2_π_c: 0.2; c1_rmc_c: 0.4; c2_rmc_c: 0.035; c1_π_nc: 0.5; c2_π_nc: 0.1; c1_rmc_nc: 0.5; c2_rmc_nc: 0.025; ρ_rp: 0.7; Δrp_ss: -2.5 - Balance of payments: - c1_tb̂_rat: 0.6; c2_tb̂_rat: 0.4; c3_tb̂_rat: 1; c4_tb̂_rat: 3; c5_tb̂_rat: 0.3; c6_tb̂_rat: 0.6; c7_tb̂_rat: 1; c1_tb̅_rat: 0.95; c2_tb̅_rat: 1; c1_ca_pother,rat: 0.25; c1_kfa_nonres,rat: 0.75; c2_kfa_nonres,rat: 0.1; tb̅_ss_rat: -2; ca_pother,rat_ss: -0.6; kfa_nonres_ss_rat: 0.4 - Monetary policy parameters: - c1_m: 0.6; c2_m: 0.5; c3_m: 1.2; c4_m: 0.6; c5_m: 0.6; c6_m: 0.4; c1_mtar: 0.7; c2_mtar: 0.2; c1_rm: 0.7; c2_rm: 2.5; c3_rm: 1.1; c4_rm: 0.2; c5_rm: 0.1; c1_Δrm̅: 1.2; c1_mci: 0.6; ρ_r: 0.7; ρ_π_t4,tar: 0.99; ρ_v: 0.9; π_ss4,tar: 4; r̅_ss: -1.5 - Exchange rate parameters: - c1_s: 0.75; c2_s: 0.1; c3_s: 0.1; c4_s: 0.06; c1_spol: 0.75; c2_spol: 0.6; c3_spol: 0.2; ρ_Δz: 0.9; Δz̅_ss: 0 - Fiscal policy parameters and steady states: - c1_pdef_rat,cyc: 0.1; c2_pdef_rat,cyc: 0.035; ρ_def: 0.5; c1_pdef_rat,struc: 0.2; c2_pdef_rat,struct: 0.4; c1_fimp_rat: 0.75; c1_tprem: 0.01; ρ_debt: 0.99; ρ_θ: 0.9; ρ_restruct: 0.7; ρ_tprem: 0.75; debt_rat_ss: 48; θ_ss: 1; prem_m,ss: 1.75; prem_nm,ss: 0; i4_nm,ss: 1 - Foreign economy parameters: - ρ_zen̅: 0.95; ρ_zen̂: 0.75; ρ_π_rw: 0.25; ρ_rs_rw: 0.9; ρ_y_rw: 0.85; ρ_r_rw: 0.95; r_ss_rw̅: 0; π_ss_rw̅: 2; Δzen_ss̅̅̅̅̅: 0 *Source: IMF Working Paper — Algeria Macroeconomic Projection Model (AMPM), 1. Introduction; chapters 2.1, 3.1, 3.3, 4.1, 4.3, 4.4, 5. Conclusion; calibration annex excerpts (wpiea2026025-source-pdf).* ### 1. Introduction ........................................................................................................ ### 1. Introduction ### Purpose and scope of AMPM - Documents the structure and properties of the Banque d’Algérie (BA)’s Quarterly Projection Model (QPM), dubbed AMPM, developed within the BA’s Forecasting and Policy Analysis System (FPAS) with IMF technical assistance. - AMPM is tailored to capture Algeria-specific features: - importance of hydrocarbon exports (oil and gas); - fiscal policy impact and monetary–fiscal interactions; - monetary aggregate targeting framework prioritizing price stability (defined by an implicit core inflation target); - managed floating exchange rate arrangement; - quasi-closed capital account. ### BA’s existing monetary policy framework (de jure and de facto) - Legal and operational context: - 2023 Monetary and Banking Law (MBL) specifies BA’s objective with price stability as the central monetary policy goal (Article 35). - De jure framework: monetary aggregate targeting, relying on base money as the operating target. - Operating instruments: liquidity management (required reserves, open market operations, bilateral absorptions), interest rates (BA policy rate, deposit and marginal lending facility rates), and exchange rate management. - Broad money supply measured by M2 and guided toward growth rate targets aligned with price stability (implicit core inflation target). - Historical practice: - M2 growth targets have had an indicative value rather than strict numerical objectives, with occasional large deviations of M2 growth from the target and significant fluctuations in interbank interest rates. - Multiple nominal anchors have been pursued, including M2 growth, the nominal exchange rate to achieve REER equilibrium, and an implicit inflation target currently defined as 4 percent core inflation. - Institutional constraints noted: - The new MBL limits Treasury overdrafts (Article 48) and provides operational tools to modernize the framework and enhance interest rate role. ### Structural impediments weakening monetary transmission - Banking system specificities with policy implications: - dominant role of state-owned banks (SOBs) accounting for over 85 percent of total assets; - SOBs practice subsidized loan rates (taux bonifiés) and benefit from government guarantees; - highly uneven distribution of liquidity, with SOBs receiving hydrocarbon export revenues; - lack of meaningful competition among financial institutions; - complex nexus between state-owned enterprises (SOEs), SOBs, and the government; - excess liquidity. - Consequences: - weakened policy transmission and undermined signaling role of BA policymaking. - During major shocks, BA extended special financial resources to support government policies (example provided below). ### Example of policy response during shock episodes - Special refinancing program (Programme Spécial de Refinancement, PSR): - Implemented to support post-pandemic recovery. - PSR was a one-year program totaling about 8.3 percent of GDP, launched in 2021 to facilitate public banks’ refinancing with the BA using government debt securities in exchange for long-term syndicated loans to state-owned enterprises purchased by the Treasury. ### Exchange rate regime and BA’s role in FX market - Regime characterization: - De jure: managed floating exchange rate arrangement (BA actively intervenes to influence level or limit fluctuations). - De facto classification noted as crawl-like in AREAER; FX market features lead to endogenous reserve adjustments. - BA interventions and mechanics: - BA is the dominant FX supplier due to mandatory conversion of commodity exporters’ FX inflows. - BA sets a nominal exchange rate reference value available to banks using a very narrow bid-ask margin. - Targeting aims: close perceived REER misalignment gradually and avoid undue volatility. - Target REER is estimated using an empirical framework with foreign and domestic variables (including oil prices and budget balance). - BA may consider deviation of core inflation from the implicit target when setting the nominal exchange rate. - Capital account and reserves: - Capital account restrictions and limited private sector FX flows imply FX reserves adjust endogenously to current account surpluses (accumulation) or deficits (decumulation). ### Modeling choices and sectoral decomposition in AMPM - GDP decomposition: - Total GDP is decomposed into extractive (mainly hydrocarbon) and non-extractive sectors. - Extractive sector: heavily impacted by global energy (oil and gas) prices and largely exogenous to domestic developments. - Non-extractive sectors: reflect domestic business cycle and demand-side inflationary pressures that policymakers can influence. - Prices and inflation: - Supply-side distinction between core and non-core (fresh agricultural goods) prices. - Central bank measures have more traction on core prices. - External balance block: - Exports primarily hydrocarbon goods driven by global energy prices. - Capital flows relatively small given quasi-closed capital account. - Current account positions counterbalanced by changes in official reserves given BA’s role as FX supplier. ### Policy rules and behavioral equations embedded in AMPM - Central bank behavior modeled via two reaction functions: - One for M2 growth (monetary aggregate target). - One for setting the nominal exchange rate reference value. - Determinants of money supply in the model include: - expected deviation of core inflation from the implicit target; - cyclical positions of output and real exchange rate; - announced or implicit M2 growth targets; - impact of not fully sterilized official reserves flows (consistent with historical correlations). - Exchange rate dynamics account for observed rigidity and BA stabilization efforts, aiming for gradual alignment with fundamental value while considering price stability. - Fiscal block: - Public domestic debt accumulation decomposed into marketable and non-marketable components. - Fiscal reaction function for the structural deficit to ensure debt sustainability and provide countercyclical support. ### Calibration, validation, and operational use - Calibration aims: - Provide economically intuitive shock propagation mechanisms and reasonable fit to Algerian data. - Model performance checks: - Impulse responses offer reasonable interpretation of economic effects and required policy reactions. - Historical decompositions provide narratives for key episodes. - In-sample simulations confirm empirical coherence. - Operationalization goals: - Establish AMPM as central element of BA’s FPAS and regular input for BA decision-making. - BA staff use specialized sectoral analysis, nowcasts, near-term forecasts, and complementary analyses to form initial conditions, construct alternative scenarios, and inform policy. - Activities organized via a day-by-day forecast calendar with tasks, deadlines, responsibilities, and deliverables. - Scenario analysis: - AMPM used to build baseline forecasts and alternative scenarios (e.g., deviations of global energy price trajectories) to analyze implications for monetary policy in a model-consistent framework. ### Model adaptability and contributions to policy debates - AMPM can be adjusted and recalibrated to simulate prospective monetary policy framework designs and strategies, including transitions: - Example application: evaluate implications of moving from a quantity-based framework to an interest rate-based one, with a clearer nominal anchor. - Relationship to literature: - Based on canonical New-Keynesian structure introduced in Berg et al. (2006) with Algeria-specific extensions and similarity to large-scale, multiple-instrument frameworks in Berg et al. (2023). - Comparable semi-structural models referenced for Jordan and the Philippines. - Contributes to limited analytical literature on Algeria; prior work includes IMF (2016, 2017), Meibel and Belarbi (2017), and Bank of Algeria (2022). *Source: IMF Working Paper — Algeria Macroeconomic Projection Model (AMPM), 1. Introduction* ### 2.1 Inflation: Core and Non-Core Components ### 2.1 Inflation: Core and Non-Core Components ### Definitions and measurement - The consumer price index (CPI) is decomposed into core and non-core components in BA’s analytical framework to reflect differing statistical properties, driving forces, and implications for monetary policy. - Non-core items: - Account for 17 percent of the expenditure basket. - Include fresh food, whose prices are impacted by weather conditions and government price control measures largely outside the BA’s influence. - Core prices: - Comprise processed food, manufactured goods, and services. - Are partly influenced by internal demand and the exchange rate. - Are used to define the medium-term price stability objective, with a 4 percent annual core inflation goal. - The base year for Algeria’s CPI basket is 2001 and the weights of CPI items are significantly outdated, complicating decision making. ### Historical dynamics and stylized facts - Over the past two decades, Algerian inflation has been influenced by: - Government spending. - Agricultural challenges and weather-related supply shocks. - The COVID-19 pandemic. - Core inflation behavior: - Relatively stable during 2003-15. - Annual core inflation averaged 3.5 percent during 2003-15. - Core inflation fluctuated between 0.5 and 6.5 percent in that period. - Sample average core inflation across the full sample is 4.2 percent. - Non-core inflation behavior: - Subdued dynamics in the mid-2000s and pre-pandemic, with significant spikes in 2008-12 and more recently. - Sample average non-core inflation across the full sample is 7.1 percent. - Recent drivers of higher inflationary pressures: - Higher import prices due to dinar depreciation. - Excess liquidity from increased current public spending. - Inefficiencies in distribution chains (e.g., logistical delays and input shortages in certain sectors). - COVID-19 effects: - Initial phase produced significant deflationary effects via weak global and domestic demand. - Subsequent reopening and recovery led to accelerated core prices due to increased demand and global supply chain disruptions. - The long-term discrepancy between core and non-core inflation (4.2 percent for core vs. 7.1 percent for non-core) indicates persistent relative price trends explicitly accounted for in AMPM. ### Analytical and policy implications - Core inflation is the target for the BA’s medium-term price stability objective (4 percent annual core inflation goal). - Non-core volatility (notably fresh food) reflects factors largely outside BA’s direct control, which justifies the decomposition in policy analysis. - The outdated CPI base year and weights (base year 2001) complicate monetary decision making and are a recognized data limitation. - AMPM explicitly incorporates the distinct dynamics of core and non-core inflation to reflect their different drivers and implications for monetary policy. *Source: IMF working paper chapter 2.1, “Inflation: Core and Non-Core Components”* ### 3.1 Internal Balance ### 3.1 Internal Balance ### 3.1.1 Real Economy and Aggregate Demand - Real GDP is disaggregated into extractive (EGDP, denoted with e; “hydrocarbon GDP” label used interchangeably) and non-extractive (NEGDP, denoted with ne; “non-hydrocarbon GDP” label used interchangeably) components. Real variables are split into gap and trend components. - Sectoral growth and gap aggregation: - ∆yt = wy_e ∆yt^e + (1−wy_e) ∆yt^ne + ε^y_t (equation (1)) - ŷ_t = wy_e ŷ_t^e + (1−wy_e) ŷ_t^ne (equation (2)) - Note: wy_e is assumed constant; measurement error ε^y_t accounts for time variation in sectoral shares and seasonal adjustment discrepancies. - NEGDP output gap (proxy for fundamental business cycle dynamics) drives domestic marginal costs and inflation pressures. - NEGDP open-economy IS curve (output gap determinants): - ŷ_t^ne = b1_y^ne ŷ_{t−1}^ne − b2_y^ne mci_t + b3_y^ne tb̂_t^rat + b4_y^ne fisc_t + ε^y_ne_t (equation (3)) - Determinants: past gap, real monetary conditions index (mci_t), trade balance gap (tb̂_t^rat), fiscal impulse (fisc_t), NEGDP demand shocks ε^y_ne_t. - Trade balance gap used instead of foreign output gap to capture external flows and terms-of-trade effects in an energy-exporting economy; allows for spillovers from extractive sector to NEGDP via fiscal impulse and demand channels. - Monetary conditions index (mci_t) definition: - mci_t = c1_mci r̂_t + (1−c1_mci)(−ẑ_t) (equation (4)) - c1_mci denotes relative weight of real interest rate component; ẑ_t is real exchange rate gap (positive/negative values indicate undervaluation/overvaluation). - Extractive sector output gap dynamics: - ŷ_t^e = c1_y^e ŷ_{t−1}^e + c2_y^e ẑen_t + ε^y_e_t (equation (5)) - Energy price composite gap: - ẑen_t = w_oil ẑoil_t + (1−w_oil) ẑgas_t (equation (6)) - Real oil and gas prices computed as nominal USD prices relative to US CPI (ptrw), then decomposed into trend and gap: - zoil_t = oil_t − ptrw_t (equation (7)) - zoil_t = zoil̅_t + ẑoil_t (equation (8)) - zgas_t = gas_t − ptrw_t (equation (9)) - zgas_t = zgas̅_t + ẑgas_t (equation (10)) - w_oil denotes the (constant) weight of oil in the energy composite. ### 3.1.2 Inflation and Aggregate Supply - Headline CPI (p_t) is a weighted average of core (p_t^c) and non-core/fresh food (p_t^nc) price indices plus a discrepancy term (disc^p_t) for log-approximation and seasonal adjustment non-additivity: - p_t = w_p^c p_t^c + (1−w_p^c) p_t^nc + disc^p_t (equation (11)) - w_p^c denotes the weight of the core component. - Relative price of core to non-core: - rp_t = p_t^c − p_t^nc (equation (12)) - rp_t = rp̅_t + r̂p_t (equation (13)) - Trend growth in relative price follows autoregressive process: - ∆rp̅_t = ρ rp̅_{t−1} + (1−ρ) ∆rp_ss + ε^∆rp̅_t (equation (14)) - ∆rp_ss denotes the steady state relative price growth; ε^∆rp̅_t is a trend shock. - Exogenous core inflation target is specified as 4 percent (reflecting the BA definition of price stability). The core target and steady-state relative price trend growth determine an implicit target for non-core inflation. - Forward-looking Phillips curve for core inflation (quarter-on-quarter annualized core inflation rate π_t^c): - π_t^c = c1_π^c E_t π_{t+1}^c + (1−c1_π^c) π_{t−1}^c + c2_π^c rmc_t^c + ε^π_c_t (equation (15)) - E_t denotes model-consistent rational expectations; rmc_t^c is real marginal cost for core; ε^π_c_t is sectoral cost-push shock. - Real marginal cost (RMC) components for core inflation: - rmc_t^c = c1_rmc^c ŷ_t^ne + c2_rmc^c ẑen_t + (1−c1_rmc^c − c2_rmc^c) ẑ_t (equation (16)) - Components: NEGDP output gap, relative energy price gap, core-based real exchange rate gap. - Non-core (fresh food) Phillips curve analogous to core but with sector-specific calibration: - π_t^nc = c1_π^nc E_t π_{t+1}^nc + (1−c1_π^nc) π_{t−1}^nc + c2_π^nc rmc_t^nc + ε^π_nc_t (equation (17)) - rmc_t^nc = c1_rmc^nc ŷ_t^ne + c2_rmc^nc ẑen_t + (1−c1_rmc^nc − c2_rmc^nc)(ẑ_t + r̂p_t) (equation (18)) - Note: non-core RMC real exchange rate incorporates relative price gap term (ẑ_t + r̂p_t) because real exchange rate is defined using core prices. ### 3.2 External Balance - Balance of payments identity (all variables as ratios to nominal GDP): - ca_t^rat + kfa_t^rat = 0 (equation (19)) - ca_t^rat comprises net exports of goods and services plus net primary and secondary income; kfa_t^rat consists of financial flows including official reserves. - Trade balance focus: Algeria’s current account is mainly driven by goods and services; external block models trade balance (tb_t^rat) determinants. Net primary and secondary income summarized by residual ca_t^other,rat. - Current account and trade balance decomposition: - ca_t^rat = tb_t^rat + ca_t^other,rat (equation (20)) - câ_t^rat = tb̂_t^rat (equation (21)) (current account gap equals trade balance gap) - tb_t^rat = tb̂_t^rat + tb̅_t^rat (equation (22)) - Trade balance gap behavioral equation: - tb̂_t^rat = c1_tb̂_rat tb̂_{t−1}^rat + (1−c1_tb̂_rat)∙( c2_tb̂_rat ẑen_t + c3_tb̂_rat ŷ_t^rw − c4_tb̂_rat ŷ_t^ne + c5_tb̂_rat ŷ_t^e + c6_tb̂_rat ẑ_t − c7_tb̂_rat (pdef_t^rat,struct − pdef_t^rat,tar) ) + ε^{tb̂}_t (equation (23)) - Determinants: real energy price gap (ẑen_t), extractive output gap (ŷ_t^e), foreign demand (ŷ_t^rw), competitiveness (ẑ_t), NEGDP output gap (import demand), public import demand proxied by (pdef_t^rat,struct − pdef_t^rat,tar). Deviations captured by ε^{tb̂}_t. - Trade balance trend dynamics: - tb̅_t^rat = c1_tb̅_rat tb̅_{t−1}^rat + (1−c1_tb̅_rat)( tb̅_ss^rat + c2_tb̅_rat(∆zen̅_t − ∆zen̅_ss) ) + ε^{tb̅}_t (equation (24)) - tb̅_ss^rat denotes steady state trade balance trend; adjusted for deviation of real energy price trend growth from steady state. - Remaining current account trends: - ca̅_t^rat = tb̅_t^rat + (ca_t^other,rat − ε^{ca_other,rat}_t) (equation (25)) - ca_t^other,rat = c1_ca_other,rat ca_{t−1}^other,rat + (1−c1_ca_other,rat) ca_ss_other,rat + ε^{ca_other,rat}_t (equation (26)) - Financial account decomposition: - kfa_t^rat = kfa_nonres_t^rat − resfin_t^rat (equation (27)) - Given Algeria’s largely closed capital account, the change in reserves covers the current account balance (kfa_t^rat = −ca_t^rat). - Endogenous non-reserve financial inflows behavior (converge to equilibrium kfa_nonres_ss^rat; sensitive to sovereign risk premium prem_t but calibrated responsiveness is small): - kfa_nonres_t^rat = c1_kfa_nonres_rat kfa_nonres_{t−1}^rat + (1−c1_kfa_nonres_rat) kfa_nonres_ss^rat ... + c2_kfa_nonres_rat (prem_t − prem_ss) + ε^{kfa_nonres}_t (equation (28)) - Law of motion for reserve assets (res_t^rat): - res_t^rat = resfin_t^rat/4 + res_{t−1}^rat (1 + ∆s_t/400) / (1 + ∆ny_t^ne/400) + ε^{res}_t (equation (29)) - resfin_t^rat are financing needs for current account transactions; ∆s_t is nominal exchange rate change; ∆ny_t^ne is nominal NEGDP growth (quarterly/annualization scaling shown by 400 factors). - Total net foreign assets: - nfa_t^rat = res_t^rat + nfa_t^priv,rat (equation (30)) - In Algeria, majority of net foreign assets are BA reserve assets due to commodity export surrender requirement. *Source: wpiea2026025-source-pdf - 3.1 Internal Balance (IMF Working Papers Algeria Macroeconomic Projection Model (AMPM))* ### 3.3 Economic Policies ### 3.3 Economic Policies ### 3.3.1 Monetary Policy - Since 2003, monetary aggregates have been central to the BA’s monetary policy formulation; AMPM specifies a forward-looking monetary policy reaction function for money supply defined in terms of the M2 aggregate as an intermediate target. - Rationale and modeling choices: - The approach models broad money (M2) directly as the central bank instrument, rather than the narrow monetary base and/or reserves. - M2 growth dynamics incorporate impacts of BA instruments (e.g., open market operations, required reserve ratios) and serve as a proxy for overall monetary policy stance. - Objectives included in the M2 reaction function: - Achieve medium-term price stability by adjusting the M2 quarter-on-quarter annualized growth rate (Δm_t) based on three-quarter ahead expected deviations of annual core inflation from its target. - Smooth business cycle fluctuations by closing the total output gap (ŷ_t) and stabilizing the interbank interest rate (r^s_t) near its neutral level (r^s_t^neutral). - Align money supply with intermediate targets for broad money annual growth (Δ^4 m_t^tar). - Account for changes in official reserves (res_t^rat − res_{t−1}^rat) if not fully sterilized. - Include persistence through Δm_{t−1} and a money supply shock (ε_t^Δm). - Core M2 reaction function (equation preserved in structure): - Δm_t = c1_m Δm_{t−1} + (1−c1_m)[Δr̅^m_t + (π^4_t^tar − (1−w_pc) Δr^p_t^tnd) ... + c2_m (r^s_t − r^s_t^neutral) − c3_m π^4_t^dev − c4_m ŷ_t − c5_m (Δm_t − Δ^4 m_t^tar)] ... + c6_m (res_t^rat − res_{t−1}^rat) + ε_t^Δm (31) - Definitions and additional relations: - Δr̅^m_t represents trend real M2 growth; the nominal counterpart is (Δr̅^m_t + (π^4_t^tar − (1−w_pc) Δr^p_t^tnd)). - Price stability determinant π^4_t^dev is defined as (E_t π^4_{t+3,c} − π^4_{t+3,tar}). - Real money growth trend specification (equation 32): - Δr̅^m_t = w_y^e ∆ȳ^e_t + (1−w_y^e) ∆ȳ^ne_t − Δv_t − c1_Δr̅^m (r̅_t − r̅_{t−1}) + ε_t^Δr̅^m - Real money definition (equation 33): r^m_t = m_t − p_t - Real money target (equation 34): r^m_t^tar = m_t^tar − p_t - Nominal money growth target dynamics (equation 35): - Δm_t^tar = c1_mtar Δm_{t−1}^tar + (1−c1_mtar) ∆nȳ^ss_ne̅̅̅ − c2_mtar (r^m_{t−1}^tar − r̅^m_{t−1}) + ε_t^{mtar} - Real money demand (real money gap r̂^m_t) depends on real interest rate gap (r̂_t), total output gap (ŷ_t), government financing needs (debt_t^rat − debt^rat,tar), liquidity effects from official reserves (res_t^rat − res_{t−1}^rat), and money demand shock (ε_t^{r^m}) (equation 36): - r̂^m_t = c1_r^m r̂^m_{t−1} − c2_r^m r̂_t + c3_r^m ŷ_t + c4_r^m (debt_t^rat − debt^rat,tar) + c5_r^m (res_t^rat − res_{t−1}^rat) + ε_t^{r^m} ### 3.3.2 Exchange Rate Policy - De jure and de facto arrangements: - AREAER classifies Algeria’s de jure exchange rate arrangement as “managed floating”. - De facto arrangement has been classified as crawl-like or stabilized in recent years. - AMPM assumptions: - Assumes an explicit nominal exchange rate depreciation target reflecting the long-term dinar depreciation trend. - With a largely closed financial account and mandatory exchange of FX income by commodity exporters to the BA, the Bank can set the exchange rate reference path. - Exchange rate determination: - The exchange rate s_t is determined primarily by the BA reference rate (policy target depreciation rate Δs_t^pol with weight c1_s) and to a complementary extent (1−c1_s) by a modified UIP condition. - Modified UIP condition links current-quarter exchange rate to next-period expected exchange rate s^F_{t+1} plus a risk-adjusted interest differential (r^s_t − r^s_{rw} − prem_t). - Other pressures on s_t include trade balance developments (t̂b_t^rat) and excessive government debt accumulation (debt_t^rat − debt^rat,tar), with coefficients c3_s and c4_s respectively; these can be interpreted as an endogenous component of the sovereign risk premium. - Exchange rate equation (preserved structure, equation 37): - s_t = c1_s {s_{t−1} + Δs_t^pol /4} + ... (1−c1_s){s^F_{t+1} − (r^s_t − r^s_{rw} − prem_t)/4} + ... (1−c1_s){−c3_s t̂b_t^rat + c4_s (debt_t^rat − debt^rat,tar)} + ε_t^s - Expectations and long-term linking to inflation (equations 38 and 39): - Exchange rate expectations: s^F_{t+1} = c2_s E_t s_{t+1} + (1−c2_s){s_{t−1} + 2/4 Δs_t^tar} (38) - Long-term PPP and exchange rate target depreciation: Δs_t^tar = Δz̅_t − π^4_{t,rw} + π^4_{t,tar} (39) - Exchange rate policy target dynamics: - BA sets the dinar reference value Δs_t^pol to mitigate excessive volatility and target a real exchange rate aligned with fundamentals (close real exchange rate gap ẑ_t), and to mitigate price pressures via expected core inflation deviation π^4_t^dev. - Policy rule for Δs_t^pol (equation 40): - Δs_t^pol = c1_spol Δs_{t−1}^pol + (1−c1_spol){Δs_t^tar − c2_spol ẑ_t − c3_spol π^4_t^dev} + ε_t^{spol} ### 3.3.3 Fiscal Policy - Fiscal block overview: - Parsimonious fiscal block anchors public debt accumulation via a fiscal reaction function for the structural primary deficit (adjusted for domestic NEGDP business and commodity price cycles). - Provides stock-flow consistency for public debt trajectory and government total and primary deficits. - Fiscal variables are expressed as shares of total nominal GDP and the framework focuses on aggregate balance measures (not detailed revenue/expenditure components). - Can be complemented with spreadsheet models or finance law forecasts for richer fiscal detail. - Capabilities of the fiscal block: - Analyze macroeconomic effects of fiscal measures (stimulus/restrictiveness). - Project public debt and required fiscal adjustments for debt sustainability. - Evaluate macro implications of financing choices: marketable debt vs non-marketable debt. - Capture monetary-fiscal interactions. - Government deficit decomposition (equation 41): - def_t^rat = pdef_t^rat + intcost_t^rat + ε_t^{defrat} - Model simplification: outstanding debt represented by generic 1-year maturity instruments. - Primary deficit decomposition (equation 42): - pdef_t^rat = pdef_t^rat,struct + pdef_t^rat,cyc + ε_t^{pdefrat} - Cyclical primary deficit (equation 43): - pdef_t^rat,cyc = −c1_pdef^rat,cyc ŷ_t^ne − c2_pdef^rat,cyc ẑ_en_t + ε_t^{pdefrat,cyc} - Captures automatic stabilizers via the non-extractive sector output gap and deviations in international real energy prices (ẑ_en_t). - Structural primary deficit reaction function (equation 44): - pdef_t^rat,struct = ρ_def (pdef_{t−1}^rat,struct − c1_pdef^rat,struct ŷ_t^ne) + (1−ρ_def)(pdef_t^rat,tar − c2_pdef^rat,struct debt̂_t^rat) + ε_t^{defstructrat} - Objectives: stabilize debt toward target (debt̂_t^rat) and smooth economic cycles (close NEGDP gap ŷ_t^ne). - Calibration includes policy smoothing via lagged term pdef_{t−1}^rat,struct. - Parameters c1_pdef^rat,struct and c2_pdef^rat,struct set relative importance of stabilization objectives; ε_t^{defstructrat} captures deviations. - Debt deviation and fiscal anchor: - Debt deviation (equation 45): debt̂_t^rat = c1_debt^rat (debt_t^rat − debt_t^rat,tar) + (1−c1_debt^rat) E_t debt̂_{t+1}^rat - Fiscal anchor: target total debt-to-GDP ratio follows autoregressive process reverting to steady state debt^rat,ss (equation 46): - debt_t^rat,tar = ρ_debt debt_{t−1}^rat,tar + (1−ρ) debt^rat,ss + ε_t^{debtrat,tar} - Fiscal impulse definition and implementation: - Fiscal impulse (equation 47) combines quarter-by-quarter change in structural deficit and deviation from steady-state structural deficit, weighted by c1_fimprat. - IS curve uses a four-quarter moving average of the fiscal impulse (equation 48): fisc_t = 1/4 (fimp_t^rat + fimp_{t−1}^rat + fimp_{t−2}^rat + fimp_{t−3}^rat) - Debt composition and accumulation: - All public debt is domestic in the model; total debt ratio (equation 49): debt_t^rat = debt_t^{m,rat} + debt_t^{nm,rat} - Debt accumulation for j ∈ {m, nm} assumes generic 1-year maturity and adjusts earlier issuances by nominal GDP changes (equation 50): - debt_t^{j,rat} = debt1y_t^{j,rat} + debt1y_{t−1}^{j,rat} exp(Δny_t/400) + debt1y_{t−2}^{j,rat} exp((Δny_t + Δny_{t−1})/400) + debt1y_{t−3}^{j,rat} exp((Δny_t + Δny_{t−1} + Δny_{t−2})/400) - New issuances reflect financing of total deficit and replacement of maturing debt (period t−4 issuances), possible reallocation between marketable and non-marketable debt (restruct_t), and measurement errors (equation 51): - debt1y_t^{j,rat} = share_t^j · def_t^rat + debt1y_{t−4}^{j,rat} exp(Δny_t /4 100) − restruct_t + ε_t^{debt1y_t^{j,rat}} - share_t^j = θ_t for j = m; share_t^j = 1 − θ_t for j = nm; θ_t proxies financing plans and varies over time. - In steady state θ_ss = 1 (deficit financed solely via marketable instruments). - Simple AR processes for θ_t and restruct_t (equations 52 and 53). - Interest expenditures and term structure: - Interest expenditures as ratio to nominal GDP for each debt type (equation 54) sum interest on all non-matured previous issuances, using 1-year bond rates i4^j lagged by issuance periods and adjusting by nominal GDP growth Δny terms. - 1-year marketable interest rate (equation 55): - i4_t^m = E_t( (r^s_t + r^s_{t+1} + r^s_{t+2} + r^s_{t+3})/4 + tprem_t^m ) - Non-marketable 1-year rate assumed fixed by authorities at i4^{nm,ss} + tprem_t^{nm} (equation 56). - Term premia dynamics (equation 57): - tprem_t^j = ρ_tprem tprem_{t−1}^j + (1−ρ_tprem)[tprem_{j,ss} + c1_tprem (debt_t^rat − debt^rat,tar)] + ε_t^{tprem_j} - Term premia capture investor risk perceptions; coefficients ρ_tprem and c1_tprem assumed identical across debt components in the model's baseline. - Other notes: - The model includes a schematic of monetary-fiscal interactions (Figure 10). - Measurement errors ε terms account for discrepancies between budget deficits and public debt data and for simplifications (1-year homogeneous debt). ### 4. Model Properties and Results — 4.1 Model Calibration (summary) - Calibration approach: - AMPM uses calibration rather than estimation due to practical considerations and data limitations. - Calibration relies on empirical evidence and theoretical benchmarks, iteratively tested to align model properties with Algerian stylized facts. - Calibration supports sensitivity analysis and controlled exploration of parameter impacts. - Rationale for calibration: - Data limitations: short maco time-series, statistical breaks, changes in policy regimes make estimation difficult or unduly prior-driven. - Calibration provides transparency and allows integration of expert knowledge. - Implementation note: - Annex C (not summarized here) presents calibrated parameter values used in AMPM. *Source: wpiea2026025-source-pdf - 3.3 Economic Policies* ### 4.1 Impulse Responses ### 4.1 Impulse Responses ### Global energy price shock (10 percent increase in global real energy price) - Shock: 10 percent increase in global (real) price of energy (oil and gas) in one quarter. - Real economy: - Expansion in extractive sector output in Algeria. - Indirect positive impact on non-extractive sector activities. - External sector and exchange rate: - Significant improvement in the current account balance. - Appreciation of the dinar (actual nominal exchange rate appreciates despite BA adopting a more depreciated nominal exchange rate reference value). - Fiscal effects: - Higher hydrocarbon export revenues allow boosting fiscal spending. - Above-trend structural fiscal deficits, consistent with historical cyclical government expenditure behavior. - Inflation and pass-through: - CPI inflation, and both core and non-core components, increase marginally. - Headline annual inflation rises close to 0.2 percentage points above the steady state. - Magnitude of pass-through is low: CPI inflation increases by maximum 0.2 percentage points for a 10 percent energy price shock. - Weak pass-through aligns with Algeria’s position as a major fuel exporter where the government regulates domestic energy prices. - Monetary policy response (BA): - BA adopts a more depreciated nominal exchange rate reference value due to real overvaluation, limited by mild above-target inflation. - BA implements a tighter M2 stance; interest rates increase to curb above-trend dynamics in the non-extractive sector and bring core inflation back to target. - Monetary tightening driven by inflation (expectations) and output above equilibria. - Model behavior: - Medium-term fluctuations and volatile interbank interest rate dynamics are typical in models approximating a monetary targeting regime with multiple policy objectives. ### Domestic non-extractive sector demand shock (positive expansionary shock) - Shock characteristics: - Reflects volatility in sectoral GDP data; short-lived impact on non-extractive output. - Real economy: - Higher demand increases domestic prices and imports. - Worsening of the current account balance and nominal depreciation pressures. - Real exchange rate becomes slightly overvalued due to higher domestic inflation and only marginal nominal depreciation. - Fiscal effects: - Countercyclical response: running lower structural deficits given above-trend output. - Reduction in the debt ratio driven by improvement in nominal GDP following the demand shock. - Monetary policy response (BA): - BA lowers M2 growth below the target as both inflation and output are above equilibrium values. - BA adopts a tighter interest rate stance. - For the nominal exchange rate, BA sets the reference value toward more depreciated levels given real overvaluation and slightly above-target inflation. ### Core prices shock (unexpected cost-push in core prices) - Real economy and prices: - Higher core inflation partially spills over into the non-core sector via relative price effects. - Moderate increase in headline CPI and real exchange rate overvaluation. - Monetary policy response: - Above-target inflation necessitates tighter monetary policy; M2 growth reduced below target in the short term. - Interest rates rise; combined with real exchange rate overvaluation, monetary conditions become restrictive. - Decline in non-extractive output helps lower price pressures and restore price stability. - Fiscal effects and risks: - Sluggish activity and reduced fiscal deficits improve the trade balance through lower import demand. - Current account-to-GDP ratio worsens due to significant price effects from real exchange rate appreciation. - Fiscal reaction indicates lower structural deficits, but significant interest rate costs (from M2 tightening) push the debt ratio upward, threatening debt sustainability. - In the short run, nominal exchange rate remains stable as depreciation pressures from higher debt and a worsening current account are counterbalanced by a higher interest rate differential and above-target inflation; BA adopts more appreciated reference values. - Policy tradeoff: - Cost-push shocks create a tradeoff for monetary authorities between stabilizing inflation and avoiding output losses. - In AMPM’s specification (approximating a framework prioritizing price stability), the central bank tightens money supply, leading to demand contraction to re-anchor inflation expectations. ### Fiscal expansion shock (1 percent of GDP increase in structural deficit over four quarters) - Shock: 1 percent of GDP increase in the structural deficit spread over four quarters. - Real economy: - Positive fiscal impulse raises non-extractive output by about 0.4 percent by the end of the first year. - Prices and external sector: - Higher demand drives price pressures; headline inflation rises with core prices increasing relatively more. - Increased import demand worsens the current account balance and exerts depreciation pressures. - Increased risk aversion due to higher public debt amplifies depreciation pressures. - Monetary and fiscal policy responses: - Central bank adopts a tighter M2 stance and increases interbank rates to reduce domestic demand-side price pressures. - BA sets a more appreciated nominal exchange rate reference value due to slight real undervaluation and above-target inflation, limiting actual nominal depreciation and containing imported cost pass-through. - The simulation suggests a reversal of fiscal policy starting in the second year because the public debt ratio increases significantly, creating debt sustainability concerns and necessitating tighter fiscal policy. - Dynamic behavior: - Medium-term fluctuations in most variables are observed as fiscal policy reverses. ### Exchange rate depreciation shock (nominal depreciation pressures) - Shock characteristics: - Unexpected but temporary reassessment leading to nominal exchange rate depreciation pressures (weaker dinar). - Real economy and prices: - Weaker dinar leads to imported cost pressures and inflationary effects. - Undervalued exchange rate strengthens the current account position, with positive spillovers on non-extractive industries, generating domestic demand-side inflationary pressures. - Policy responses: - BA reduces M2 growth to tighten the interest rate stance and create more restrictive monetary conditions as output and core inflation are above targets. - Fiscal policy acts countercyclically by reducing the structural deficit and generating negative fiscal impulses. - Exchange rate dynamics: - The nominal exchange rate stabilizes at the new level around two years after the shock, partly due to a more appreciated reference value set by BA in reaction to the initially weaker real exchange rate and above-target inflation. - Modeling note: - Semi-structural gap models like AMPM generally assume depreciation shocks are expansionary via a favorable net exports channel (undervalued real exchange rate enhances price competitiveness). - Alternative identifications (e.g., depreciation from capital outflows or risk-off pressures) may be contractionary, especially in EMDEs with open capital accounts. *IMF Working Papers — Algeria Macroeconomic Projection Model (AMPM), Section 4.1 Impulse Responses* ### 4.3 Model-Based Historical Interpretations ### 4.3 Model-Based Historical Interpretations ### Methodology and scope - Uses observed data and Kalman filtration to decompose observed variables into unobserved components (gaps and trends) and to identify structural shocks. - Provides a macroeconomic narrative for Algeria covering 2003Q1 to 2024Q4. - Findings cross-referenced with Algerian historical developments and framed through the main blocks of AMPM, emphasizing unobserved variables. ### Economic activity in extractive and non-extractive sectors - Direct contribution of the extractive sector to economic growth based on National Accounts data is not dominant, but the sector remains a critical driver of internal and external balance. - Aggregate activity and output gap in AMPM are divided into extractive and non-extractive sectors to capture distinct dynamics and impacts. - Historical patterns: - Pre-global financial crisis: gradually declining growth tendency; output gap neutral on average (demand-driven inflationary pressures subdued). - 2010–2019: economic growth slowed, averaging around 2.8 percent, and dipped to around 2 percent just before the COVID-19 pandemic; accompanied by slowing potential growth. - COVID-19 pandemic: severe contraction across extractive and non-extractive sectors, resulting in a significantly negative output gap. - Post-2020: strong rebound restoring pre-pandemic consumption levels and generating inflationary pressures through a positive output gap. - Extractive sector: - Post-global financial crisis extractive output exhibited a clear downward trend. - Extractive output gap closely tracked oil and gas prices, global demand, and Algerian extraction capacities; turned negative during 2013–2016 commodity price declines and during the COVID-19 pandemic. - Non-extractive sector: - Non-extractive GDP (services, industry, construction and agriculture) grew continuously over the past 20 years except during the pandemic; recovered to about 4 percent on average after 2021. - Non-extractive output gap is a key driver of domestic inflationary pressures. - Early period drivers: trade balance (TB) gap and fiscal policy; fiscal policy relatively restrictive between 2003 and 2008. - Post-2013 fiscal policy shifted to a largely expansionary stance due to successive large fiscal deficits amid lower oil and gas prices. - Monetary policy effects captured via real interest rate (RIR) and real exchange rate (RER) gaps forming the monetary conditions index. - Monetary policy adopted an expansionary stance when output gap was negative in 2015–18 and during COVID-19 (relatively later in the second episode, starting 2021 only); otherwise broadly neutral or acting primarily via the real exchange rate. - Trade balance gap contributed prominently negative to the domestic business cycle in 2015–17 and during COVID-19; post-pandemic improvements in the trade balance, alongside fiscal and monetary policies, contributed to economic overheating. - Low persistence of non-extractive output gap reflected in relatively small lag parameter in the IS curve: b1y ne =0.4. - Relatively small contribution of monetary policy through RER and RIR indicates weak transmission of monetary policy to non-extractive activity and inflation. ### Balance of payments and external balance - Algeria heavily reliant on international trade; oil and gas exports shape government revenues, foreign exchange reserves, and exchange rate dynamics. - Key balance of payments components (as shares of nominal GDP) are the trade balance position of the current account and the official reserves financing component of the financial account; other items treated as residuals in AMPM accounting. - Trade balance trends: - A decade of surpluses pre-2013 transitioned into deficits thereafter, with recent stabilization around a relatively balanced position. - Trade balance gap decomposition accounts for export and import determinants per equation (23). - Exports driven by global energy prices and extractive production; foreign demand has an overall marginal effect on non-hydrocarbon exports. - Imports driven by non-extractive output gap and fiscal deficit; both export and import sides affected by RER gap. - Decomposition confirms dominant effects of hydrocarbon exports (real oil prices and extractive output gap) and heavy dependence of imports on domestic demand and government acquisitions. - Historical episodes: - Pre-2008: trade balance surplus driven by strong commodity exports, improved competitiveness (undervalued RER), and high foreign demand. - Post-2008 crisis and mid-2014 to pre-COVID: TB gap negative due to declining oil and gas exports amid depressed commodity prices; weak non-extractive domestic demand and temporarily undervalued RER mitigated deterioration. - COVID-19 and persistent oil price shocks further exacerbated trade deficits and led to a sharp decline in foreign reserves. - Post-2021: trade balance improved, driven primarily by higher oil prices and some recovery in non-hydrocarbon exports; by 2023 trend stabilized close to zero and the gap balanced. ### Fiscal policy and government debt - Fiscal policy strongly influenced by extractive sector export revenues; pivotal for driving the real economy. - Government debt-to-GDP ratio trajectory: - Relatively stable at around 7 percent after the global financial crisis. - Began rising steadily in 2015, climbed continuously, peaking at close to 60 percent, before recently stabilizing below 50 percent. - Increase captured primarily by non-marketable debt instruments. - Rise in debt reflects persistent government deficits, high volatility of oil revenues, long-term decline in real extractive GDP, and rising current public expenditures (which began to rise significantly after 2015). - Interest costs remained relatively low, reflecting the non-marketable nature of the debt. - Structural deficit (cyclically adjusted primary deficit) generally followed the trajectory of government debt. - Structural fiscal deficit ties to the non-extractive output gap per equation (3); relationship illustrated at annual frequency. - Fiscal behavior: - Rising deficits in response to adverse shocks indicate countercyclical behavior in certain periods. - In other years, fiscal policy exhibited pro-cyclical tendencies, reflecting challenges of countercyclical management in an economy reliant on volatile hydrocarbon revenues. ### Monetary policy: money targeting and exchange rate - Bank of Algeria (BA) influences real activity and inflation via monetary conditions: real interest rate gap and real exchange rate gap. - Trend real interest rate shaped by foreign real interest rate trend, country risk premium, and expected appreciation trend of the real exchange rate. - Real exchange rate trend: appreciation tendency before the late-2010s global financial crisis, followed by a period of real depreciation. - Signals: - Negative real interest rate gap and/or positive real exchange rate gap (indicating undervaluation) are stimulative and generate inflationary pressures. - Nominal exchange rate developments: - Driven largely by nominal depreciation of the dinar against the US dollar; several depreciation episodes since 2003 driven by global energy price fluctuations. - Notable nominal depreciations occurred in 2009 and 2015; in 2015 a 34 percent dinar depreciation contributed to inflationary pressures. - In 2023 the dinar appreciated by nearly 5 percent, supported by a recovery in hydrocarbon prices and current account surpluses. - Real interest rate gap history: - Below-trend real interest rate (negative gap) evident during 2014–17 (gradual nominal interbank rate reductions) and during the COVID-19 pandemic. - Money supply (M2) policy: - M2 growth follows a BA reaction function per equation (31), responding to expected deviation of annual core inflation from the target, the non-extractive output gap, and misalignment between actual M2 growth and its target, with a smoothing term tied to long-term fundamentals. - Decomposition of quarterly M2 growth shows the smoothing mechanism as the dominant factor; BA reacts to expected inflation deviations; frequent and important deviations from the reaction function identified historically. - M2 growth targets are set explicitly or implicitly and considered exogenous for the QPM; intra-year targets are determined by the model. - Exchange rate policy: - BA operates a managed floating regime with exporters required to convert FX receipts into dinar with the central bank, giving BA a quasi-monopolistic FX market position. - AMPM assumes Bank sets reference exchange rate per a nominal exchange rate rule (equation (40)) to eliminate misalignments, align with fundamentals, and reduce short-term volatility. - Decomposition shows BA adjusts exchange rate primarily in response to RER misalignments while avoiding excessive volatility; small contributions from exchange rate policy shocks occur during oil price fluctuations or large government spending changes. ### Inflation - BA prioritizes core inflation, which represents 83 percent of the consumption basket and excludes volatile fresh food prices; medium-term core inflation target of 4 percent. - Historical inflation dynamics: - 2003–05: inflation relatively low, averaging around 4 percent. - 2006–2009: inflationary pressures rose, peaking in 2009 due to sharp fresh food price increases. - 2012: inflation reached 9 percent driven by food and manufactured goods prices and excess liquidity from elevated public spending. - Late 2013–early 2014: brief period of low inflation. - 2015: inflation rose, spurred by higher import prices and a 34 percent dinar depreciation; surged to 6.4 percent in 2016 before moderating by 2019. - Initial COVID-19: inflation dropped to 2.4 percent reflecting moderation in core inflation and declining non-core prices. - 2021–2023: inflation accelerated significantly, exceeding 9 percent at its peak, driven by drought conditions, base effects, global supply chain disruptions, and heightened post-pandemic demand. - 2024: core inflation evolved close to or below the 4 percent target; recent pressures appear primarily short-term with non-core prices (fresh vegetables) rising faster than core prices. - Drivers of core inflation: - Positive association between core-specific real marginal costs and core inflation. - Real marginal costs decomposed into domestic demand (non-extractive output gap), foreign price effects (RER gap), and global commodity (oil and gas) prices. - Foreign prices and energy prices are primary drivers of real marginal cost variability and thus core inflation fluctuations. - Significant inflationary episodes in 2005–06, 2015–16, and 2022–23 were primarily driven by real exchange rate movements followed by increases in core inflation. *IMF Working Paper — Chapter 4.3: Model-Based Historical Interpretations (AMPM)* ### 4.4 In-Sample Simulations ### wpiea2026025-source-pdf - 4.4 In-Sample Simulations ### Purpose and methodology - In-sample recursive forecasts constructed conditionally on foreign variables (e.g., foreign economic indicators, oil and gas prices) and domestic targets (inflation and money growth targets). - Assumes perfect foresight of conditioned variables; excludes historical revisions, domestic trends, and expert judgments. - Recursive projections span eight quarters ahead, originating at each point in time from 2010Q1 to 2024Q3, and are compared against observed data. - Evaluation tools include impulse response functions, historical equation decompositions, in-sample simulations, RMSE statistics, and comparison versus a random walk (RW) benchmark. ### Main findings on fit and dynamics - General performance: - Forecast simulations for main macroeconomic variables closely follow observed data. - QPM-based simulations show no systematic deviations; forecast errors remain relatively small (in absolute value) and balanced over the simulation horizon in direction. - Real GDP: - Real GDP grew slightly below 5 percent on average over the sample period. - QPM predicts non-extractive growth relatively well. - Hydrocarbon sector introduces variability and biases, particularly between 2016 and 2020, due to commodity market volatility. - AMPM does not fully capture the pandemic-induced downturn and subsequent recovery. - Inflation: - Model-based headline inflation forecasts are slightly more volatile than actual data. - Discrepancy arises because inflation forecasts are derived from core inflation and the rest of the CPI basket (fresh food); fresh food exhibits higher forecast errors. - Integrating an exogenous outlook for the fresh food subcomponent could improve overall inflation data fit. - Monetary policy variables: - Nominal exchange rate forecasts align closely with actual data except during extraordinary depreciation events in 2015 and 2022. - Simulated M2 money growth is generally in line with observed data, reflecting a smoothing effect consistent with the policy reaction function and convergence to implicit money growth targets. - External and fiscal variables: - Model effectively replicates current account dynamics: captures deterioration starting 2012, gradual improvement starting 2016, and marked recovery in 2021-22. - Government debt forecasts, generated endogenously by the fiscal block, are unbiased with no systematic or persistent deviations from observed data. ### Numerical evaluation (RMSE and RMSE relative to RW) - Interpretation note: A ratio below 1 indicates AMPM outperforms the RW model for the given forecast horizon. RMSEs increase with forecast horizon; AMPM consistently outperforms RW for all variables across all horizons. - AMPM RMSE (by horizon) - GDP growth (% yoy): 1-q 1.44; 2-q 1.85; 3-q 2.04; 4-q 2.33; 5-q 2.42; 6-q 2.44; 7-q 2.35; 8-q 2.31 - CPI inflation (% yoy): 1-q 0.92; 2-q 1.42; 3-q 1.76; 4-q 2.23; 5-q 2.21; 6-q 2.20; 7-q 2.15; 8-q 2.12 - Exchange rate (DZD per USD): 1-q 2.01; 2-q 3.16; 3-q 3.94; 4-q 4.51; 5-q 4.87; 6-q 5.06; 7-q 5.11; 8-q 5.13 - M2 growth (% yoy): 1-q 1.20; 2-q 2.11; 3-q 2.98; 4-q 3.74; 5-q 3.87; 6-q 3.91; 7-q 3.87; 8-q 3.79 - Current account balance (% of GDP): 1-q 2.77; 2-q 3.50; 3-q 3.88; 4-q 4.05; 5-q 4.12; 6-q 4.17; 7-q 4.18; 8-q 4.21 - Government debt (% of GDP): 1-q 2.59; 2-q 3.17; 3-q 3.87; 4-q 4.38; 5-q 5.05; 6-q 5.73; 7-q 6.49; 8-q 7.29 - AMPM RMSE relative to RW RMSE (by horizon) - GDP growth (% yoy): 1-q 0.68; 2-q 0.62; 3-q 0.61; 4-q 0.62; 5-q 0.66; 6-q 0.70; 7-q 0.65; 8-q 0.63 - CPI inflation (% yoy): 1-q 0.68; 2-q 0.70; 3-q 0.67; 4-q 0.70; 5-q 0.64; 6-q 0.61; 7-q 0.57; 8-q 0.56 - Exchange rate (DZD per USD): 1-q 0.80; 2-q 0.73; 3-q 0.66; 4-q 0.60; 5-q 0.55; 6-q 0.50; 7-q 0.46; 8-q 0.42 - M2 growth (% yoy): 1-q 0.55; 2-q 0.57; 3-q 0.60; 4-q 0.63; 5-q 0.59; 6-q 0.56; 7-q 0.53; 8-q 0.52 - Current account balance (% of GDP): 1-q 0.93; 2-q 0.77; 3-q 0.69; 4-q 0.63; 5-q 0.57; 6-q 0.52; 7-q 0.48; 8-q 0.44 - Government debt (% of GDP): 1-q 0.82; 2-q 0.78; 3-q 0.74; 4-q 0.75; 5-q 0.74; 6-q 0.75; 7-q 0.76; 8-q 0.76 ### Implications and ongoing work - Usefulness: - AMPM is useful for medium-term forecasts, notably for the nominal exchange rate, current account ratio, and debt ratio where RMSE relative to RW declines further into the future. - For inflation and money growth, RMSE ratios remain broadly stable across horizons. - Planned enhancements: - Regular ex-post (out-of-sample) evaluations to compare historical forecasts with realized data to identify biases and improve forecast approaches. - Decompositions of forecast changes to understand factors driving forecast adjustments round-by-round and to strengthen policy advice robustness. - Consider integrating an exogenous outlook for the fresh food CPI subcomponent to improve inflation fit. *Source: 4.4 In-Sample Simulations (wpiea2026025-source-pdf)* ### 5. Conclusion ### 5. Conclusion ### Modernization and FPAS development - The Banque d’Algérie (BA) is modernizing its monetary policy framework with IMF technical assistance (TA), developing a Forecasting and Policy Analysis System (FPAS) centered on Algeria Macroeconomic Projection Model (AMPM). - The advanced analytical framework supports forward-looking policymaking at the BA by accurately capturing key features of the Algerian economy, including: - the monetary-fiscal policy landscape; - the managed floating exchange rate arrangement; and - exposure to global energy price fluctuations. ### AMPM structure and core features - AMPM is built on a New-Keynesian semi-structural trend-gap framework and incorporates: - core domestic blocks and an exogenous external block; - disaggregation of real GDP into extractive and non-extractive components, emphasizing the hydrocarbon sector’s importance for aggregate activity; - decomposition of CPI into core and non-core components to understand sectoral price dynamics and interlinkages; - a fiscal sector representation driven by global energy prices, with fiscal balance and domestic public debt decomposed into marketable and non-marketable components; - a balance of payments block representing current account and capital and financial account, with the trade balance driven primarily by oil and gas exports. - The comprehensive structure captures weak policy transmission due to shallow financial markets and the dominance of terms of trade and supply shocks, enabling simulation of different policy scenarios and regular macroeconomic projections. ### Economic policy representation in the model - Monetary policy: - The model specifies a forward-looking monetary policy reaction function for money supply (M2). - The BA uses various instruments to achieve price stability, with broad money supply as the main intermediate target. - The BA assumes an active exchange rate policy, guiding the nominal exchange rate to close real exchange rate deviations from the medium-run trend. - Fiscal policy: - The fiscal block anchors public debt accumulation through a fiscal reaction function for the structural deficit. - External sector: - The external balance incorporates current account dynamics heavily impacted by hydrocarbon exports, changes in reserve-related financing in the absence of large private financial flows, and corresponding FX reserves accumulation. ### Model validation, IRFs, and historical interpretation - Impulse Response Functions (IRFs): - IRFs illustrate how endogenous variables react to temporary structural shocks, providing insights into the model’s transmission mechanisms. - Findings from IRFs generally align with theoretical expectations, demonstrating model robustness in capturing key interlinkages and policy responses. - Examples from IRFs: - An increase in global energy prices leads to an expansion in the extractive sector, positive spillovers to non-extractive activities, an improved current account balance, and dinar appreciation. - A fiscal expansion increases non-extractive output and price pressures, prompting the BA to adopt a tighter M2 stance. - Model-based historical interpretations: - Using observed data and Kalman filtration, AMPM provides a comprehensive narrative of Algeria’s recent macroeconomic developments. - These interpretations highlight key drivers behind historical evolutions and assess specification, calibration, and theoretical consistency. - Findings are cross-referenced with historical data and stylized facts to ensure accuracy and relevance. ### Strengths, limitations, and potential improvements - Strengths: - AMPM is a robust tool for policy analysis and forecasting, capturing the current key features of the economy and policy framework. - It offers valuable insights into transmission channels and the effects of monetary policy actions, fiscal policy stance, inflationary pressures, and external position. - Limitations and challenges: - The model faced challenges in accurately capturing the pandemic-induced downturn and subsequent recovery. - It exhibited volatility in inflation forecasts. - It showed limitations in hydrocarbon sector forecasts due to global commodity market volatility. - Potential improvement: - Integrating an exogenous outlook for fresh food prices could enhance headline inflation accuracy. ### Operational use, adaptability, and complementary efforts - AMPM as described reflects main theoretical and empirical findings at the time of publication; BA staff operational use involves incorporation of complementary assessments and judgments during regular forecasting rounds to ensure relevance and performance. - AMPM can be adapted to policymakers’ demands under the existing regime and a prospective modernized one. - Complementary FPAS development efforts include: - strengthening internal communications among relevant BA departments for fully consistent analytical work in support of policy processes; - regular interactions with BA’s decision-making body; and - enriching external communications to support forward-looking policy formulation. *Source: https://www.imf.org/-/media/files/publications/wp/2026/english/wpiea2026025-source-pdf.pdf* ### 0.4 Non-extractive GDP gap lag term Iterative adjustment ### 0.4 Non-extractive GDP gap lag term Iterative adjustment ### Real economy — key parameters and calibration - b2_y_ne: 0.4 — Non-extractive GDP gap lag term — Iterative adjustment - b3_y_ne: 0.2 — Monetary conditions elasticity — Iterative adjustment - b4_y_ne: 0.1 — Trade balance gap elasticity — Iterative adjustment - c1_y_e: 0.3 — Fiscal impulse elasticity — Iterative adjustment - c2_y_e: 0.6 — Extractive GDP gap lag term — Iterative adjustment - w_y_e: 0.05 — Energy price elasticity — Iterative adjustment - ρ_Δy_e: 0.1 — Weight of extractive GDP — Historical data - ρ_Δy_ne: 0.8 — Persistence of extractive GDP trend growth — Iterative adjustment - ρ_Δny_ne: 0.85 — Persistence of non-extractive GDP trend growth — Iterative adjustment - Δy̅_ss_e: 0 — Persistence of non-extractive nominal GDP trend growth — Iterative adjustment — Steady States - Δy̅_ss_ne: 0.5 — Extractive GDP trend growth — Historical data - Δny̅_ss_ne: 4.5 — Non-extractive GDP trend growth — Historical data - Non-extractive nominal GDP trend growth: 8 — Historical data ### Aggregate Supply — code, value, parameter role, calibration choice and justification - w_p_c: 0.83082 — Weight of core items — ONS data - c1_π_c: 0.5 — Core inflation lead term — Iterative adjustment - c2_π_c: 0.2 — RMC elasticity of core inflation — Iterative adjustment - c1_rmc_c: 0.4 — Weight of extractive GDP gap in core RMC — Iterative adjustment - c2_rmc_c: 0.035 — Weight of energy prices in core RMC — Iterative adjustment - c1_π_nc: 0.5 — Non-core inflation lead term — Iterative adjustment - c2_π_nc: 0.1 — RMC elasticity of non-core inflation — Iterative adjustment - c1_rmc_nc: 0.5 — Weight of extractive GDP gap in non-core RMC — Iterative adjustment - c2_rmc_nc: 0.025 — Weight of energy prices in non-core RMC — Iterative adjustment - ρ_rp: 0.7 — Relative core price trend growth lag — Iterative adjustment — Steady State - Δrp_ss: -2.5 — Relative core price growth — Historical data ### Balance of Payments — code, value, parameter role, calibration choice and justification - c1_tb̂_rat: 0.6 — Trade balance gap lag term — Iterative adjustment - c2_tb̂_rat: 0.4 — Energy price elasticity — Iterative adjustment - c3_tb̂_rat: 1 — Foreign output gap elasticity — Iterative adjustment - c4_tb̂_rat: 3 — Non-extractive GDP gap elasticity — Iterative adjustment - c5_tb̂_rat: 0.3 — Extractive GDP gap elasticity — Iterative adjustment - c6_tb̂_rat: 0.6 — Real exchange rate gap elasticity — Iterative adjustment - c7_tb̂_rat: 1 — Deviation of structural primary deficit elasticity — Iterative adjustment - c1_tb̅_rat: 0.95 — Trade balance trend lag term — Iterative adjustment - c2_tb̅_rat: 1 — Energy prices trend elasticity — Iterative adjustment - c1_ca_pother,rat: 0.25 — Other current account items lag term — Iterative adjustment - c1_kfa_nonres,rat: 0.75 — Endogenous non-reserve financial flows lag term — Iterative adjustment - c2_kfa_nonres,rat: 0.1 — Risk premium elasticity — Iterative adjustment - tb̅_ss_rat: -2 — Trade balance — Historical data — Steady States - ca_pother,rat_ss: -0.6 — Other current account items — Historical data — Steady States - kfa_nonres_ss_rat: 0.4 — Endogenous non-reserve financial inflows — Historical data — Steady States ### Monetary Policy — code, value, parameter role, calibration choice and justification - c1_m: 0.6 — Money growth lag term — Iterative adjustment - c2_m: 0.5 — Deviation of interbank rate from neutral rate — Iterative adjustment - c3_m: 1.2 — Deviation of core inflation from target — Iterative adjustment - c4_m: 0.6 — Output gap — Iterative adjustment - c5_m: 0.6 — Deviation of money growth from target — Iterative adjustment - c6_m: 0.4 — Change in (non-sterilized) FX reserves — Iterative adjustment - c1_mt ar: 0.7 — Nominal money growth target lag term — Iterative adjustment - c2_mtar: 0.2 — Deviation of real money target from trend — Iterative adjustment - c1_rm: 0.7 — Real money demand gap lag term — Iterative adjustment - c2_rm: 2.5 — Real interest rate gap elasticity — Iterative adjustment - c3_rm: 1.1 — Output gap elasticity — Iterative adjustment - c4_rm: 0.2 — Debt deviation from target elasticity — Iterative adjustment - c5_rm: 0.1 — Change in reserves elasticity — Iterative adjustment - c1_Δrm̅: 1.2 — Change in the real interest rate trend elasticity — Iterative adjustment - c1_mci: 0.6 — Weight of real interest rate gap in RMCI — Iterative adjustment - ρ_r: 0.7 — Persistence real interest rate trend — Iterative adjustment - ρ_π_t4,tar: 0.99 — Persistence core inflation target — Iterative adjustment - ρ_v: 0.9 — Persistence money velocity growth — Iterative adjustment - π_ss4,tar: 4 — Core inflation target — BA strategy — Steady States - r̅_ss: -1.5 — Real interest rate trend — Historical data — Steady States ### Exchange Rate — code, value, parameter role, calibration choice and justification - c1_s: 0.75 — UIP weight of BA reference rate — Iterative adjustment - c2_s: 0.1 — Forward-looking expectations — Iterative adjustment - c3_s: 0.1 — Trade balance gap elasticity — Iterative adjustment - c4_s: 0.06 — Debt deviation from target elasticity — Iterative adjustment - c1_spol: 0.75 — Policy exchange rate depreciation lag term — Iterative adjustment - c2_spol: 0.6 — Real exchange rate gap elasticity — Iterative adjustment - c3_spol: 0.2 — Inflation deviation from target elasticity — Iterative adjustment - ρ_Δz: 0.9 — Persistence real exchange rate trend growth — Iterative adjustment - Δz̅_ss: 0 — Real exchange rate trend growth — Historical data — Steady State ### Fiscal policy — code, value, parameter role, calibration choice and justification - c1_pdef_rat,cyc: 0.1 — Non-extractive output gap cyclical elasticity — Iterative adjustment - c2_pdef_rat,cyc: 0.035 — Energy price cyclical elasticity — Iterative adjustment - ρ_def: 0.5 — Smoothing of fiscal reaction function — Iterative adjustment - c1_pdef_rat,struc: 0.2 — Non-extractive output gap elasticity of fiscal policy — Iterative adjustment - c2_pdef_rat,struct: 0.4 — Debt deviation elasticity of fiscal policy — Iterative adjustment - c1_fimp_rat: 0.75 — Fiscal impulse weight on the deficit change — Iterative adjustment - c1_tprem: 0.01 — Debt deviation elasticity of term premia — Iterative adjustment - ρ_debt: 0.99 — Public debt target lag term — Iterative adjustment - ρ_θ: 0.9 — Share of marketable financing lag term — Iterative adjustment - ρ_restruct: 0.7 — Restructuring of outstanding debt lag term — Iterative adjustment - ρ_tprem: 0.75 — 1-year term premium lag term — Iterative adjustment - debt_rat_ss: 48 — Public debt target — Expert judgement — Steady States - θ_ss: 1 — Share of marketable financing — Expert judgement — Steady States - prem_m,ss: 1.75 — 1-year term premium for marketable debt — Historical data — Steady States - prem_nm,ss: 0 — 1-year term premium for non-marketable debt — Historical data — Steady States - i4_nm,ss: 1 — 1-year non-marketable debt interest rate — Expert judgement — Steady States ### Foreign Economy — code, value, parameter role, calibration choice and justification - ρ_zen̅: 0.95 — Relative price of energy trend lag term — Iterative adjustment - ρ_zen̂: 0.75 — Relative price of energy gap lag term — Iterative adjustment - ρ_π_rw: 0.25 — US inflation lag term — Iterative adjustment - ρ_rs_rw: 0.9 — US interest rate lag term — Iterative adjustment - ρ_y_rw: 0.85 — US output gap lag term — Iterative adjustment - ρ_r_rw: 0.95 — US real interest rate lag term — Iterative adjustment - r_ss_rw̅: 0 — US real interest rate — Historical data — Steady State - π_ss_rw̅: 2 — US inflation — Fed strategy — Steady State - Δzen_ss̅̅̅̅̅: 0 — Relative price of energy growth — Historical data — Steady State *IMF WORKING PAPERS Algeria Macroeconomic Projection Model (AMPM) Working Paper No. WP/2026/25 — INTERNATIONAL MONETARY FUND* --- _Source: https://www.imf.org/-/media/files/publications/wp/2026/english/wpiea2026025-source-pdf.pdf_