## Annex 1. Solving the Quarterly Projection Model Using EViews 13 68

## Source details

**Canonical URL:** [Annex 1. Solving the Quarterly Projection Model Using EViews 13 68](https://www.imf.org/-/media/files/publications/tnm/2026/english/tnmea2026003.pdf)

## Other formats

- [Markdown version](/-/media/files/publications/tnm/2026/english/tnmea2026003.pdf.md)
- [Structured JSON version](/-/media/files/publications/tnm/2026/english/tnmea2026003.pdf.json)

---

### Model description and role
- The QPM is a semistructural New Keynesian gap model designed for monetary policy analysis and forward-looking monetary policy formulation (FPAS).
- Core structure: four behavioral equations — aggregate demand (output gap), Phillips curve (inflation), uncovered interest parity (exchange rate), and a Taylor rule (policy rate).
- Model features:
  - Gap framework for output, inflation, real interest rate, and real exchange rate.
  - Monetary policy neutral in the long term (policy affects paths around equilibrium but not medium-term output level).
  - Monetary conditions index (mci_t) combines deviations of the real interest rate from its neutral rate, optionally a credit premium (cr_prem_t), and real effective exchange rate gap (z^t).
  - Forward-looking Phillips curve with real marginal costs (rmc_t) combining output gap and real exchange rate gap; includes food and energy subcomponents.
  - UIP includes an extrapolative (backward-looking) component controlled by parameter e1.
  - Taylor rule sets nominal policy rate it as a function of persistence, expected inflation deviations from the target, and output gap; neutral nominal rate equals trend real rate plus model-consistent inflation expectations.

### Canonical calibration (Country Z example) — steady states and key parameters
- Steady-state targets and long-term parameters:
  - Potential output growth: 2.5 percent.
  - Domestic inflation target: 2 percent.
  - Foreign inflation target (euro area): 2 percent.
  - Trend real exchange rate appreciation: 1.5 percent per year.
  - Domestic real neutral interest rate: 2 percent (implies nominal neutral rate of 4 percent).
  - Foreign real neutral interest rate: 0.75 percent (implies nominal neutral rate of 2.75 percent).
- Selected parameter values (as calibrated for Country Z):
  - Aggregate demand: b1 = 0.7; b2 = 0.2; b3 = 0.3; b4 = 0.7; b5 = 0.7.
  - Phillips curve (core): a1 = 0.5; a2 = 0.25; a3 = 0.65.
  - Food subcomponent: a21 = 0.50; a22 = 0 .10; a23 = 0.25.
  - Energy subcomponent: a31 = 0.50; a32 = 0.05; a33 = 1.0.
  - Taylor rule: g1 = 0.75; g2 = 0.25; g3 = 0.25.
  - Imperfect control over money market: h1 = 0.0.
  - UIP: e1 = 0.20; degree of control over exchange rate h2 = 0.0.
  - Exchange rate target: f1 = 0.50; f2 = 0.0; f3 = 0.0.
- Steady-state target series and values:
  - π_Tt (inflation target q-on-q) = 2.0
  - r̄_t (domestic real neutral rate) = 0.5
  - Δz̄_t (annualized trend real exchange rate change) = −1. 5
  - Δȳ_t (annualized potential output change) = 2.5
  - π̄*_t (rest of the world target inflation) = 2.0
  - r̄*_t (rest of the world target real interest rate) = 0.75
  - log(p_tfood) − log(p_t) = 0.0
  - log(p_tcore) − log(p_t) = 0.0
  - π*_t (rest of the world annualized inflation) = 2.0
  - π_toil (annualized world oil inflation) = 5.0
- Speed-of-adjustment parameters (to steady state):
  - ρ1 = 0.75; ρ2 = 0.75; ρ3 = 0.5; ρ4 = 0.8; ρ5 = 0.8; ρ6 = 0.8; ρ7 = 0.8; ρ8 = 0.8; ρ9 = 0.8; ρ10 = 0.8; ρ11 = 0.25; ρ12 = 0.5; ρ13 = 0.9; ρ14 = 0.5; ρ15 = 0.9; ρ16 = 0.5.

### Calibration assessment and impulse responses
- Suitability of calibration is assessed by generating impulse responses (one-period unitary shock = 1 percent) to structural errors and comparing with country stylized facts.
- Standard shocks used for assessment: nominal exchange rate shock, aggregate demand shock, cost-push shock, monetary shock.
- Impulse responses in the package are produced by introducing one-period unitary shocks to specific structural errors (shk_ variables) and solving the model with EViews solver to compare alternative solutions vs baseline.

### Practical EViews implementation: seven-step workflow (minimum steps for baseline)
- Step 1 — Data preparation:
  - Raw data file: data.csv with fixed short names (GDP, CPI_U, CPIF_U, CPIE_U, CPIXFE_U, S, RS, GDP_RW, CPI_RW, D4L_CPI_TAR, WFOOD_U, WOIL_U, S_CROSS, w_CPIF, w_CPIE).
  - Example sample: first observation 1996:1; last observation 2022:3; projection start 2022:4; projection end 2025:4 (13 periods).
- Step 2 — Import into EViews workfile and rename page file to Country_Z.
- Step 3 — Edit driver program 0_qpm.prg:
  - Set parameters: %source_dir (page file with raw data, e.g., "Country_Z"), %baseline tag (max three characters, e.g., "cz"), %target_dir (page file to store solution, e.g., temp_ev).
  - Set %sstart, %slast, %simstart, %simend values (e.g., %sstart = 1996:1; %slast = 2022:3; %simstart = 2022:4; %simend = 2025:4).
- Step 4 — Run 1_makedata.prg to initialize endogenous and exogenous variables, deseasonalize as needed, compute potential (_bar) and gap series via Hodrick–Prescott filter.
- Step 5 — Run 2_definemodel.prg to create model object fpas_b (83 equations, mostly identities) and set scalars to calibrated values.
  - Note: 26 shock variables (shk_*) defined and set to zero by default for baseline.
  - Adjust UIP forward/backward (e1 between 0 and 1; 0 = completely forward-looking, 1 = completely backward-looking).
  - Modify monetary policy rule as needed (default: central bank full control and forward-looking expectations, default horizon three periods ahead).
- Step 6 — Run 3_solve_model.prg to compute baseline solution using EViews solver (E-QNewton recommended). Initial values for future-dated variables are set (mostly to steady-state) to allow iterative convergence.
  - Baseline solution series are saved with suffix based on %baseline (e.g., l_gdp_cz).
- Step 7 — Create alternative scenarios by assigning non-zero values to shk_* variables for specified dates and re-running solver; alternative solution saved with %Alternative suffix (e.g., _czr).

### Example scenarios and outputs (illustrative Country Z exercises)
- Simple monetary policy shock:
  - Four-period 100 basis point increase in policy rate starting 2022:4 (implemented by setting shk_rs_unc for 2022:4–2023:3).
  - Baseline suffix: "cz"; alternative suffix: "czr".
  - Observed effects: alternative (czr) vs baseline (cz) — inflation falls with a lag; nominal exchange rate appreciates; monetary conditions index increases; output gap more negative.
  - Comparison objects: COMPARE_CZR (plots baseline vs alternative for nine key variables) and DIFFERENCES_CZR (plots alternative − baseline).
- Constrained baseline example (7_policy_shock_baseline.prg):
  - Foreign inflation (DLA_CPI_RW), foreign interest rate (RS_RW), and foreign output gap (L_GDP_RW_GAP) constrained during 2022:4–2024:3.
  - Domestic inflation components constrained for 2022:4: core = 13.7 percent; food = 15.6 percent; energy = 22.0 percent.
  - Constrained solution tag: "_cza". Comparing _cza vs _cz highlights the impact of constraints on domestic and foreign variables.
- Confidence shock example (7_policy_shock_baseline_con.prg):
  - Negative shock to real GDP gap (l_gdp_gap) of −2.5 percent each period during 2022:4–2023:3 (first forecast year) relative to cza baseline.
  - Alternative tag: "con". Effects: lower inflation and nominal interest rate relative to cza; nominal exchange rate depreciates.
- Oil shock example (7_policy_shock_baseline_oil.prg):
  - Global oil prices increase by 10 percent relative to cza baseline during 2022:4–2023:3; foreign output gap declines by 0.25 percent over same period.
  - Solver excludes DLA_WOIL and L_GDP_RW_GAP for 2022:4–2023:3 so those paths are fixed; resultant domestic output gap becomes more negative relative to cza.
- Example decomposition result (illustrative):
  - Core inflation projected at 8.3 percent in 2022:4 with contributions: persistence 6.7 percentage points; expectations 2.4 percentage points; imported real marginal costs −0.8 percentage points.

### EViews package utilities and outputs
- Driver: 0_qpm.prg (initialization and run control).
- Core programs:
  - 1_makedata.prg: data processing, deseasonalization, potential/gap computation.
  - 2_definemodel.prg: parameter assignment, builds model object fpas_b (83 equations).
  - 3_solve_model.prg: initializes future-dated variables and solves model (E-QNewton).
- Utility programs:
  - 4_stylized_facts.prg: produces stylized facts graphs (stylized_facts_p1 through stylized_facts_p7).
  - 5_produce_forecast_graphs.prg: creates forecasts_{%Target} graph (six key variables).
  - 6_generate_data_for_forecast_tables.prg: assembles table_data_{%Target} group for Excel template QPM_Output.xlsx.
  - 8_model_properties.prg: generates impulse responses across parameter combinations (e.g., varying b1 and b2).
  - 9_impulse_response_graphs.prg and 10_compare.prg: impulse response graphs and scenarios comparison tables/graphs (COMPARE_*, DIFFERENCES_*).
- Output workflow for Excel reporting:
  - Use table_data_{%Target} group (e.g., table_data_cz), copy to clipboard, paste into QPM_Output.xlsx Data worksheet (Paste Special → Text).
  - QPM_Output.xlsx contains worksheets including Main Indicators, Inflation Decomposition, Demand and Supply, Policy Decomposition, Foreign Variables.

### Solver, numerical notes, and series conventions
- EViews solving options:
  - E-QNewton used in manual (recommended for robustness). Other methods available: Fair-Taylor, E-Newton.
  - Initial guesses for forward-looking variables set (mostly to steady-state) across solve horizon to enable iterative convergence.
- Series naming conventions:
  - l_x prefix = 100 * log(x)
  - d4l_x = log(x) − log(x(−4)) (annual growth rate via four-period log difference)
  - dla_x = 4 * log(x) − log(x(−1)) (annualized one-period difference)
  - _bar suffix = potential/smoothed value
  - _gap suffix = deviation from potential
  - Target and shock series follow naming conventions shown in Table 3.

### EViews 13 specifics and MCE package
- EViews 13 cannot solve models with forward-looking variables directly.
- Required package: “MCE_SOLVE_LIBRARY” (FRB/US MCE_SOLVE_LIBRARY routines for EViews).
  - Purpose: Enables model-consistent expectations solution for forward-looking variables in EViews 13.
  - The EV13 folder in the EViews package for this manual contains the code to solve the QPM using MCE.
  - Implementation note: Line numbers in the MCE/EV13 code do not correspond to those referenced in specific programs in the technical manual, although the structure and steps are very similar.
- Contrast with EViews 14: EViews 14 can solve models with forward-looking variables directly; EViews 13 requires the MCE package and additional commands to invoke the FRB MCE.

### Practical recommendations and best practices (implicit)
- Calibrate the model to match the country’s stylized facts; validate calibration via impulse responses to standard shocks.
- Use historical data through the end of the latest available period; set projection horizon explicitly (%simstart and %simend).
- Keep baseline shocks at zero; introduce non-zero values to shk_* variables only for alternative scenarios and revert to zero afterwards.
- Use E-QNewton solver for robustness, and provide reasonable initial conditions for forward-looking variables.
- Use the provided EViews program suite (0_qpm.prg through 10_compare.prg) and the Excel template (QPM_Output.xlsx) to generate forecasts, decompositions, and comparison graphs.

*Source: ANNEX 1. Solving the Quarterly Projection Model Using EViews 13 (tnmea2026003).*

### Annex 1. Solving the Quarterly Projection Model Using EViews 13 68

### Annex 1. Solving the Quarterly Projection Model Using EViews 13 68

### Document unit
- Title: Annex 1. Solving the Quarterly Projection Model Using EViews 13
- Page: 68
- Source PDF filename: tnmea2026003 - Annex 1. Solving the Quarterly Projection Model Using EViews 13 68

*https://www.imf.org/-/media/files/publications/tnm/2026/english/tnmea2026003.pdf*

### References

### Solving the Canonical Quarterly Projection Model Using EViews (References chapter)

### Model description and role
- The QPM is a semistructural New Keynesian gap model designed for monetary policy analysis and forward-looking monetary policy formulation (FPAS).
- Core structure: four behavioral equations — aggregate demand (output gap), Phillips curve (inflation), uncovered interest parity (exchange rate), and a Taylor rule (policy rate).
- Model features:
  - Gap framework for output, inflation, real interest rate, and real exchange rate.
  - Monetary policy neutral in the long term (policy affects paths around equilibrium but not medium-term output level).
  - Monetary conditions index (mci_t) combines deviations of the real interest rate from its neutral rate, optionally a credit premium (cr_prem_t), and real effective exchange rate gap (z^t).
  - Forward-looking Philips curve with real marginal costs (rmc_t) combining output gap and real exchange rate gap; includes food and energy subcomponents.
  - UIP includes an extrapolative (backward-looking) component controlled by parameter e1.
  - Taylor rule sets nominal policy rate it as a function of persistence, expected inflation deviations from the target, and output gap; neutral nominal rate equals trend real rate plus model-consistent inflation expectations.

### Canonical calibration (Country Z example) — steady states and key parameters
- Steady-state targets and long-term parameters:
  - Potential output growth: 2.5 percent.
  - Domestic inflation target: 2 percent.
  - Foreign inflation target (euro area): 2 percent.
  - Trend real exchange rate appreciation: 1.5 percent per year.
  - Domestic real neutral interest rate: 2 percent (implies nominal neutral rate of 4 percent).
  - Foreign real neutral interest rate: 0.75 percent (implies nominal neutral rate of 2.75 percent).
- Selected parameter values (as calibrated for Country Z):
  - Aggregate demand: b1 = 0.7; b2 = 0.2; b3 = 0.3; b4 = 0.7; b5 = 0.7.
  - Phillips curve (core): a1 = 0.5; a2 = 0.25; a3 = 0.65.
  - Food subcomponent: a21 = 0.50; a22 = 0 .10; a23 = 0.25.
  - Energy subcomponent: a31 = 0.50; a32 = 0.05; a33 = 1.0.
  - Taylor rule: g1 = 0.75; g2 = 0.25; g3 = 0.25.
  - Imperfect control over money market: h1 = 0.0.
  - UIP: e1 = 0.20; degree of control over exchange rate h2 = 0.0.
  - Exchange rate target: f1 = 0.50; f2 = 0.0; f3 = 0.0.
- Steady-state target series and values:
  - π_Tt (inflation target q-on-q) = 2.0
  - r̄_t (domestic real neutral rate) = 0.5
  - Δz̄_t (annualized trend real exchange rate change) = −1. 5
  - Δȳ_t (annualized potential output change) = 2.5
  - π̄*_t (rest of the world target inflation) = 2.0
  - r̄*_t (rest of the world target real interest rate) = 0.75
  - log(p_tfood) − log(p_t) = 0.0
  - log(p_tcore) − log(p_t) = 0.0
  - π*_t (rest of the world annualized inflation) = 2.0
  - π_toil (annualized world oil inflation) = 5.0
- Speed-of-adjustment parameters (to steady state):
  - ρ1 = 0.75; ρ2 = 0.75; ρ3 = 0.5; ρ4 = 0.8; ρ5 = 0.8; ρ6 = 0.8; ρ7 = 0.8; ρ8 = 0.8; ρ9 = 0.8; ρ10 = 0.8; ρ11 = 0.25; ρ12 = 0.5; ρ13 = 0.9; ρ14 = 0.5; ρ15 = 0.9; ρ16 = 0.5.

### Calibration assessment and impulse responses
- Suitability of calibration is assessed by generating impulse responses (one-period unitary shock = 1 percent) to structural errors and comparing with country stylized facts (responses to nominal exchange rate shock, aggregate demand shock, cost-push shock, monetary shock).
- Impulse responses in the package are produced by introducing one-period unitary shocks to specific structural errors (shk_ variables) and solving the model with EViews solver to compare alternative solutions vs baseline.

### Practical EViews implementation: seven-step workflow (minimum steps for baseline)
- Step 1 — Data preparation:
  - Raw data file: data.csv with fixed short names (GDP, CPI_U, CPIF_U, CPIE_U, CPIXFE_U, S, RS, GDP_RW, CPI_RW, D4L_CPI_TAR, WFOOD_U, WOIL_U, S_CROSS, w_CPIF, w_CPIE).
  - Example sample: first observation 1996:1; last observation 2022:3; projection start 2022:4; projection end 2025:4 (13 periods).
- Step 2 — Import into EViews workfile and rename page file to Country_Z.
- Step 3 — Edit driver program 0_qpm.prg:
  - Set parameters: %source_dir (page file with raw data, e.g., "Country_Z"), %baseline tag (max three characters, e.g., "cz"), %target_dir (page file to store solution, e.g., temp_ev).
  - Set %sstart, %slast, %simstart, %simend values (e.g., %sstart = 1996:1; %slast = 2022:3; %simstart = 2022:4; %simend = 2025:4).
- Step 4 — Run 1_makedata.prg to initialize endogenous and exogenous variables, deseasonalize as needed, compute potential (_bar) and gap series via Hodrick–Prescott filter.
- Step 5 — Run 2_definemodel.prg to create model object fpas_b (83 equations, mostly identities) and set scalars to calibrated values (see Table 1 values above).
  - Note: 26 shock variables (shk_*) defined and set to zero by default for baseline.
  - Adjust UIP forward/backward (e1 between 0 and 1; 0 = completely forward-looking, 1 = completely backward-looking).
  - Modify monetary policy rule as needed (default: central bank full control and forward-looking expectations, default horizon three periods ahead).
- Step 6 — Run 3_solve_model.prg to compute baseline solution using EViews solver (E-QNewton recommended). Initial values for future-dated variables are set (mostly to steady-state) to allow iterative convergence.
  - Baseline solution series are saved with suffix based on %baseline (e.g., l_gdp_cz).
- Step 7 — Create alternative scenarios by assigning non-zero values to shk_* variables for specified dates and re-running solver; alternative solution saved with %Alternative suffix (e.g., _czr).

### Example scenarios and outputs (illustrative Country Z exercises)
- Simple monetary policy shock:
  - Four-period 100 basis point increase in policy rate starting 2022:4 (implemented by setting shk_rs_unc for 2022:4–2023:3).
  - Baseline suffix: "cz"; alternative suffix: "czr".
  - Observed effects: alternative (czr) vs baseline (cz) — inflation falls with a lag; nominal exchange rate appreciates; monetary conditions index increases; output gap more negative.
  - Comparison objects: COMPARE_CZR (plots baseline vs alternative for nine key variables) and DIFFERENCES_CZR (plots alternative − baseline).
- Constrained baseline example (7_policy_shock_baseline.prg):
  - Foreign inflation (DLA_CPI_RW), foreign interest rate (RS_RW), and foreign output gap (L_GDP_RW_GAP) constrained during 2022:4–2024:3.
  - Domestic inflation components constrained for 2022:4: core = 13.7 percent; food = 15.6 percent; energy = 22.0 percent.
  - Constrained solution tag: "_cza". Comparing _cza vs _cz highlights the impact of constraints on domestic and foreign variables.
- Confidence shock example (7_policy_shock_baseline_con.prg):
  - Negative shock to real GDP gap (l_gdp_gap) of −2.5 percent each period during 2022:4–2023:3 (first forecast year) relative to cza baseline.
  - Alternative tag: "con". Effects: lower inflation and nominal interest rate relative to cza; nominal exchange rate depreciates.
- Oil shock example (7_policy_shock_baseline_oil.prg):
  - Global oil prices increase by 10 percent relative to cza baseline during 2022:4–2023:3; foreign output gap declines by 0.25 percent over same period.
  - Solver excludes DLA_WOIL and L_GDP_RW_GAP for 2022:4–2023:3 so those paths are fixed; resultant domestic output gap becomes more negative relative to cza.

### EViews package utilities and outputs
- Driver: 0_qpm.prg (initialization and run control).
- Core programs:
  - 1_makedata.prg: data processing, deseasonalization, potential/gap computation.
  - 2_definemodel.prg: parameter assignment, builds model object fpas_b (83 equations).
  - 3_solve_model.prg: initializes future-dated variables and solves model (E-QNewton).
- Utility programs:
  - 4_stylized_facts.prg: produces stylized facts graphs (stylized_facts_p1 through stylized_facts_p7).
  - 5_produce_forecast_graphs.prg: creates forecasts_{%Target} graph (six key variables).
  - 6_generate_data_for_forecast_tables.prg: assembles table_data_{%Target} group for Excel template QPM_Output.xlsx.
  - 8_model_properties.prg: generates impulse responses across parameter combinations (e.g., varying b1 and b2).
  - 9_impulse_response_graphs.prg and 10_compare.prg: impulse response graphs and scenarios comparison tables/graphs (COMPARE_*, DIFFERENCES_*).
- Output workflow for Excel reporting:
  - Use table_data_{%Target} group (e.g., table_data_cz), copy to clipboard, paste into QPM_Output.xlsx Data worksheet (Paste Special → Text).
  - QPM_Output.xlsx contains worksheets including Main Indicators, Inflation Decomposition, Demand and Supply, Policy Decomposition, Foreign Variables.
  - Example decomposition result: core inflation projected at 8.3 percent in 2022:4 with contributions: persistence 6.7 percentage points; expectations 2.4 percentage points; imported real marginal costs −0.8 percentage points (example from output description).

### Solver and numerical notes
- EViews solving options:
  - E-QNewton used in manual (recommended for robustness). Other methods available: Fair-Taylor, E-Newton.
  - Initial guesses for forward-looking variables set (mostly to steady-state) across solve horizon to enable iterative convergence.
- Series naming conventions:
  - l_x prefix = 100 * log(x)
  - d4l_x = log(x) − log(x(−4)) (annual growth rate via four-period log difference)
  - dla_x = 4 * log(x) − log(x(−1)) (annualized one-period difference)
  - _bar suffix = potential/smoothed value
  - _gap suffix = deviation from potential
  - Target and shock series follow naming conventions shown in Table 3.

### Practical recommendations for users (implicit best practices from manual)
- Calibrate the model to match the country’s stylized facts; validate calibration via impulse responses to standard shocks.
- Use historical data through the end of the latest available period; set projection horizon explicitly (%simstart and %simend).
- Keep baseline shocks at zero; introduce non-zero values to shk_* variables only for alternative scenarios and revert to zero afterwards.
- Use E-QNewton solver for robustness, and provide reasonable initial conditions for forward-looking variables.
- Use the provided EViews program suite (0_qpm.prg through 10_compare.prg) and the Excel template (QPM_Output.xlsx) to generate forecasts, decompositions, and comparison graphs.

*Source: Solving the Canonical Quarterly Projection Model Using EViews (Technical Note / Manual excerpt).*

### ANNEX 1. Solving the Quarterly Projection

### ANNEX 1. Solving the Quarterly Projection Model Using EViews 13

### Overview
- EViews 13 cannot solve models with forward-looking variables directly.
- To solve such models in EViews 13, the model-consistent expectations (MCE) package called “MCE_SOLVE_LIBRARY” (written by the Federal Reserve Board) must be downloaded and invoked.
- The EV13 folder included with the EViews package for this manual contains the code required to solve the quarterly projection model (QPM) using the MCE package.
- The overall approach to solving the QPM using EViews 13 and MCE is very similar to the approach described in the main manual, with additional commands to invoke the FRB MCE.

### Required package and invocation
- Package name: “MCE_SOLVE_LIBRARY” (FRB/US MCE_SOLVE_LIBRARY routines for EViews).
- Purpose: Enables model-consistent expectations solution for forward-looking variables in EViews 13.
- Note in text: “The FRB/US MCE_SOLVE_LIBRARY routines for EViews can be downloaded by clicking here.” (download instruction as presented in the source).

### EV13 folder and main programs
- The EV13 folder in the EViews package for this manual contains the code to solve the QPM using MCE.
- Main programs used to set up and solve the QPM:
  - 0_qpm.prg
  - 1_makedata.prg
  - 2_definemodel.prg
  - 3_solve_model
- Implementation note: Line numbers in the MCE/EV13 code do not correspond to those referenced in specific programs in the technical manual, although the structure and steps are very similar.

### Differences and practical notes
- Contrast with EViews 14: EViews 14 can solve models with forward-looking variables directly; EViews 13 requires the MCE package.
- The addition to the workflow for EViews 13 is the invocation of the FRB MCE routines; otherwise, the core QPM solution approach remains the same.

### References and identifiers
- Berg, A., P. D. Karam, and D. Laxton. 2006. “A Practical Model-Based Approach to Monetary Policy Analysis—Overview.” IMF Working Paper 06/80, International Monetary Fund, Washington, DC.
- Lindquist, W. 2023. “Monetary Policy Analysis Using the Quarterly Projections Model.” IMF Country Report 23/190, International Monetary Fund, European Department, Washington, DC. ht tps://doi.org/10.5089/9798400244612.002
- Numeric identifiers present in source:
  - 13
  - 14
  - 0_qpm.prg
  - 1_makedata.prg
  - 2_definemodel.prg
  - 3_solve_model
  - 06/80
  - 23/190
  - 035439798229
  - ISBN 9798229035439
  - TNM/2026/03

*Source: ANNEX 1. Solving the Quarterly Projection Model Using EViews 13 (tnmea2026003).*

---


_Source: https://www.imf.org/-/media/files/publications/tnm/2026/english/tnmea2026003.pdf_