## PREFACE — Thailand Technical Assistance on Macroeconomic Frameworks (tarea2025082-source-pdf)

## Source details

**Canonical URL:** [PREFACE — Thailand Technical Assistance on Macroeconomic Frameworks (tarea2025082-source-pdf)](https://www.imf.org/-/media/files/publications/tar/2025/english/tarea2025082-source-pdf.pdf)

## Other formats

- [Markdown version](/-/media/files/publications/tar/2025/english/tarea2025082-source-pdf.pdf.md)
- [Structured JSON version](/-/media/files/publications/tar/2025/english/tarea2025082-source-pdf.pdf.json)

---

### Project overview and objectives
- Capacity development project between ICD and the Ministry of Finance of Thailand initiated at the request of the Fiscal Policy Office (FPO).
- Objective: develop and operationalize a structural model (in particular, a dynamic stochastic general equilibrium model) for macroeconomic policy analysis between end-2024 and mid-2026.
- Dual goals:
  - Convey the project team assessment on the current toolkit, capacity, and institutional mechanisms to conduct fiscal policy and macrostructural analysis.
  - Register the action plan agreed with the authorities to guide technical assistance delivery and work commitments by the Fund and MOF.

### Scoping mission: timing, team, and activities
- Location and dates: Bangkok, Thailand, from November 18 to 26, 2024.
- Mission team:
  - Alberto Soler (Project Chief and Senior Economist)
  - Michal Andrle (Deputy Division Chief)
  - Daniel Baksa (Economist)
  - Mirza Gelashvili (Economist)
- Key activities:
  - Diagnostic meetings with MOF and other stakeholders.
  - Introductory training on DSGE models with 14 staff from different divisions within the Fiscal Policy Office.

### Institutional forecasting arrangements and core tools
- Publication and coordination practices:
  - Each main forecasting body (FPO, BOT, NESDC) publishes annual-frequency projections independently; publication months: January – MOF, February – NESDC, and March – BOT.
  - MOF publishes short-run estimates quarterly (April, July, October and January) and a 5-year MTFF once a year; MTFF nominal GDP path shared with TPB and PDMO; macro projections approved by FPC and published in December.
- FPO core forecasting tool: MAMUANG
  - Large-scale econometric quarterly model with over 300 equations.
  - Relies on error correction mechanisms estimated using a univariate OLS, with dummy variables and time trends.
  - Key exogenous inputs: GDP growth in the major trading partners; nominal exchange rate vs. US$; Dubai oil prices; tourism expenditure; government primary expenditure.
  - Implemented in EViews; outputs exported to Excel and tabulated for management discussion.
  - Endogenous variables projected by combining equation forecasts with add-factors reflecting expert judgement and quarterly projections from BOT and NESDC.
  - Public debt projections by PDMO coherently with GDP used by FPO, similar exchange rate assumptions and their own interest rate forecasts.
- Other institutions:
  - BOT: Bank of Thailand’s Macroeconometric Model (BOTMM) with 94 behavioral equations and 76 identities; semi-structural gap model (MPMIPA); ongoing work on DSGE (QIPF) with IMF MCM since 2020.
  - NESDC: time-series CQM for nowcasting and QFM for beyond very near term (around 30 endogenous variables); CGE framework for structural policy simulations (12 production sectors).
  - FPO: early-stage ADB technical assistance to develop a CGE model.

### Identified limitations of MAMUANG and current practices
- Monetary and exchange rate treatment:
  - Monetary policy does not play any role in transmission to long-term interest rates.
  - Spot exchange rate is entirely exogenous and not sensitive to foreign interest rates or risk premia.
- Macroeconomic–fiscal simultaneity and estimation concerns:
  - Lack of simultaneous determination of macroeconomic and fiscal variables can create inconsistencies (e.g., disposable income calculation; divergence between government expenditure implicit inflation and model-projected inflation).
  - Estimated coefficients may be policy-dependent and unstable; coefficients re-estimated every 4-5 years.
  - Sole use of estimated OLS coefficients poses endogeneity risks and can affect dynamic properties.
- Usability and institutional knowledge:
  - Model is difficult to comprehend and use for policy and scenario analysis despite being useful for many purposes.
  - Training and institutional knowledge gaps: user manual exists but none of the original project members remain; staff find interpretation of underlying economics difficult.

### Role envisaged for a DSGE model relative to MAMUANG
- DSGE intended to complement, not replace, MAMUANG:
  - Use DSGE for scenario analysis and policy analysis where structural transmission mechanisms and forward-looking expectations matter.
  - DSGE policy scenarios could be incorporated into MAMUANG via add-factors or modifications in exogenous variable paths.
- Frequency and integration:
  - Given fiscal policy focus, annual frequency for the DSGE is most suitable and aligns with MAMUANG frequency.

### Rationale and characteristics of STAMP (Structural Analysis of Macroeconomic Policies)
- Rationale: DSGE models are dynamic, feature endogenous forward-looking expectations, embed uncertainty, and ensure general equilibrium linking all agents and markets.
- Model class and base features:
  - State-of-the-art New Keynesian DSGE with sticky prices and frictions.
  - Small-open economy with fiscal and monetary policy blocks; two types of households; one domestic productive sector.
  - Long-term growth exogenously driven by productivity and population growth.
  - Law of one price for imports; augmented UIP with a country-risk premium.
- Box 1 — STAMP main building blocks (summary):
  1. Household block:
     - Intertemporal optimizing overlapping generation (OLG) consumers (savers) and liquidity constrained (LIQ) consumers (non-savers).
     - Savers access capital and domestic/external bonds; non-savers consume current income and receive government transfers.
     - Real rigidities: external habit formation and nominal wage rigidities.
  2. Production block:
     - Capital producers allocate domestic savings and foreign credits; investment adjustment costs; perfect competition and flexible prices.
     - Intermediate firms: CES technology, monopolistic competition, nominal price rigidities.
     - Retailers combine intermediate outputs and imports under perfect competition.
  3. Government block:
     - Intertemporal budget constraint; deficits can be covered by issuing debt.
     - Distortionary taxes on labor income, capital income, consumption, plus lump-sum taxes; current expenditures and targeted transfers.
     - Access to domestic and foreign debt; fiscal policy modeled as a forward-looking rule: primary surplus responds to current and future deviations of total public debt with respect to a target; baseline adjustments to primary balance achieved by adjusting current expenditures.
  4. Central bank and rest of the world:
     - Central bank: inflation targeting under a fully flexible nominal exchange rate; Taylor rule for policy.
     - Rest of the world: foreign inflation, external interest rates, foreign GDP modeled exogenously as autoregressive processes; foreign interest rate embeds country risk premium increasing with net foreign asset to GDP ratio.

### Customization priorities and analytical use cases
- FPO policy priority areas:
  - Role of productivity-enhancing public investment and fiscal policies.
  - Unemployment dynamics and labor skills heterogeneity.
  - Tax reforms (VAT, tariffs).
  - Redistributive policies (redesign of transfers, minimum wage increases, energy subsidies, debt moratoria).
  - Sensitivity analysis of key parameters associated with the fiscal reaction function.
- Trade-off: preserve tractability while enabling practical use; capacity development and full ownership by FPO staff are objectives.
- Proposed initial extensions:
  - Introduction of public investment.
  - Incorporation of labor heterogeneity with a large low-skilled segment.
  - Modeling minimum wage effects, energy subsidies, VAT reforms, negative income tax, ageing impacts.

### Illustrative scenarios and selected results (Box 2)
- Scope of illustrative exercises includes:
  - Expansionary monetary policy shocks with different interest rate elasticities.
  - Fiscal stimulus (financial transfers and current expenditures) with varying shares of liquidity-constrained households.
  - Labor income tax changes under different planning horizons.
  - Risk-off shocks, foreign demand shocks, sticky price importance under fiscal shocks.
- Social reform illustrative scenario (stylized emerging economy; Inflation Targeting and floating exchange rate):
  - Policy: increased financial transfers to low-income households in 2025 equal to 1 percent of the initial GDP, fading out gradually.
  - Financing: higher social spending financed through public debt issuance.
  - Results:
    - In the first year, real GDP increases by 0.6 percent.
    - Second-round effects: higher domestic demand raises labor demand and wages, generating additional inflationary pressure and prompting contractionary monetary policy.
    - Higher consumption and employment generate additional fiscal revenue, but stimulus is not sufficient to cover increased expenditures, resulting in a permanent shift of public debt.
  - Note: "Primary deficit and debt in percent of GDP as deviations in levels from steady state. GDP and private consumption in relative deviations from the steady state."
- Financing-options scenario comparing consumption tax increase vs labor income tax increase to keep debt-to-GDP near baseline:
  - Both options equally efficient in maintaining public debt ratio near baseline.
  - Higher consumption taxes reduce consumption, offsetting positive impact of transfers.
  - Higher labor income taxes discourage labor market participation, decreasing labor supply and generating a strong negative impact on real GDP.
  - Reduced economic performance and lower labor supply reduce firms' demand for domestic capital and postpone private investment, undermining potential growth.

### Reporting system, software environment, and outputs
- Reporting objectives:
  - Teach MOF staff how to communicate policy messages and enhance model storytelling.
  - Reporting system has two layers:
    - (1) Ready-to-use collection of figures and tables for presentations embedding clear charts from model simulations.
    - (2) Detailed diagnostic including each agent's decision functions and inputs, with additional explanatory figures on request.
- Standard report contents:
  - GDP and its components; key nominal variables (inflation, nominal interest rates, nominal exchange rates); external balance variables (current account, net exports, net foreign asset position, real exchange rates); fiscal block (primary balance, expenditure and revenue sides, debt accumulation); foreign variables and exogenous assumptions.
- Software environment and toolboxes:
  - MATLAB (2022b or older) — MATLAB licenses recommended to be purchased via MathWorks.
  - DYNARE (version 4.5.4)
  - IRIS Toolbox (version 20200607)
  - CompEcon Toolbox
  - DYNARE, IRIS Toolbox and CompEcon Toolbox are free; MATLAB licenses should be procured. Temporary shared 2-month licenses used during scoping mission.
  - Toolbox roles: DYNARE (solve and simulate dynamic equilibrium), CompEcon (calibration and steady state), IRIS (visualize and report results; export to Excel/CSV).

### Training, staffing, and continuity requirements
- Scoping mission training outcomes:
  - 14 FPO staff participated; eight from DMP.
  - None but one participant had prior exposure to a DSGE model.
  - Most participants had bachelor’s or master’s degrees in Economics.
  - Many participants learned to manipulate DYNARE codes and identify transmission channels in IRFs.
- Time commitment and staffing:
  - Core staff should dedicate 2-3 hours per day (approximately between 25% and 35% of their time) to project activities.
  - Core group size agreed: six participants (all from Division of Macroeconomic Policy — 4 to MMFS and 2 to Innovation and Research Section).
  - Core group participants listed: Youthapoom Charusreni; Chanon Limpasitipon; Tanut Puangnuam; Kawin Iamtrakul; Ittipat Prapaprasert; Puttisombat Pahnkul.
- Continuity mechanisms and retention:
  - Biweekly virtual meetings between missions required.
  - Documentation of the tool from mission 2 onward.
  - Mitigation incentives (subject to performance): reverse missions to Fund’s HQ; stays in HQ; selection for IMF regional training; co-authored working paper.
  - Required overlap if turnover: new member should overlap with departing member by at least four months.

### Project structure, timeline, missions, and deliverables
- Project aim: build FPO capacity to customize, operate and integrate a DSGE-based simulation tool primarily for macroeconomic policy and structural analysis.
- Tentative mission schedule and key deliverables:
  - Mission 1 (In-person) November 2024 — scoping, conceptual intro, STAMP overview, training, finalization of scoping draft report.
  - In-between work December 2024 - March 2025 — exercises, calibration data identification.
  - Mission 2 (In-person) March 2025 — linearized system, first calibration, reporting infrastructure presentation, define documentation.
  - In-between work March 2025 - June 2025 — exercises, prepare policy packages for mission 3.
  - Mission 3 (In-person) June 2025 — second version presentation, simulate policy packages, internal presentation to DMP management.
  - In-between work May 2025 - September 2025 — reporting and scenario refinement, proposed recalibration.
  - Mission 4 (virtual) September 2025 — analysis of third version, presentations to management, stocktaking.
  - In-between work September 2025 - December 2025 — refine model, integrate shocks into baseline, adjust reporting.
  - Mission 5 (In-person) December 2025 — finalize model, hands-on practice, begin integration into briefings and fiscal risk statement, discuss co-authored paper.
  - In-between work December 2025 - March 2026 — draft paper, prepare DMP Briefings and fiscal risk statement sections.
  - Mission 6 (Virtual) March 2026 — presentation to PDMO and other divisions; finalize scenarios and fiscal risk statement sections; continue paper drafting.
  - In-between work March 2026 - June 2026 — paper drafting continues; design reporting structure for NESDC; last exercises.
  - Mission 7 (In-person) June 2026 — final presentation, finalize documentation, finalize draft paper, end-project report delivery.
  - Buffer time June 2026- November 2026 for contingencies and additional activities.
- Notes:
  - Mission modality (in-person vs. virtual) indicative.
  - Documentation begins in mission 2 and continues until project end.
  - Integration of tool outputs in Fiscal Risk Statement subject to DFP agreement.
  - Co-authored paper subject to adequate performance.

### Commitments, risks, and mitigation measures
- FPO Management commits to:
  1. Ensuring core group members dedicate at least 25 per cent of standard working time daily to project activities; other TA projects should not detract from this minimum.
  2. Promoting active participation of core group members in missions and biweekly follow-ups.
  3. Incentivizing stable participation; in case of turnover, authorities shall grant a minimum 4-month overlap between new and exiting members.
  4. Ensuring MATLAB licenses are available to the core group during the whole project and beyond.
- Key risks and mitigations:
  - Risk: Insufficient time availability for staff.
    - Mitigations: authorities’ commitment to create space for staff; select core members with appropriate background.
  - Risk: High staff turnover.
    - Mitigations: advance documentation; select motivated staff closely related to the tool; incentives (publication, stays, reverse missions); encourage participation in related trainings.
  - Additional mitigation: participation in DSGE courses at HQ or regional centers when available.

### Logframe highlights (selected)
- Objective: Develop capacity in macroeconomic forecasting and policy analysis to support policy decision making and communications - MFR.
- Outcome 1: Analytical models and forecasting tools are developed and operational.
  - Annual Assessment Rating: 1 Not Achieved — Outcome Rating Date: 12/2/2024.
  - Outcome Indicator: Macroeconomic Projection Tool (MPT) is developed in the form of a fully structural DSGE model — Baseline Value: 1 — Target Value: 4 — Current Assessment Value: 1.
  - Milestones (selected):
    - DSGE structure customized to Thai economy — Target Completion Date: 12/31/2025 — Milestone Rating: 1 Not Achieved.
    - DSGE calibration completed — Target Completion Date: 3/31/2026 — Milestone Rating: 1 Not Achieved.
    - Standardized reporting structure developed — Target Completion Date: 9/30/2024 — Milestone Rating: 1 Not Achieved.
- Outcome 2: Improved analytical skills and better macroeconomic forecasting capacity.
  - Annual Assessment Rating: 1 Not Achieved — Outcome Rating Date: 12/2/2024.
  - Outcome Indicator: Relevant staff demonstrates capability to independently operate the MPT — Baseline Value: 1 — Target Value: 4 — Current Assessment Value: 1.
  - Milestones (selected):
    - Core team produces several policy and shock scenarios with the tool — Target Completion Date: 9/30/2025 — Milestone Rating: 1 Not Achieved.
    - Core team able to maintain, update and customize the DSGE model — Target Completion Date: 6/30/2026 — Milestone Rating: 1 Not Achieved.
    - Documentation of the DSGE model finalized — Target Completion Date: 6/30/2026 — Milestone Rating: 1 Not Achieved.
- Outcome 3: Authorities have a baseline understanding of forecasting capabilities and opportunities for improvement.
  - Annual Assessment Rating: 4 Fully Achieved — Outcome Rating Date: 12/2/2024.
  - Outcome Indicator: Diagnostic/scoping produced baseline and opportunities for improvement — Baseline Value: 1 No diagnostic in place — Target Value: 4 An action plan elaborated and agreed — Current Assessment Value: 4.
  - Milestone: Authorities discussed capacity assessment and agreed on action plan — Target Completion Date: 11/30/2024 — Milestone Actual Completion Date: 11/26/2024 — Milestone Rating: 4 Fully Achieved.
- Note: This logframe is provisional pending capacity assessment and discussions with the authorities.

*IMF | Thailand Technical Assistance on Macroeconomic Frameworks — PREFACE and selected sections (tarea2025082-source-pdf)*

### PREFACE ................................................................................................................

### PREFACE

### Project overview
- Begins a capacity development project between ICD and the Ministry of Finance of Thailand.
- Project initiated at the request of the Fiscal Policy Office (FPO) of the Ministry of Finance.
- Objective: develop and operationalize a structural model (in particular, a dynamic stochastic general equilibrium model) for macroeconomic policy analysis between end-2024 and mid-2026.
- Dual goal of the report:
  - Convey the project team assessment on the current toolkit, capacity, and institutional mechanisms to conduct fiscal policy and macrostructural analysis.
  - Register the action plan for the project, agreed with the authorities during the mission, to guide the delivery of technical assistance and enshrine work commitments on both the Fund and MOF’s side.

### Scoping mission: timing, team, and activities
- Scoping mission location and dates: Bangkok, Thailand, from November 18 to 26, 2024.
- Mission team:
  - Alberto Soler (Project Chief and Senior Economist)
  - Michal Andrle (Deputy Division Chief)
  - Daniel Baksa (Economist)
  - Mirza Gelashvili (Economist)
- Key activities:
  - Diagnostic meetings with MOF and other stakeholders.
  - Introductory training on DSGE models with 14 staff from different divisions within the Fiscal Policy Office.

### Meetings and main interlocutors
- Meetings held with:
  - Mr. Warotai Kosolpisitkul, International Economic Advisor of the Fiscal Policy Office
  - Mr. Pisit Puapan, Executive Director of the Division of Macroeconomic Policy
  - Mr. Norabajra Asava-Ballobv, Section Chief of Economic Data Innovation and Research
  - Mr. Youthapoom Charusreni, Section Chief of Macroeconomic Modeling and Forecasting
  - Mr. Surach Tanboon, Senior Director of the Monetary Policy Department in the Bank of Thailand
  - Mr. Sirus Pussayanavin, from the Public Debt Management Office

### Institutional context and timeline
- Project inclusion and funding:
  - Request from MOF received on March 22, 2024 (see later text for formal request reference).
  - Project included in ICD’s FY25 work program and obtained funding assurances from the Government of Japan.
- Pre-scoping meeting: May 29, 2024, led by Mr. Juan Sebastián Corrales (ICD); attendees from MOF included Mr. Pisit Puapan and Mr. Chanon Limpasitipon.
- Agreed timeline in report: model development and operationalization between end-2024 and mid-2026.

### Expected benefits and trade-offs
- Expected returns to MOF from capacity building in the DSGE model:
  - Enhanced solidity of the fiscal strategy and its communication.
  - Significant strengthening of human capital within MOF.
  - Positive spillovers to improvement of other elements of the FPO analytical toolkit.
- Caveat: project requires a strong initial investment of time and effort to realize benefits.
- Key success factors envisaged in the Action Plan:
  - Ensure sufficient time for the FPO team to work on the project.
  - Promote continuous engagement during and between missions.
  - Provide effective incentives for stable participation.

### Acknowledgements
- Gratitude expressed to the Fiscal Policy Office for hospitality, organization, and logistical support.
- Special thanks to:
  - Mr. Chanon Limpasitipon, Senior Economist, Section of Macroeconomic Modeling and Forecasting
  - Dr. Youthapoom Charusreni, Chief of the same section
  - Country team of Thailand from the Asia and Pacific Department (APD), led by Ms. Corinne C. Delechat
  - Mrs. Eteri Kvintradze, Director of CDOT
  - Collaborating MCM and FAD teams involved in mission preparation

*IMF | Thailand Technical Assistance on Macroeconomic Frameworks — PREFACE*

### 8.      All of three main national forecasting bodies –the FPO, BOT, and NESDC—use a portfolio

### 8.      All of three main national forecasting bodies –the FPO, BOT, and NESDC—use a portfolio of models as inputs into their projections

### Institutional forecasting arrangements and publication practices
- Each institution publishes only annual-frequency projections, which are independent and not shared with the others until officially published.
- Publication months: January – MOF, February – NESDC, and March – BOT.
- Forecasting teams regularly interact to discuss common assumptions (foreign outlook, government policies, ...) and openly share data among themselves.
- The MOF publishes short-run estimates quarterly (April, July, October and January) and a 5-year MTFF once a year; the nominal GDP path from the MTFF is shared with the TPB and the PDMO and the whole set of macro projections is approved by the FPC and published on the MOF website in December.

### FPO core forecasting tool: MAMUANG (Macroeconomic Forecasting Model of Ministry of Finance of Thailand)
- Model type and frequency:
  - Large-scale econometric quarterly model with over 300 equations.
  - Relies on error correction mechanisms estimated using a univariate OLS, with dummy variables and time trends.
- Key exogenous inputs explicitly used:
  - GDP growth in the major trading partners
  - Nominal exchange rate vs. US$
  - Dubai oil prices
  - Tourism expenditure
  - Government primary expenditure
- Model implementation and workflow:
  - System of equations solved in EViews; outputs exported to Excel and tabulated for discussion with Ministry management.
  - Endogenous variables projected by combining equation forecasts with add-factors reflecting expert judgement and quarterly projections from the BOT and the NESDC.
  - Government revenues forecast by the Division of Fiscal Policy using information provided by the DMP and collaborating with relevant tax collection departments.
  - Public debt projections made by the Public Debt Management Office (PDMO), coherently with the GDP used by the FPO, similar exchange rate assumptions and their own interest rate forecasts.
- Complementary tools and practices:
  - Sectoral experts in the MMFS have developed satellite models for major trade partner GDP growth, exchange rate, tourism expenditure, oil prices and public expenditure.
  - Nowcasting and Near-Term Forecasting (NTF) tools rely on econometric and machine learning models; results are used as inputs for MAMUANG.
  - Satellite models also produce bottom-up aggregates (e.g., inflation by aggregation of sub-components) for cross-validation with system-of-equations headline inflation.

### Identified limitations and concerns with MAMUANG
- Monetary policy and exchange rate treatment:
  - Monetary policy does not play any role in transmission to long-term interest rates.
  - Spot exchange rate is entirely exogenous and not sensitive to foreign interest rates or risk premia.
- Macroeconomic–fiscal simultaneity:
  - Lack of simultaneous determination of macroeconomic and fiscal variables can create inconsistencies, including:
    - Calculation of disposable income, as the model does not determine direct tax collection.
    - Potential divergence between inflation implicit in government expenditure assumptions and inflation projected by the model after introduction of those assumptions.
- Estimation and stability issues:
  - Estimated coefficients may be policy-dependent and unstable.
  - Coefficients are currently re-estimated every 4-5 years, which may not effectively mitigate instability.
  - Sole use of estimated OLS coefficients poses risks related to endogeneity and can affect dynamic properties of the model.
- Usability for policy and scenario analysis:
  - Model is difficult to comprehend and use for policy and scenario analysis despite being useful for many purposes.
  - Training and institutional knowledge gaps: project left a user manual and the team can operate the model, but none of the original project members remain, and staff find it difficult to interpret some underlying economics.

### Role envisaged for a DSGE model relative to MAMUANG
- The DSGE is intended to complement, not replace, MAMUANG.
- Use cases:
  - Supplement MAMUANG for scenario analysis and policy analysis where structural transmission mechanisms and forward-looking expectations matter.
  - DSGE policy scenarios could be incorporated into MAMUANG via add-factors or modifications in paths of exogenous variables; EViews implementation facilitates such tunes.
- Medium-term considerations:
  - FPO could consider streamlining its forecasting model, studying intersectoral consistency, reviewing dynamic properties and calibration, leveraging experience gained during the DSGE project.

### Data management, availability, and coordination
- Data frequency and coverage:
  - Most macroeconomic data (national accounts, BOP, labor market, monetary sector, etc.) are available at least quarterly with a regular publication calendar.
  - Quarterly GDP data are published by the NESDC, alongside an updated projection.
  - No large data gaps identified for calibrating a DSGE, though discrepancies between budget accounting and GFS were noted.
- Databases and version control:
  - Some large-scale fiscal and microdata databases maintained by IT staff using professional database software.
  - Macroeconomic database in FPO maintained in MS Excel on an internal shared drive; vintages are monthly archived.
  - Sector-assigned staff update sector data; IT staff back up internal share drive regularly, reducing risk of different vintages being used.
- Access issues:
  - Problems arise mainly where detailed micro data or specialized provider data are needed; available data may be too expensive or subject to administrative/privacy barriers.

### Other institutions’ models and capabilities
- Bank of Thailand (BOT):
  - Uses a suite of economic, econometric, and machine-learning models.
  - Core model: Bank of Thailand’s Macroeconometric Model (BOTMM) with 94 behavioral equations and 76 identities; estimated error-correction equations.
  - BOTMM covers four sectors: real, monetary, balance of payments and fiscal sector.
  - BOT publishes short-term quarterly projections by the end of every quarter.
  - BOT also uses a semi-structural gap model (Monetary Policy Model for an Integrated Policy Analysis – MPMIPA) extended for detailed financial and macroprudential sector; fiscal sector in that model is highly stylized and aggregated.
  - BOT has developed a structural DSGE in the past, not currently used; since 2020 has been working with IMF MCM to operationalize an estimated DSGE (QIPF), with continuation of work and training up to November 2024.
- NESDC:
  - Toolkit includes time-series CQM (Current Quarter Model) for nowcasting and near-term forecasting (2 quarters) and a semi-structural macroeconometric model QFM beyond the very near term.
  - QFM features around 30 endogenous variables and a similar number of exogenous variables.
  - Potential GDP projections informed by growth accounting (production function) and assumptions on future growth drivers.
  - Publishes short-term projections every second month of all quarters (estimates for current year and one-year ahead by last quarter of the present year).
  - Medium-term GDP projections partly informed by MOF and BOT views and serve as the basis for the MTFF.
  - NESDC specializes in medium- and long-term structural and sectoral policy priorities; uses CGE framework for structural policy simulations (12 production sectors, calibrated using the social accounting matrix).
- FPO and external assistance:
  - FPO is in an early stage of ADB technical assistance to develop a CGE model.

### Macroeconomic risk analysis, stress testing, and policy scenario practices
- FPO scenarios and limitations:
  - MMFS elaborates alternative scenarios by modifying exogenous variable paths or introducing add-factors; scenarios are discussed internally but lack pre-established production frequency.
  - Scenario outputs presented as numerical differences from the baseline.
  - Existing model limitations: lack of real integration of macroeconomic and fiscal outcomes and missing key transmission mechanisms.
  - Recommendation: alternative scenarios should be produced regularly (at least quarterly), be based on a structural or semi-structural model, and include analysis of shock transmission mechanisms rather than only numerical results.
- Fiscal policy assessment at MMFS:
  - Core forecasting model used for assessment; suited to policies with direct budgetary impact.
  - Revenue measures assessment guided by indicative multipliers; model lacks parametrization of tax rates, limiting comparison of financing mechanisms and analysis of non-fiscal structural policies.
  - Absence of explicit forward-looking expectations limits analysis of anticipated fiscal measures.
- PDMO stress testing:
  - PDMO conducts sensitivity exercises and stress tests on its baseline projections 3 times per year to define short- and medium-term targets.
  - Shocks include GDP, interest rates, exchange rates, and materialization of explicit and implicit contingent liabilities.
  - Results over a 5-year horizon inform Annual Debt Management Plans and calibrate buffers in the MTFF debt stock path.
  - Simulations used to update debt-related limits of the fiscal rule, considering cyclical position and viability of discretionary consolidations.
- Longer-term DSAs:
  - DSAs over a 10-20 period offer complementary insights (e.g., ageing-related expenditure, tax base evolution) and should extend baseline projections with deterministic shocks and probabilistic analysis.
  - Recommendation: conduct DSAs at least once per year and include them in MTFF discussions and MTDS updates; involve FPO, DFP and the BB in designing deterministic scenarios.
- BOT policy modeling emphasis:
  - BOT’s semi-structural gap model demonstrates benefits of fiscal-monetary coordination and macroprudential policies.
  - Model includes nonlinearities from interactions when policy rate hits the zero lower bound, protracted low interest rates effects on debt and credit quality, and exchange rate appreciation impacts on export-driven sectors.

### CGE and sectoral modeling
- NESDC and Ministry of Commerce maintain in-house CGE models for sectoral and distributional macro analysis.
- NESDC’s CGE:
  - Incorporates intra- and inter-temporal dimensions within a neoclassical growth model.
  - Features 12 production sectors and is fully calibrated using the social accounting matrix.
- FPO is developing a CGE model with ADB technical assistance.

*IMF | Thailand Technical Assistance on Macroeconomic Frameworks*

### 27.        MTFF preparation could benefit from a more systematic and deeper discussion of scenario

### 27.        MTFF preparation could benefit from a more systematic and deeper discussion of scenario analyses.

### MTFF and scenario analysis
- Finding: While policy scenarios based on major measures have occasionally been discussed within the government, a systematic scenario-based analysis of policy options has not yet been incorporated to the MTFF process.
- Recommendation: Incorporate policy scenarios that include both macroeconomic and fiscal outcomes to reflect strategic alternatives, such as:
  - the pace of fiscal consolidation,
  - funding options for social spending measures,
  - productivity-enhancing reforms.
- Institutional proposal: Discuss these scenarios in detail by a technical working group of the FPO.
- Complementarity: DSAs, which often capture more in detail some long-term factors than structural models, could be a good complement to scenario discussions.
- Risk balancing: During the elaboration of the medium-term macroeconomic projections, MOF, NESDBC and BOT could make a more intensive use of macroeconomic shock scenarios to better balance the risks implicit in the baseline.

### Fiscal Risk Statement and macroeconomic risk analysis
- Current practice: The government publishes an annual Fiscal Risk Statement jointly prepared by MOF, NESDC and BOT that contains a data-based qualitative analysis of specific fiscal risks to revenue, expenditure and public debt, including explicit and implicit contingent liabilities.
- Gap: The analysis of macroeconomic risks and their potential fiscal consequences is largely absent from the report.
- Suggested enhancements:
  - Include a macroeconomic risk matrix reflecting the origin of risks, their probability and possible effects.
  - Add several quantitative scenarios in general equilibrium to show interrelations between macroeconomic, fiscal and financial variables and identify amplifying/mitigating factors.
  - Use the DSGE model to enrich the Fiscal Risk Statement along these lines.

### Rationale for the adoption of a DSGE model
- Suitability: DSGE models are dynamic, feature endogenous forward-looking expectations, embed uncertainty, and ensure general equilibrium linking all agents and markets.
- Advantages over existing tools:
  - Rich transmission mechanisms and microfoundations for simulation.
  - Ability to translate policymakers' preferences into model variables and reflect country characteristics.
  - Capacity to warn about longer-horizon consequences of policy measures and inform preventive/corrective action.
- Intended role: The structural macroeconomic model will not generate baseline projections but can inform them and steer discussion between FPO advisors and policymakers on the nature of macroeconomic shocks and potential policy implications.
- Frequency: Given the primary focus on fiscal policy planned on an annual basis, an annual frequency for the DSGE is the most suitable option.

### Characteristics of the STAMP (Structural Analysis of Macroeconomic Policies)
- Model class: State-of-the-art New Keynesian DSGE with sticky prices and frictions enabling fiscal and monetary policies to yield short-term effects on aggregate demand.
- Base features:
  - Small-open economy with fiscal and monetary policy blocks.
  - Two types of households and one domestic productive sector.
  - Long-term growth exogenously driven by productivity and population growth.
  - Law of one price for imports; imperfect integration into international capital markets via an augmented UIP with a country-risk premium.
- Intended implementation: Annual frequency to align with fiscal policy focus and MAMUANG frequency.

- Box 1 — STAMP main building blocks:
  1. Household block:
     - Intertemporal optimizing overlapping generation (OLG) consumers (savers) and liquidity constrained (LIQ) consumers (non-savers).
     - Savers access capital and domestic/external bonds; non-savers consume current income and receive government financial transfers.
     - Real rigidities include external habit formation and nominal wage rigidities.
  2. Production block:
     - Capital producers allocate domestic savings and foreign credits; face investment adjustment costs; operate under perfect competition and flexible prices.
     - Intermediate firms use CES technology combining capital and labor; face monopolistic competition and nominal price rigidities.
     - Retailers combine intermediate outputs and imports to produce aggregate output; operate under perfect competition and flexible prices.
  3. Government block:
     - Intertemporal budget constraint; deficits can be covered by issuing debt.
     - Distortionary taxes on labor income, capital income, consumption, plus lump-sum taxes; current expenditures and targeted transfers.
     - Access to domestic and foreign debt; financing plan is a function of exogenous decisions.
     - Fiscal policy modeled as a forward-looking rule: primary surplus responds to current and future deviations of total public debt with respect to a target. In the baseline, adjustments to primary balance are achieved by adjusting current expenditures until debt returns to target; other revenue/expenditure instruments are exogenous AR processes or optional instruments.
     - Central bank: inflation targeting under a fully flexible nominal exchange rate; Taylor rule used for policy decisions; UIP condition driven by interest differential adjusted by risk premium and expected nominal exchange rate.
  4. Rest of the world block:
     - Foreign inflation, external interest rates, and foreign GDP modeled exogenously as autoregressive processes with stochastic shocks.
     - Foreign GDP affects export demand; foreign interest rate embeds a country risk premium increasing with the net foreign asset to GDP ratio.
- Equilibrium: Market equilibrium conditions determine equilibrium prices and quantities and external balance consistently determines changes in net foreign debt with net exports and foreign income.

### Customization priorities and operational considerations
- Policy priority areas identified by FPO managers:
  - Role of productivity-enhancing public investment and fiscal policies,
  - Unemployment dynamics and labor skills heterogeneity,
  - Tax reforms (VAT, tariffs),
  - Redistributive policies (redesign of existing transfers, introduction of new transfers, minimum wage increases, energy subsidies, debt moratoria).
- Additional area of interest: sensitivity analysis of key parameters associated with the fiscal reaction function.
- Balance: Preserve tractability while enabling practical use; capacity development and full ownership by FPO staff are objectives of ICD.
- Scoping mission proposed customization rounds and initial aspects raised by staff for model extension:
  - Introduction of public investment,
  - Incorporation of labor heterogeneity with a large low-skilled segment possibly earning the minimum wage or lower salaries in the informal sector,
  - Economic impact of raising minimum wages,
  - Modeling government subsidies to energy prices,
  - Reforms in the tax system including VAT rises and a negative income tax,
  - Economic impact of an ageing society.

### Illustrative scenarios and analytical possibilities (Box 2)
- Scope of illustrative exercises: expansionary monetary policy shocks with different interest rate elasticities; fiscal stimulus (financial transfers and current expenditures) with varying shares of liquidity-constrained households; labor income tax changes under different planning horizons; sticky price importance under fiscal shocks; risk-off shocks with different import substitution levels; foreign demand shocks with varying risk premium elasticities; financial transfer shocks with different fiscal consolidations.
- Social reform illustrative scenario (calibrated to a stylized emerging economy with Inflation Targeting and floating exchange rate):
  - Policy: increased financial transfers to low-income households in 2025 equal to 1 percent of the initial GDP, fading out gradually in following years.
  - Financing: higher social spending financed through public debt issuance.
  - Results:
    - In the first year, real GDP increases by 0.6 percent as low-income households increase consumption.
    - Second-round effects: higher domestic demand raises labor demand, shifting wages upward and generating additional inflationary pressure, prompting contractionary monetary policy.
    - Higher consumption and employment generate additional fiscal revenue, but stimulus is not sufficient to cover increased expenditures, resulting in a permanent shift of public debt.
  - Footnote clarification: "Primary deficit and debt in percent of GDP as deviations in levels from steady state. GDP and private consumption in relative deviations from the steady state."
- Financing-options scenario comparing two financing strategies to keep debt-to-GDP around baseline:
  - Increase in consumption taxes vs increase in labor income taxes.
  - Findings:
    - Both options equally efficient in maintaining public debt ratio near baseline.
    - Higher consumption taxes reduce consumption, offsetting positive impact of transfers.
    - Higher labor income taxes discourage labor market participation, decreasing labor supply and generating a strong negative impact on real GDP.
    - Reduced economic performance and lower labor supply (especially with labor tax increases) reduce firms' demand for domestic capital and postpone private investment, undermining potential growth.
  - Analytical strength: STAMP provides consistent narrative across agents and policies, quantifies impacts across macro variables, and helps decision-makers understand costs and benefits of financing schemes and consolidation plans.

*Source: IMF | Thailand Technical Assistance on Macroeconomic Frameworks — sections 27–36.*

### 37.        The development of a standardized reporting system will be an important objective of this

### The development of a standardized reporting system will be an important objective of this project.

### Reporting system: design, outputs, and customization
- The project emphasizes teaching MOF staff how to communicate policy messages and enhancing the model's storytelling capacities.  
- The reporting system must be "good, streamlined and informative" to enable the team to deliver clear and relevant recommendations to policy makers and produce effective inputs for presentations to senior management at the MOF in the form of charts and tables.
- The reporting system has two layers:
  - (1) A ready-to-use collection of figures and tables as direct input for presentation that embeds clear and simple charts from the model simulation.
  - (2) A detailed diagnostic of the model that includes each agent's decision functions and inputs for their decision-making, providing a deep understanding of channels and transmission mechanisms, and upon request, additional explanatory figures.
- The project team will provide templates and an initial version of reports and model-based presentations; the reporting system is a living document and MOF staff will learn how to adjust figures and tables and add variables when needed.
- Standard report contents include:
  - GDP and its components (expenditure, supply-side, or income-side decomposition).
  - Key nominal variables: inflation, nominal interest rates, nominal exchange rates.
  - External balance variables: current account, net exports, net foreign asset position, real exchange rates.
  - Fiscal block: primary balance, expenditure and revenue sides, debt accumulation, deficit path consistent with the assumed fiscal framework behavior.
  - Additional pages listing foreign variables and exogenous assumptions.
- Reporting capabilities:
  - Compare short and medium-term impacts of scenarios and provide exact numbers in tables along with figures.
  - Reports are automatically generated by the procedure and saved in PDF format.
  - Results from the model can also be exported to other databases, like Excel or CSV files.

### Software environment and licensing
- STAMP codes combine different MATLAB toolboxes; procurement of MATLAB licenses is strongly recommended as soon as possible to avoid relying on temporary ones.
- Annotated code infrastructure includes and combines:
  - MATLAB (2022b or older)
  - DYNARE (version 4.5.4)
  - IRIS Toolbox (version 20200607)
  - CompEcon Toolbox
- DYNARE, IRIS Toolbox and CompEcon Toolbox are free software and available on the web; MATLAB licenses should be purchased via the MathWorks.
- Temporary solution used during scoping mission: shared 2-month licenses with the FPO team.
- Toolbox functions (as described in footnote):
  - DYNARE: solves dynamic equilibrium conditions and simulates the model; allows analysis of dynamic properties, e.g., transmission mechanisms of shocks and policies.
  - CompEcon Toolbox: perform calibration and solve for the steady state in relatively complex models (e.g., interactions between heterogenous agents).
  - IRIS Toolbox: visualize and report model results; save simulation results into consistent databases; export results to external software (Microsoft Excel or other statistical software).

### Training, capacity building, and staffing requirements
- Scoping mission training outcomes:
  - 14 FPO staff participated in the training, eight of which were from the DMP.
  - None but one of these participants had previously been exposed to a DSGE model.
  - Most participants had bachelor’s or master’s degrees in Economics.
  - Many participants quickly learned to manipulate DYNARE codes of basic DSGE models and identify transmission channels in IRFs.
- Time commitment required:
  - Customizing STAMP and learning to operate it autonomously will take significant staff time.
  - Staff should dedicate 2-3 hours per day (approximately between 25% and 35% of their time) to project activities.
  - Releasing core team members’ daily work time is a must for project success; workload in the MMFS is high, with core tasks concentrated on 4-5 staff.
  - Maintenance and full incorporation of the tool may warrant hiring additional staff beyond the project timeframe.
- Continuity mechanisms:
  - Biweekly virtual meetings between missions to: (i) present and discuss core team solutions to ICD-provided exercises/tasks; (ii) resolve technical questions. Discontinuation of these meetings would render the project infeasible.
  - Documentation of the tool (operational and analytical aspects) to be developed from the second mission onward.
- Turnover and retention:
  - High staff turnover at the FPO is a critical project risk (structural characteristic of the Thai MOF).
  - Mitigation measures:
    - From the second mission, the core group will elaborate detailed documentation under ICD guidance.
    - ICD incentives (subject to adequate staff performance) could include:
      - (i) organization of reverse missions to the Fund’s HQ,
      - (ii) stays of some core group members in HQ working on the model with ICD support,
      - (iii) selection of some core group members to participate in training courses organized by IMF regional training center,
      - (iv) preparation of a working paper jointly between IMF staff and core group members to be published by the MOF.
    - If turnover is unavoidable, the new team member should overlap with the departing member by at least four months.

### Core group composition and role
- Core group criteria considered: (i) background and experience related to the project; (ii) relation of staff assignments to operationalization and/or analytical exploitation of the tool; (iii) prospects of staff stability at the FPO during the project and embedding period.
- The agreed core group size and composition:
  - Six participants, all belonging to the Division of Macroeconomic Policy (4 of them to the MMFS and 2 to the Innovation and Research Section).
- Core group participants (Table 1):
  - Youthapoom Charusreni — Macroeconomic Policy — Director of Macroeconomic Modeling and Forecasting Section
  - Chanon Limpasitipon — Macroeconomic Policy — Economist, Professional Level (Coordinator)
  - Tanut Puangnuam — Macroeconomic Policy — Economist, Professional Level
  - Kawin Iamtrakul — Macroeconomic Policy — Economist, Professional Level
  - Ittipat Prapaprasert — Macroeconomic Policy — Economist, Practitioner Level
  - Puttisombat Pahnkul — Macroeconomic Policy — Economist, Practitioner Level
- Participation guidance:
  - The number of participants in the introductory training seemed too high for the core group; an intermediate level of English is a minimum requirement to develop fluent communication with the ICD team.
  - Consideration for ad-hoc members from other divisions (Tax Policy, Fiscal Policy, Macroeconomic Analysis Section) to participate in design and interpretation of specific policy scenarios.
  - Ad-hoc membership can leverage expertise for tax or expenditure reform scenarios and customize the fiscal block.

### Project structure, timeline, and deliverables
- Project aim:
  - Build capacity in the FPO to customize, operate and integrate in policy analysis a simulation tool based on a dynamic stochastic general equilibrium (DSGE) model; tool to be used primarily for macroeconomic policy and structural analysis at the MOF and complement FPO’s forecasting tools.
- Coordination and operation:
  - Project coordinated from the DMP of the FPO; tool principally operated by the Macroeconomic and Modeling Forecasting Section.
- Customization and calibration plan:
  - After scoping mission, DSGE model will be customized and calibrated in successive rounds beginning in mission 2 and tentatively finalizing in mission 5.
  - Mission 2 starting point: canonical STAMP with a first calibration for Thailand; present broad structure and analyze steady state and model properties to identify customization and recalibration needs.
  - Priority given first to customization elements achievable without large structural changes to the canonical model; replicate some empirical properties via combinations of shocks where possible.
- Presentations schedule:
  - End of the third mission: first presentation to FPO officials showing model properties and collecting improvement views.
  - Fourth mission: presentation highlighting impact of key macroeconomic policy measures selected by the core team.
  - Final presentation by end of project presenting a wide range of results with the final model version.
  - Other presentations to external stakeholders included in the work program.
- Exercises and learning path:
  - Core team will solve exercises starting with simple DSGE models sharing characteristics with STAMP, then focus on the customized Thailand model.
  - Exercises include derivation and linearization of equations, calibration review, derivation of steady states, analysis of alternative scenarios.
  - Exercise results presented during in-between sessions and missions and followed by project team feedback.
- Integration into policy decision-making:
  - Documentation of the model begins in the second mission and continues through every mission until project end.
  - Reporting structure of results developed from the second mission.
  - From the fifth mission, the core group will work on integrating DSGE outputs into key macro-fiscal reports and notes (published and internal).
  - Subject to adequate performance, the work program contemplates co-authoring a paper with the core team during the second half of the project; central scenario envisions publication by the MOF.
- Missions and contingencies:
  - Work program tentatively includes seven missions, including the scoping one that has just finalized.
  - Delivery modalities combine in-person and virtual CD; in-person missions frontloaded during the first half of the project.
  - Virtual missions are indicative and to be discussed with the team during CY2025.
  - Reverse missions (FPO staff traveling to Fund’s HQ) could be considered subject to good performance.
  - In-between mission engagements will always be virtual.
  - A 6-month "buffer time" between June and November 2026 is included for contingencies and additional activities; the ICD team will discuss during CY2026Q1 the eventual need to utilize this buffer.

*Source: IMF | Thailand Technical Assistance on Macroeconomic Frameworks — excerpts from the provided chapter/section.*

### 57.        Importantly, the effective length of the project depends to a great extent on the core

### 57.        Importantly, the effective length of the project depends to a great extent on the core group’s efforts and progress. It could be expected that, the more time is regularly dedicated to working on the project, and the more continuous the group’s engagement with the ICD team, the less time it should take to get positive results.

### Project timeline and CD activities (Table 2: CD activities during the project)
- Mission 1 (In-person) November 2024
  - Capacity and analysis system assessment
  - Training: conceptual introduction to DSGE models, economic foundations
  - Overview of STAMP characteristics and functionalities
  - Finalization of scoping draft report
- In-between mission work December 2024 - March 2025
  - Series of exercises to be resolved by the core group, based on the presentations of mission 1 training
  - The core team identifies relevant time series for the calibration of big ratios and assessment of basic characteristic of the Thai economy.
  - Discussion on immediate customization needs continues.
- Mission 2 (In-person) March 2025
  - Structure of STAMP: presentation of the linearized system of equations, structural assumptions and implicit transmission mechanisms. Review of big ratios and calibration criteria.
  - Analysis of the properties of the model with the first calibration, with emphasis on the fiscal balance and public debt path. Discussion of recalibration needs within the discussion about the characteristics of the Thai economy.
  - Presentation of a reporting infrastructure and automation of input data relevant to the model.
  - Definition of the information to include in the documentation of the tool
- In-between mission work March 2025 - June 2025
  - The core group completes series of exercises on deriving and solving a simple general equilibrium model
  - Development and interpretation of scenarios with the homework model
  - Discussion of the detailed composition of policy packages to simulate during mission 3. Emphasis on fiscal policy reforms and policies to boost trend growth.
- Mission 3 (In-person) June 2025
  - Presentation of the second version of the model based on the decisions made in the previous mission. Study of the properties of the recalibrated version of the model.
  - Simulation of policy packages proposed by the core group. Distillation of conclusions and preparation of a narrative.
  - Preparation of an internal presentation for DMP management based on the second version of the model and preparation of a narrative for scenarios.
  - Presentation to management and discussion of feedback. Stocktaking of possible recalibration or customization needs.
- In-between mission work May 2025 - September 2025
  - The core group completes work on reporting and policy scenarios with the second version of the model.
  - Core group members exercise themselves in the design and explanation of scenarios reflecting relevant economic phenomena.
  - The core group presents its proposed recalibration for a number of parameters, based on a compilation of empirical studies or analyses of their own.
  - The core group replicates the derivation of some selected parts of the model.
- Mission 4 (virtual) September 2025
  - Analysis of the properties of the third version of the model
  - Preparation and delivery of a presentation to DMP and DFP management that assesses relevant policy measures through the second version of the model.
  - Stocktaking of recalibration and extension needs.
  - Discussion on options for an integrated narrative for baseline projections and alternative scenarios.
  - Discussion on possible ways to integrate alternative scenarios produced with the model in internal and external reports produced by FPO, and analytical discussions with the NESDC
  - Identification of further customization needs in the reporting infrastructure.
- In-between mission work September 2025 - December 2025
  - The core group continues the works on refining the model.
  - Further practice with the derivation of some parts of the model.
  - The core group conducts a series of exercises to integrate shocks simulated in model into its baseline forecast.
  - Adjustments in the reporting infrastructure.
- Mission 5 (In-person) December 2025
  - Finalization of the model based on the feedback from the previous mission, management and policy questions.
  - Additional hands-on practice to introduce the outcomes of simulated shocks and policy measures as alternative scenarios to baseline projections.
  - Introduction of changes in the reporting infrastructure. The team begins the preparation of some alternative scenarios and their narrative in the DMP’s briefings and/or the fiscal risk statement.
  - Discussion on the structure of a publication co-authored by the team on the final version of the model calibrated for Thailand.
- In-between mission work December 2025 - March 2026
  - The core group starts drafting the paper.
  - The core group continues the preparation of sections for the DMP Briefings and/or fiscal risk statement.
- Mission 6 (Virtual) March 2026
  - Technical presentation of the functionalities of the tool to the PDMO and other remaining divisions within the FPO.
  - Discussion on possible improvements in the reporting system or discussion of the tool’s outputs to ensure its relevant role in policy-making decisions.
  - Preparation of updated scenarios in view of the policy priorities for next fiscal year.
  - The team concludes the relevant sections of the fiscal risk statement, and develops prepares material for the presentation of alternative scenarios in the FPO, together with its baseline projections.
  - Drafting of the paper continues.
- In-between mission work March 2026 - June 2026
  - Paper drafting continues
  - The core team designs a reporting structure for presentation of the tool results to the NESDC and elaborates a calendar for the utilization of the tool.
  - Last round of exercises on shock for discussion with management.
- Mission 7 (In-person) June 2026
  - Final presentation to management and other divisions of the FPO of the refined outputs of the tool and pilot report
  - Finalization of the model documentation
  - Finalization of the draft paper
  - Wrap-up and delivery of the end-project report
- Buffer time June 2026- November 2026
  - Resolution of additional difficulties in the operation of the tool
  - Follow-up on the integration of the tool in the policy analysis systen
  - Possible additional missions

Notes related to schedule and deliverables:
- 1/ The modality of missions (in-person vs. virtual) is for now indicative.
- 2/ Preparation of the model documentation will start in mission 2 and continue until the end of the project.
- 3/ Most relevant reports would the Fiscal Risks Statement, and internal documentation for presentations of projections within the FPO and multi-annual projection discussions with the NESDC.
- 4/ Inclusion of the tool outputs in the Fiscal Risk Statement is subject to the agreement of the Division of Fiscal Policy, tasked with the coordination of this document.
- 5/ The preparation of this paper, to be published by the MOF, would be subject to an adequate performance and progress by the team.
- Recommendation during final missions: Subject to the FPO authorities’ agreement, set up presentations on the functionalities of the tool to the NESDC and BOT.

### Commitments of the authorities (Section 58)
- FPO Management commits to:
  1. Ensuring that core group members are able to devote at least 25 per cent of their standard working time, on a daily basis, to those activities directed to build their technical capacity within the project. Other ongoing or future TA projects in which core group members participate should not detract from this minimum time.
  2. Promoting the active participation of the core group members in missions and in-between mission follow-up meetings every two weeks.
  3. Incentivizing the stable participation of the core group members in the whole project and the consolidation of the tool in the FPO. In case of core group’s members turnover, authorities shall grant a minimum 4-month overlap of the new group member with the exiting one, to facilitate the training of the former.
  4. Ensuring that MATLAB licenses are available to the core group during the whole project and beyond, to allow for the operation of the tool.

### Risks of the project (Section 59) and mitigation measures
- Risk: Insufficient time availability for staff. A high workload could compromise an adequate dedication to the project by the core group.
  - Mitigation measures:
    - (i) The authorities’ commitment to creating space for the participating staff should largely neutralize this risk;
    - (ii) The selection of core group members with the right background should also reduce the possibilities of inefficiently large time requirements to meet the project’s needs.
- Risk: High staff turnover. Entry costs to a project of these characteristics could be high for new staff, and the replacement of core group members result in considerable delays of the project.
  - Mitigation measures:
    - (i) The tool documentation should steadily advance during the project, with specific time slots in each mission and in-between mission follow-ups allocated to the review of material;
    - (ii) The selection of staff for the core group rest on their motivation and professional closeness to the tool;
    - (iii) The work program includes some possible incentives to discourage turnover, such as the publication of the paper or staff’s stays and reverse missions at HQ;
    - (iv) It is in FPO’s direct interest to consolidate a solid human capital in the wake of this project, and Management is encouraged to consider possible incentive schemes to retain as much as possible the most capable staff.
  - Additional consideration: Participation of core group members in related trainings and courses delivered in HQ or the Fund’s regional centers (e.g., DSGE courses when available).

### Logframe highlights (ANNEX 1)
- Objective:
  - Develop capacity in macroeconomic forecasting and policy analysis to support policy decision making and communications - MFR
- Outcome 1:
  - Analytical models and forecasting tools are developed and operational
  - Annual Assessment Rating: 1 Not Achieved
  - Outcome Rating Date: 12/2/2024
  - Outcome Indicator: Macroeconomic Projection Tool (MPT) is developed in the form of a fully structural DSGE model
    - Baseline Value: 1
    - Target Value: 4
    - Current Assessment Value: 1
  - Milestones:
    - DSGE structure is customized to account for the specific characteristics of the Thai economy — Target Completion Date: 12/31/2025 — Milestone Rating: 1 Not Achieved
    - DSGE calibration is completed — Target Completion Date: 3/31/2026 — Milestone Rating: 1 Not Achieved
    - A standardized reporting structure for the tool’s outputs is developed — Target Completion Date: 9/30/2024 — Milestone Rating: 1 Not Achieved
- Outcome 2:
  - Improved analytical skills, and better macroeconomic forecasting and policy analysis capacity
  - Annual Assessment Rating: 1 Not Achieved
  - Outcome Rating Date: 12/2/2024
  - Outcome Indicator: Relevant staff demonstrates the capability to independently operate the MPT and other auxiliary analytical models and tools
    - Baseline Value: 1
    - Target Value: 4
    - Current Assessment Value: 1
  - Milestones (selected):
    - Core team members demonstrate their ability to produce several policy and shock scenarios in with the tool — Target Completion Date: 9/30/2025 — Milestone Rating: 1 Not Achieved
    - Core team members demonstrate their ability in maintaining, updating and customizing the DSGE model — Target Completion Date: 6/30/2026 — Milestone Rating: 1 Not Achieved
    - Core team presents the DSGE model to its management, and show its functionalities for policy scenario analysis — Target Completion Date: 12/31/2025 — Milestone Rating: 1 Not Achieved
    - Core team conducts a pilot integration of DSGE outputs into its analytical report system — Target Completion Date: 6/30/2026 — Milestone Rating: 1 Not Achieved
    - Documentation of the DSGE model is finalized — Target Completion Date: 6/30/2026 — Milestone Rating: 1 Not Achieved
- Outcome 3:
  - Authorities have a baseline understanding of their forecasting and analytical capabilities and opportunities for improvement
  - Annual Assessment Rating: 4 Fully Achieved
  - Outcome Rating Date: 12/2/2024
  - Outcome Indicator: Diagnostic or scoping has produced baseline and opportunities for improvement
    - Baseline Value: 1 No diagnostic in place
    - Target Value: 4 An action plan for improvement has been elaborated and agreed with the authorities
    - Current Assessment Value: 4
  - Milestone:
    - Authorities discuss the capacity assessment with the scoping mission team and agrees on an action plan — Target Completion Date: 11/30/2024 — Milestone Actual Completion Date: 11/26/2024 — Milestone Rating: 4 Fully Achieved

- Note: 1/ This logframe is only provisional, pending the capacity assessment and discussions with the authorities in the upcoming mission.

*Source: https://www.imf.org/-/media/files/publications/tar/2025/english/tarea2025082-source-pdf.pdf*

---


_Source: https://www.imf.org/-/media/files/publications/tar/2025/english/tarea2025082-source-pdf.pdf_
