## Preface and Executive Summary — TA Project on Strengthening Macroeconomic Modeling in Support of FPAS at the NBU

## Source details

**Canonical URL:** [Preface and Executive Summary — TA Project on Strengthening Macroeconomic Modeling in Support of FPAS at the NBU](https://www.imf.org/-/media/files/publications/tar/2026/english/tarea2026010-source-pdf.pdf)

## Other formats

- [Markdown version](/-/media/files/publications/tar/2026/english/tarea2026010-source-pdf.pdf.md)
- [Structured JSON version](/-/media/files/publications/tar/2026/english/tarea2026010-source-pdf.pdf.json)

---

### Project background and scope
- At the request of the National Bank of Ukraine (NBU), a Technical Assistance (TA) Project on Strengthening Macroeconomic Modeling in Support of FPAS at the NBU was launched by ICD in May 2024.
- Project focused on two complementary workstreams:
  - reviewing the NBU’s existing Quarterly Projection Model (QPM), further developing the QPM and strengthening modeling and forecasting capacity; and
  - reviewing existing FPAS processes and further integrating model-based forecasting and analysis into NBU’s policymaking and external communications.
- Engagement modalities:
  - A total of four in-person missions were conducted in FY25 (all held in Vienna, Austria), and regular virtual office hours in between the missions.
  - IMF TA project team: Natan Epstein, Iaroslav Miller, Daniel Baksa, Mariia Sydorovych (all ICDMM), and Karel Musil (short-term expert).

### Key institutional details
- Monetary Policy and Economic Analysis Department (MPEAD) implements the NBU FPAS and employs around 40 officials.
- MPEAD organizational components: Monetary Policy division; Macroeconomic Modeling and Forecasting division (MMFD); Economic Analysis division; Research unit.
- MMFD comprises QPM operators and is responsible for model maintenance and forecasting.
- NBU primary objective under its policy framework: inflation at 5% target; main instrument: the short-term policy rate.

### Progress, model development and QPM+
- By the end of the first year of the FPAS TA project the NBU modeling team made remarkable progress in capacity development, including in data management, macroeconomic modeling and coding.
- Existing QPM reviewed, recalibrated and extended to reflect wartime shocks and structural features.
- Key model extensions completed or progressed:
  - relative prices block to better interpret distinct core and non-core drivers of inflation;
  - more detailed monetary policy block to allow a more comprehensive assessment of policy trade-offs under Inflation Targeting with increase in exchange rate and capital flow management during the ongoing war;
  - decomposition of GDP into agricultural and non-agricultural sectors to better reflect their distinct implications for inflationary pressures;
  - revisions to external sector and labor blocks to reflect change in composition of trade partners and growing labor market pressures due to conscription and emigration.
- Revised QPM has been used as the core forecasting model by the NBU since early-2025 and was officially announced as QPM+ in the NBU’s April 2025 Inflation Report.
- Ongoing challenges: staffing shortages, staff turnover and key person risk.

### Executive Summary — Main achievements in year one
- Four in-person missions (Vienna) and regular virtual office hours were held starting in May 2024.
- Significant progress on QPM and modeling capacity:
  - QPM reviewed, recalibrated and extended to reflect wartime shocks and structural features.
  - Model extensions focused on relative prices, monetary policy block detail, GDP sectoral decomposition, external and labor sector revisions.
- NBU modeling and forecasting team advanced data management, coding and modeling capacity.

### Executive Summary — Remaining priorities and next steps
- Further model extensions to reflect direct effects of the balance of payments (BOP) and fiscal policy.
- Strengthen communications between the forecasting team and the Monetary Policy Committee (MPC) on development of initial conditions and preliminary forecast.
- Design and integrate a template for MPC initial conditions presentation and roadmap to formally include it in the forecasting calendar.
- Address staffing shortages and mitigate key person risk.

### Forecasting process and calendar (current practice)
- Forecasting round follows an internally pre-announced calendar that starts 4 weeks prior to the final monetary policy decision.
- Forecasting round sequence:
  - preparation of data, outlook for fiscal and foreign sectors, and CPI and GDP nowcasts (Economic Analysis division);
  - integration of inputs into the QPM by the modeling team to produce initial conditions assessment presented to MPEAD;
  - baseline forecast reviewed by the Deputy Governor overseeing monetary policy and presented to the Ministry of Finance and the Ministry of Economy;
  - preliminary baseline presented to the MPC by end of second week for review and discussion of alternative scenarios;
  - two weeks are available to incorporate MPC feedback into the final baseline and produce alternative scenarios presented to the MPC during the policy decision meeting.

### Identified FPAS practice weaknesses and recommendations
- Initial conditions presentation:
  - Missing standalone meeting on model-based initial conditions with the MPC; current first presentation focuses on preliminary baseline forecast.
  - Recommendation: introduce initial conditions presentation as standard practice (initial conditions, preliminary forecast, final baseline).
- Analytical toolkit and outputs:
  - Coding capacity focused on operating existing tools; development needed to build a richer analytical toolkit (e.g., multiple model outputs, equation decomposition).
  - NBU produced shock decomposition and trend-gap reports but did not produce richer reports explaining impacts of initial conditions and forecast assumptions on baseline projection.
- Discussions with MPC:
  - Modeling team should focus discussions with MPC on business cycle pressures (underlying trends and gaps) using model-based interpretations.
- FPAS coordination:
  - Recommendation to introduce an FPAS Coordinator role within MPEAD (possibly at MPEAD deputy director level) to organize and manage forecasting workflow, oversee modelers and sectoral experts, and communicate with senior NBU management.

### Potential model extensions requested by NBU
- Four additional blocks identified for QPM:
  - relative prices block to capture distinct long-term dynamics of CPI subcomponents and second round spillovers;
  - balance of payments (BoP) block (based on Berg et al (2023)) to capture pressures from external imbalances affecting domestic variables and policy decisions;
  - explicit fiscal block to better capture monetary-fiscal interactions and impact of debt management strategies on inflationary and exchange rate pressures;
  - credit sector block to better account for pressures stemming from bank lending activities.
- TA mission team devised a medium-term model plan with flexibility to accommodate evolving NBU needs.

### Adjusted Model Development Plan (milestones and timing)
- Development of relative price extension and regime switching in QPM. — 2024Q3 (completed)
- Launch development of the aggregated Fiscal block. — 2024Q4
- Launch Expenditure-side decomposition of GDP. — Spring 2025 (omitted)
- Launch the development of BoP extension. — Fall 2025
- Launch the development of a credit block. — Winter 2026

### Project design, delivery and FY2025 outcomes
- TA engagement planned as recurring over three years combining in-person TA missions and virtual support, including dedicated training on model-based monetary policy analysis and forecasting (MPAF).
- FY2025 outcomes:
  - TA team largely achieved model development objectives for FY2025.
  - Four in-person TA missions held in Vienna in FY2025.
  - Focus: review of existing NBU QPM, assessment of model properties and forecasting performance, recalibration to reflect wartime shocks.
  - Revised QPM used as core forecasting model since early-2025 and officially announced as QPM+ in April 2025 Inflation Report.

### Capacity building, coding, modeling and forecasting
- TA mission largely achieved objective of enhancing the NBU team’s coding, modeling and forecasting capacity to produce forward-looking policy recommendations.
- Engagement modalities:
  - In-person missions plus frequent virtual office hours for NBU staff to discuss coding and modeling exercises.
- Focus areas:
  - (i) enhancing coding skillset and modeling intuition of NBU staff;
  - (ii) strengthening data management procedures;
  - (iii) reviewing existing nowcasting and near-term forecasting (NTF).

### Staffing, turnover, resourcing and operational risk mitigation
- At project start, NBU modeling and forecasting team (within MMFD) included two model operators, one of whom was a team manager.
- During the first year:
  - One operator was promoted to a different role;
  - The other operator was promoted to lead the modeling team (both within MMFD).
- Staffing adjustments:
  - NBU hired one full-time operator;
  - Another economist joined the forecasting team on a part-time basis.
- Additional support requested:
  - NBU requested IMF to hire a peripatetic expert to further support forecast rounds and TA projects; the expert is expected to join the project in November 2025.
- Operational risk mitigation recommendations:
  - Increase number of model operators; consider internal mobility and new hires;
  - Ensure backup staff and document FPAS tools and processes;
  - Practice MMFD staff rotation to broaden skillsets;
  - Spread analytical knowledge and skills broadly and make use of documentation materials (including published Working Papers).

### Nowcasting, NTF tools and data constraints
- Existing nowcasting and near-term forecasting (NTF) tools: high quality but limited by data quality constraints.
- NBU analytical instruments:
  - Broad suite of empirical-based models for nowcasts and NTFs for key macroeconomic variables to inform medium-term forecasting.
  - Data-driven forecasts use high-frequency data (e.g., google trends analytics, night light intensity, online data on job vacancies and wages) that correlate well with current-quarter GDP dynamics.
- Limitations exacerbated by Russian invasion:
  - (i) higher volatility and publication lags for conventional data;
  - (ii) larger heterogeneity on regional level;
  - (iii) structural breaks and changes in correlations.

### QPM history, features, requested revisions and implementation status
- Regular medium-term macroeconomic projections produced using semi-structural QPM since 2016.
- QPM features:
  - New-Keynesian four-block core, along lines of Berg et al. (2006), adapted and extended for the Ukrainian economy.
  - Calibrated to match Inflation Targeting framework; policy regime shifted toward larger emphasis on exchange rate management following Russian invasion in 2022.
- TA-supported extensions and updates:
  - Monetary policy block elaborated to introduce an exchange rate volatility smoothing into the QPM, allowing parameter changes in modified UIP equation and policy reaction function and reflecting gradual adjustments back to pure Inflation Targeting.
  - Inflation block augmented to include relative prices to capture diverging drivers of four distinct inflation components: core, raw food, petrol and administrative prices.
  - Foreign economy block streamlined to focus on main trading partners and enhanced the euro’s role to reflect re-orientation toward Euro Area economies.
  - Existing fiscal and terms-of-trade blocks removed and to be replaced by new fiscal and BoP blocks later.
- NBU-requested revisions:
  - Extend QPM with labor block and decomposition of GDP into agricultural and non-agricultural sectors; postpone fiscal block development.
  - Rationale: capture pressures from large outward migration and associated wage and price pressures; interpret distinct inflationary pressures from agricultural sector performance (e.g., bumper harvest explaining simultaneous high GDP growth and low inflation).
- Expenditure-side GDP decomposition:
  - Will not be implemented due to data quality limits (low quality of sectoral GDP components, especially GFCF and inventories; difficulty reconciling business cycle based on overall GDP with sectoral components).
  - Recommendation by Dr. Nikolaychuk: omit this extension and prioritize BoP extension.

### Reporting infrastructure and data management improvements
- Developed reporting infrastructure to automatically compile:
  - (i) comprehensive impulse response report to assess model properties during calibration;
  - (ii) detailed historical interpretation report decomposing behavior of key model equations and illustrating drivers of business cycle dynamics and country fundamentals for initial conditions discussion;
  - (iii) forecast report structured like historical interpretation report but focused on economic outlook to help build forecasting narrative;
  - (iv) scenario builder script to assess forecast risks transparently.
- Data management improvements:
  - Overhauled infrastructure to eliminate risk of data pollution, ad-hoc processing solutions and to improve code readability and maintainability.

### Virtual office hours, junior modeler training and FPAS integration
- Regular virtual office hours enhanced coding practices and experimentation with model structures and calibrations.
- Emphasis on strengthening modeling intuition and selecting model versions suited to current monetary policy regime.
- Sustained engagement valuable due to addition of junior modelers to the NBU team.
- FPAS practices largely align with international standards; MMFD model-based analysis follows a pre-announced forecasting calendar and manages deliverables.
- MMFD critical tasks:
  - Identify and perform tasks related to quarterly forecasting cycles; integrate medium-term and near-term forecasts; identify risks and build alternative scenarios; compile macroeconomic forecasts by reaching collective views; draft materials for internal and external communications; present analysis and interact with management; build forecasting narrative based on model outputs and present policy recommendations consistent with monetary policy objectives.

### Forecasting calendar gaps and explicit recommendations
- Identified gaps and recommended actions:
  - (i) extend the forecast calendar with the MPC initial conditions presentation;
  - (ii) include more explicit model-based analysis of business cycle position into MPC presentations;
  - (iii) appoint an FPAS Coordinator to organize workflow, focus analytical discussions and communicate with senior management.
- Initial conditions presentation recommended as fundamental for building consistent narrative about current economic conditions and to obtain senior decision-makers’ feedback.
- Modeling team advised to elicit MPC input on underlying trends and gaps rather than limiting discussion to headline numbers.

### Forthcoming engagements and resident advisor
- Upcoming TA focus:
  - Model extensions for balance of payments (BoP) and fiscal blocks; prior decision: do not pursue expenditure-side GDP decomposition due to data limitations.
  - Continue developing MMFD model intuition and capacity for BoP and fiscal blocks; continue practicing preparation of shadow forecasts using QPM+.
  - Design template for an MPC initial conditions presentation and roadmap to integrate it into forecasting calendar.
- Macroeconomic Frameworks Resident Advisor:
  - Newly hired Resident Advisor will reside in Vienna and travel to Kyiv on quarterly frequency to join MMFD during forecasting rounds in-person.
  - Responsibilities: support FPAS, advise on model-based forecasting and analytical framework, re-design forecasting calendar, strengthen forecasting procedures and processes, assist MMFD staff with analysis and forecast preparations using QPM+.
  - Advisor will commence work in November 2025 and will join FPAS TA missions held in Vienna.
- Coordination with IMF Ukraine country team:
  - TA team remained in close contact with EUR Ukraine country team and will remain in close communication to ensure alignment with objectives under IMF lending program with Ukraine.

*Source: https://www.imf.org/-/media/files/publications/tar/2026/english/tarea2026010-source-pdf.pdf*

### Preface ................................................................................................................

### Preface

### Project background and scope
- At the request of the National Bank of Ukraine (NBU), a Technical Assistance (TA) Project on Strengthening Macroeconomic Modeling in Support of FPAS at the NBU was launched by ICD in May 2024.
- The project focused on two broad and complementary workstreams:
  - reviewing the NBU’s existing Quarterly Projection Model (QPM), further developing the QPM and strengthening modeling and forecasting capacity; and
  - reviewing the existing FPAS processes and further integration of model-based forecasting and analysis into NBU’s policymaking and external communications.
- A total of four in-person missions were conducted in FY25 (all held in Vienna, Austria), and regular virtual office hours in between the missions.
- The IMF TA project team consists of Natan Epstein, Iaroslav Miller, Daniel Baksa, Mariia Sydorovych (all ICDMM), and Karel Musil (short-term expert).

### Key institutional details
- The Monetary Policy and Economic Analysis Department (MPEAD) implements the NBU FPAS and employs around 40 officials.
- MPEAD organizational components: Monetary Policy division; Macroeconomic Modeling and Forecasting division (MMFD); Economic Analysis division; Research unit.
- The Macroeconomic Modeling and Forecasting division (MMFD) comprises QPM operators and is responsible for model maintenance and forecasting.
- The NBU’s primary objective under its policy framework is inflation at 5% target; the main instrument is the short-term policy rate.

### Progress and constraints
- By the end of the first year of the FPAS TA project the NBU modeling team made remarkable progress in capacity development, including in data management, macroeconomic modeling and coding.
- The existing QPM model has been reviewed, recalibrated and extended to better reflect key features of the Ukrainian economy and improve analysis of policy trade-offs in an environment of exceptionally high uncertainty due to the ongoing Russian invasion.
- Key model extensions completed or progressed:
  - relative prices block to better interpret distinct core and non-core drivers of inflation;
  - more detailed monetary policy block to allow a more comprehensive assessment of policy trade-offs under Inflation Targeting with increase in exchange rate and capital flow management during the ongoing war;
  - decomposition of GDP into agricultural and non-agricultural sectors to better reflect their distinct implications for inflationary pressures;
  - revisions to external sector and labor blocks to reflect change in composition of trade partners and growing labor market pressures due to conscription and emigration.
- The updated version of the QPM—called QPM+—has been used as a core forecasting model since early-2025 to inform MPC decision making on the key policy rate.
- Ongoing challenges include staffing shortages, staff turnover and key person risk.

---

### Executive Summary

### Main achievements in year one
- Four in-person missions (Vienna) and regular virtual office hours were held starting in May 2024.
- Significant progress along the QPM and modeling capacity development workstream:
  - QPM reviewed, recalibrated and extended to reflect wartime shocks and structural features.
  - Model extensions focused on relative prices, monetary policy block detail, GDP sectoral decomposition, external and labor sector revisions.
- NBU modeling and forecasting team advanced data management, coding and modeling capacity.

### Remaining priorities and next steps
- Further model extensions to reflect direct effects of the balance of payments (BOP) and fiscal policy.
- Strengthening communications between the forecasting team and the Monetary Policy Committee (MPC) on development of initial conditions and preliminary forecast.
- Design and integrate a template for MPC initial conditions presentation and roadmap to formally include it in the forecasting calendar.
- Address staffing shortages and mitigate key person risk.

---

### Overview of the first year — Scoping mission findings

### Forecasting process and calendar
- Forecasting round follows an internally pre-announced calendar that starts 4 weeks prior to the final monetary policy decision.
- The forecasting round includes:
  - preparation of data, outlook for fiscal and foreign sectors, and CPI and GDP nowcasts (Economic Analysis division);
  - integration of inputs into the QPM by the modeling team to produce initial conditions assessment presented to MPEAD;
  - baseline forecast reviewed by the Deputy Governor overseeing monetary policy and presented to the Ministry of Finance and the Ministry of Economy;
  - preliminary baseline presented to the MPC by end of second week for review and discussion of alternative scenarios;
  - two weeks are available to incorporate MPC feedback into the final baseline and produce alternative scenarios presented to the MPC during the policy decision meeting.

### Identified FPAS practice weaknesses
- Initial conditions presentation:
  - A standalone meeting to discuss model-based initial conditions with the MPC is missing from the forecast calendar, leading to relatively less attention on cyclical position and trend-gap decomposition.
  - Under standard FPAS, three presentations to the MPC during every forecasting round are typical: initial conditions, preliminary forecast, and final baseline forecast.
- Analytical toolkit and outputs:
  - NBU coding capacity focused on operating existing tools and needed development to build a richer analytical toolkit.
  - The NBU team focused analytical discussions mainly on shock decomposition reports or reports on underlying trends and gaps, and did not produce richer reports explaining the impact of initial conditions and forecast assumptions on baseline projection.
  - Standard FPAS practice uses multiple model outputs (e.g., equation decomposition) to assess initial conditions and build consistent economic narratives.
- Discussions with MPC:
  - The modeling team could benefit from focusing discussions with the MPC around business cycle pressures (underlying trends and gaps dynamics) using model-based interpretations.
- FPAS coordination:
  - Introduction of an FPAS Coordinator role within MPEAD was recommended to organize and manage workflow during forecasting rounds; such role is currently missing.
  - The FPAS Coordinator could be at the level of the MPEAD deputy director to oversee modelers and sectoral experts and communicate with senior NBU management.

### Potential model extensions requested by NBU
- Four additional blocks identified for the QPM:
  - relative prices block to capture distinct long-term dynamics of CPI subcomponents and second round spillovers;
  - balance of payments (BoP) block (based on Berg et al (2023)) to capture pressures from external imbalances affecting domestic variables and policy decisions;
  - explicit fiscal block to better capture monetary-fiscal interactions and impact of debt management strategies on inflationary and exchange rate pressures;
  - credit sector block to better account for pressures stemming from bank lending activities.
- The TA mission team devised a medium-term model plan with flexibility to accommodate evolving NBU needs.

---

### Adjusted Model Development Plan (milestones and timing)
- Development of relative price extension and regime switching in QPM. — 2024Q3 (completed)
- Launch development of the aggregated Fiscal block. — 2024Q4
- Launch Expenditure-side decomposition of GDP. — Spring 2025 (omitted)
- Launch the development of BoP extension. — Fall 2025
- Launch the development of a credit block. — Winter 2026

---

### Objectives and status of the NBU FPAS TA Project

### Project design and delivery
- At the end of the Scoping mission, the TA team proposed a recurring engagement over three years combining in-person TA missions and virtual support, including dedicated training on model-based monetary policy analysis and forecasting (MPAF).
- The project concentrated on:
  - reviewing and extending the existing QPM and strengthening modeling and forecasting capacity; and
  - reviewing existing NBU FPAS processes and further integrating model-based analytical work into NBU policy processes.

### FY2025 outcomes
- The TA team largely achieved its model development objectives for FY2025.
- Four in-person TA missions were held in Vienna in FY2025.
- Focus areas: review of the existing NBU QPM, assessment of model properties and forecasting performance, recalibration to improve forecasting performance and to reflect wartime shocks.
- The revised QPM has been used as the core forecasting model by the NBU since early 2025 and was officially announced as QPM+ in the NBU’s April 2025 Inflation Report.

*Source: https://www.imf.org/-/media/files/publications/tar/2026/english/tarea2026010-source-pdf.pdf*

### 9. The TA mission team also largely achieved its objective of enhancing the NBU team’s coding,

### 9. The TA mission team also largely achieved its objective of enhancing the NBU team’s coding,

### Capacity building, coding, modeling and forecasting
- TA mission largely achieved objective of enhancing the NBU team’s coding, modeling and forecasting capacity to produce forward-looking policy recommendations.
- Engagement modalities:
  - In-person mission plus frequent virtual office hours for NBU staff to discuss coding and modeling exercises.
- Focus areas of exercises:
  - (i) enhancing the coding skillset and modeling intuition of the NBU staff;
  - (ii) strengthening data management procedures;
  - (iii) reviewing existing nowcasting and near-term forecasting (NTF).

### Staffing, turnover and resourcing
- At project start, NBU modeling and forecasting team (within the MMFD) included two model operators, one of whom was a team manager.
- During the first year of the project:
  - One operator was promoted to a different role;
  - The other operator was promoted to lead the modeling team (both within MMFD).
- Staffing adjustments:
  - NBU hired one full-time operator;
  - Another economist joined the forecasting team on a part-time basis.
- Additional support requested:
  - NBU requested IMF to hire a peripatetic expert to further support forecast rounds and TA projects; the expert is expected to join the project in November 2025.
- Operational risk mitigation discussed with NBU:
  - Amount of model operators is insufficient → extend it by hiring more staff; consider both internal mobility and new hires options;
  - Risk of excessive turnover → make sure there are backup staff; document FPAS tools and processes; practice MMFD staff rotation to broaden the skillset across individual staff;
  - Key person risks → make sure analytical knowledge and skills are spread more broadly and there are backup positions; make use of documentation materials (including published Working Papers) to initiate new staff.

### Nowcasting, NTF tools and data constraints
- Existing nowcasting and near-term forecasting (NTF) tools are of high quality but limited by data quality constraints.
- NBU analytical instruments:
  - Broad suite of empirical-based models for producing nowcasts and NTFs for key macroeconomic variables to inform medium-term forecasting.
  - Data-driven forecasts explore high-frequency data (e.g., google trends analytics, night light intensity, online data on job vacancies and wages) that correlate well with current-quarter GDP dynamics.
- Limitations exacerbated by Russian invasion:
  - (i) higher volatility and publication lags for conventional data;
  - (ii) larger heterogeneity on regional level;
  - (iii) structural breaks and changes in correlations.

### QPM history, extensions and QPM+
- Regular medium-term macroeconomic projections produced using semi-structural QPM since 2016.
- QPM features:
  - New-Keynesian four-block core, along lines of Berg et al. (2006), adapted and extended for the Ukrainian economy.
  - Calibrated to match Inflation Targeting framework; policy regime shifted toward larger emphasis on exchange rate management following Russian invasion in 2022.
- TA-supported extensions and updates:
  - Monetary policy block elaborated to introduce an exchange rate volatility smoothing into the QPM, allowing parameter changes in modified UIP equation and policy reaction function and reflecting gradual adjustments back to pure Inflation Targeting.
  - Inflation block augmented to include relative prices to capture diverging drivers of four distinct inflation components: core, raw food, petrol and administrative prices.
  - Foreign economy block streamlined to focus on main trading partners and enhanced the euro’s role to reflect re-orientation toward Euro Area economies.
  - Existing fiscal and terms-of-trade blocks removed and will be replaced by new versions of fiscal and BoP blocks later.
- NBU-requested revisions:
  - NBU requested extending QPM with labor block and decomposition of GDP into agricultural and non-agricultural sectors, postponing fiscal block development.
  - Rationale: capture pressures from large outward migration and associated wage and price pressures; interpret distinct inflationary pressures from agricultural sector performance (e.g., bumper harvest explaining simultaneous high GDP growth and low inflation).
- Implementation status:
  - Revised QPM has been used as core forecasting model by the NBU since early-2025.
  - Officially introduced as QPM+ in the NBU’s April 2025 Inflation Report (see Figure 3).
- Expenditure-side GDP decomposition:
  - Will not be implemented due to data quality limits (low quality of sectoral GDP components, especially GFCF and inventories; difficulty reconciling business cycle based on overall GDP with sectoral components).
  - Recommendation by Dr. Nikolaychuk: omit this extension and prioritize BoP extension.

### Reporting infrastructure and data management
- Mission assisted with development of reporting infrastructure and streamlining data management procedures.
- NBU expanded reporting infrastructure to automatically compile:
  - (i) comprehensive impulse response report to assess model properties during calibration;
  - (ii) detailed historical interpretation report describing dynamics of key variables, decomposing behavior of key model equations and illustrating drivers of business cycle dynamics and country fundamentals for initial conditions discussion;
  - (iii) forecast report structured like historical interpretation report but focused on economic outlook to help build forecasting narrative;
  - (iv) scenario builder script to assess forecast risks transparently.
- Data management improvements:
  - Overhauled infrastructure to eliminate risk of data pollution, ad-hoc processing solutions and to improve code readability and maintainability.

### Virtual office hours and junior modeler training
- Regular virtual office hours provided supplementary time to enhance coding practices and experiment with model structures and calibrations.
- Particular emphasis on strengthening modeling intuition and finding model versions best suited to reflect current monetary policy regime.
- Sustained engagement was particularly valuable due to addition of junior modelers to the NBU team.

### Integration of FPAS into internal processes and communications
- NBU FPAS practices largely align with international standards.
- MMFD model-based analysis follows a pre-announced forecasting calendar and manages deadlines and responsibilities for deliverables.
- Critical MMFD tasks:
  - Identify and perform tasks related to quarterly forecasting cycles;
  - Integrate medium-term and near-term forecasts, identify risks, build alternative scenarios;
  - Compile macroeconomic forecasts by reaching collective and consensual views;
  - Draft materials destined for internal and external communications;
  - Present analysis and interact on them with management;
  - Build a forecast narrative based on model outputs and present policy recommendations consistent with monetary policy objectives.
- MMFD organization:
  - MMFD includes two full-time model operators (one is head of the modeling unit) who maintain, use and develop QPM infrastructure and produce model-based analysis.
  - A separate division supplies sectoral expertise and forecast judgements (e.g., inflation NTFs, agricultural outlook, fiscal outlook, external sector outlook).

### Forecasting calendar, workflow and identified gaps
- Forecasting process and analytical workflow occur on quarterly basis; NBU policymaking process is conducted over four MPC sessions per year (out of eight).
- Practical needs for forecasting rounds:
  - Timing of meetings, deadlines, responsibilities and deliverables;
  - Tasks performed by staff, structure to overall workflow and internal communications (horizontally and vertically);
  - Coordination within MMFD and with management.
- Identified gaps in FPAS practices and recommendations:
  - (i) extend the forecast calendar with the MPC initial conditions presentation;
  - (ii) include more explicit model-based analysis of business cycle position into MPC presentations;
  - (iii) appoint an FPAS Coordinator.
- Initial conditions presentation:
  - Missing at NBU; current first presentation to MPC focuses on preliminary baseline forecast rather than initial conditions.
  - TA mission view: initial conditions presentation is fundamental for building consistent narrative about current economic conditions and to obtain senior decision-makers’ feedback.
- Interaction with MPC:
  - Modeling team limits discussion to headline numbers rather than eliciting MPC input on underlying trends and gaps.
  - Standard FPAS practice calls for discussion of business cycle position and inflationary pressures using model-based interpretation to build consistent economic narrative and anchor MPC on key inflation drivers.
- FPAS Coordinator:
  - Missing in MPEAD; recommended to establish position (possibly at MPEAD deputy director level) to organize workflow during forecasting round, identify key forecast issues, focus analytical discussions between modelers and sectoral experts, and communicate with senior NBU management.

### Forthcoming engagements and resident advisor
- Upcoming TA missions will focus on model extensions for balance of payments (BoP) and fiscal blocks.
  - Prior decision: do not pursue expenditure-side GDP decomposition due to data limitations; focus modeling TA on BoP and fiscal blocks to capture pressures from external imbalances and monetary-fiscal interactions.
  - TA will continue developing MMFD model intuition and capacity to analyze additional variables and policy interactions that will be reflected in BoP and fiscal blocks.
  - TA will continue practicing preparation of shadow forecasts using QPM+ with the NBU team.
  - TA will design a template for an MPC initial conditions presentation and roadmap to integrate it into forecasting calendar.
- Macroeconomic Frameworks Resident Advisor:
  - Newly hired Resident Advisor will reside in Vienna and travel to Kyiv on quarterly frequency to join MMFD during forecasting rounds in-person.
  - Responsibilities: support FPAS, advise on model-based forecasting and analytical framework, re-design forecasting calendar to incorporate enhanced presentation development and delivery, strengthen forecasting procedures and processes, assist MMFD staff with analysis and forecast preparations using QPM+.
  - Advisor will commence work in November 2025 and will join FPAS TA missions held in Vienna.
- Coordination with IMF Ukraine country team:
  - TA team remained in close contact with EUR Ukraine country team; communicated prior to every mission and briefed outcomes after each mission.
  - Scheduling of TA missions informed by country team visits and engagements, including in context of ongoing 4-year IMF Extended Fund Facility arrangement.
  - TA team will remain in close communication and cooperation with IMF country team to ensure TA recommendations align with objectives under IMF lending program with Ukraine.

*IMF | Technical Report – NBU TA on Developing FPAS at the NBU*

---


_Source: https://www.imf.org/-/media/files/publications/tar/2026/english/tarea2026010-source-pdf.pdf_