## Section I. Detailed Technical Assessment and Recommendations ## Source details **Canonical URL:** [Section I. Detailed Technical Assessment and Recommendations](https://www.imf.org/-/media/files/publications/tar/2024/english/tarea2024085-print-pdf.pdf) ## Other formats - [Markdown version](/-/media/files/publications/tar/2024/english/tarea2024085-print-pdf.pdf.md) - [Structured JSON version](/-/media/files/publications/tar/2024/english/tarea2024085-print-pdf.pdf.json) --- ### A. Introduction - TA mission conducted from July 15–19, 2024 to assist authorities in Ukraine to develop new processes and methods for the compilation of the House Price Index (HPI). This was the second mission of a project that commenced in April 2024. - Mission objectives: - Develop R scripts to clean listings data received from an online real estate platform. - Implement updated methods for index compilation. - Increase capacity of SSSU staff in methods and tools. - Mission location: Joint Vienna Institute (JVI) in Austria due to the ongoing war. - Risk note: Ukraine under martial law since early 2022; uncertainty over the war creates a significant risk to the workplan. ### B. Data Sources — findings and recommended actions Findings - Two data sources used for the HPI: - New apartments: survey data collected by SSSU regional offices (from real estate agents and property developers). - Existing apartments: online listings from OLX (OLX agreed to provide data directly to SSSU on a quarterly basis). - Current data processing: - OLX raw data are currently processed and filtered using MS Excel; each quarter this takes about two weeks. - List prices are almost exclusively denominated in USD; for 2024 Q2 only 1 percent of observations on OLX are denominated in a currency other than USD. Foreign currency prices are converted to Ukrainian hryvnia using the prevailing daily exchange rate; HPI calculated using local currency prices. - OLX data are filtered heavily; typical filters include removal of small towns/rural areas, apartments with >3 rooms, allowable floor areas by room count, allowable price range, removal of new apartments, removal of “construction in progress”, removal of single-story dwellings, and retention of observations in the middle month of the quarter. - Sample sizes (Number New and Existing Apartments per Quarter, 2023 Q1 to 2024 Q2): - 2023 Q1: New apartments 893; Existing apartments 24,278 - 2023 Q2: New apartments 926; Existing apartments 30,070 - 2023 Q3: New apartments 925; Existing apartments 33,703 - 2023 Q4: New apartments 912; Existing apartments 33,529 - 2024 Q1: New apartments 923; Existing apartments 35,556 - 2024 Q2: New apartments 1873; Existing apartments 31,417 - On the OLX file for 2024 Q2, there are just over 10,000 existing apartments categorized as “construction in progress”. - SSSU expanded data collection for new apartments in 2024 Q2 (from about 900 observations per month to over 1,800), but unusual price movements were evident for the latest quarter (e.g., median prices significantly increased in one NUTS region). Operational notes on filtering (filters preserved) - Filter 1: Apartments in small towns and rural areas are removed, leaving just larger cities. - Filter 2: Apartments with greater than three rooms are removed. - Filter 3: Allowable floor area depending on number of rooms: - If number of rooms is 1, then floor area should be between 30 and 100 square meters. - If number of rooms is 2, then floor area should be between 40 and 180 square meters. - If number of rooms is 3, then floor area should be between 50 and 250 square meters. - Filter 4: Allowable range for price in local currency is between 200,000 and 50,000,000. - Filter 5: New apartments are removed, leaving just existing apartments. - Filter 6: Apartments categorized as “construction in progress” are removed. - Filter 7: Single story dwellings are removed, leaving only apartments in buildings with more than one floor. - Filter 8: Observations in the middle month of the quarter are retained e.g. for Q2, May is retained and the other months are removed. - Duplicate properties removed within a quarter; duplicates not removed across quarters. Specific data investigations required - Determine treatment of apartments “transacted during the construction process or even before construction has begun”. - Consult Eurostat to clarify definition of new properties to meet relevant EU statistical legislation. - Organize meeting with OLX to: - Investigate extending data to include houses (current OLX file only includes apartments). - Collect more detailed information on the location of properties for sale. - Apply R scripts to historical OLX files once raw files for previous quarters are located. Recommended actions (explicit) - Identify appropriate IT supports for staff to upskill in the R statistical package. - Locate the raw data files received from OLX for each quarter from 2023 Q1. - Investigate the unusual price changes for 2024 Q2 for new apartments. - Implement the new R scripts for the regular processing and cleaning of the OLX data. - Gather more information on apartments categorized as “construction in progress” to decide how they should be treated during the data processing. - Relax the data filtering over time to expand the coverage of the HPI. - Experiment with using the OLX data, instead of survey data, for new apartments. - Extend the data coverage to include houses. - Collect more detailed information on the location of the apartments for sale. - Apply R scripts to historical OLX files once raw files are located. ### C. Compilation Methods — findings, experimental methods, and workplan actions Findings - In May 2024, SSSU published an improved HPI for reference quarter Q1 2024 with main changes: - New data source for existing apartments (OLX) with more detailed geographical coverage. - Use of list prices instead of transaction prices. - Improved compilation methods: quality adjustment carried out using regression-based (hedonic) methods instead of stratified means. - The current SSSU method is a non-standard approach; more optimal hedonic methods exist and should be applied to better follow international standards. - New stratification developed using broad NUTS regional classification; real estate market split into sixteen strata: eight NUTS regions (Kyiv City separated) further split into new and existing apartments. - Experimental quality-adjusted hedonic indices for the strata were compiled during the mission. - Regional results: western regions experiencing significant price increases; war-impacted eastern regions show little growth or falling prices. - A compilation spreadsheet in MS Excel was designed to aggregate stratum-level indices using expenditure weights (weights calculated as sum of expenditure for calendar year 2023). - National HPI for Ukraine shows price change of about 15 percent over six quarters from 2023 Q1 to 2024 Q2. Workplan and methodological milestones (selected) - Data sources for the RPPI: - H Identify appropriate IT supports for staff to upskill in the R statistical package. — 30 Sep, 2024 - H Locate the raw data files received from OLX for each quarter from 2023 Q1. — 30 Sep, 2024 - H Investigate the unusual price changes for 2024 Q2 for new apartments. — 30 Sep, 2024 - H Implement the new R scripts for the regular processing and cleaning of the OLX data. — 31 Dec, 2024 - M Gather more information on apartments categorized as “construction in progress” to decide how they should be treated during the data processing. — 31 Dec, 2024 - M Relax the data filtering over time to expand the coverage of the HPI. — 30 Jun, 2025 - H Experiment with using the OLX data, instead of survey data, for new apartments. — 30 Jun, 2025 - M Extend the data coverage to include houses. — 31 Dec, 2025 - M Collect more detailed information on the location of the apartments for sale. — 31 Dec, 2025 - Compilation methods: - M Implement the time-dummy hedonic (TDH) method for index calculation. — 31 Mar, 2025 - M Once familiar with the TDH method, extend the method by using a rolling window. — 31 Mar, 2025 - M Use Cook’s Distance for outlier detection. — 31 Mar, 2025 - M Use the OLX data for the expenditure weights. — 31 Mar, 2025 - M Practice using the compilation spreadsheet for aggregating the stratum-level price indices. — 31 Mar, 2025 - M Become familiar with the requirements under relevant EU statistical legislation. — 31 Mar, 2025 Key methodological recommendations - Implement more optimal hedonic methods aligned with international standards (e.g., time-dummy hedonic method). - Use rolling-window extensions once staff are familiar with TDH. - Implement Cook’s Distance for outlier detection. - Use OLX data for constructing expenditure weights. - Ensure staff ownership of R scripts: SSSU staff to amend and update scripts after upskilling. Key statistics and data counts (Number of Observations per Strata for New and Existing Apartments for All Quarters, 2023 Q1 to 2024 Q2) - UA1 (Poltavska Oblast, Sumska Oblast, Kharkivska Oblast, Chernihivska Oblast): New apartments 935; Existing apartments 26,844 - UA3 (Dnipropetrovska Oblast, Zaporizka Oblast, Kirovohradska Oblast): New apartments 808; Existing apartments 22,627 - UA4 (Mykolaivska Oblast, Odeska Oblast, Khersonska Oblast): New apartments 698; Existing apartments 49,009 - UA5 (Vinnytska Oblast, Ternopilska Oblast, Khmelnytska Oblast): New apartments 540; Existing apartments 9,194 - UA6 (Kyivska Oblast, Cherkaska Oblast): New apartments 818; Existing apartments 14,515 - UA63 (M. Kyiv): New apartments 1,050; Existing apartments 48,595 - UA7 (Zakarpatska Oblast, Ivano-Frankivska Oblast, Lvivska Oblast, Chernivetska Oblast): New apartments 1,056; Existing apartments 12,822 - UA8 (Volynska Oblast, Zhytomyrska Oblast, Rivnenska Oblast): New apartments 547; Existing apartments 4,898 Aggregation and national index - Expenditure weights calculated as sum of expenditure (i.e., price in local currency) for calendar year 2023. - National HPI shows price change of about 15 percent over six quarters from 2023 Q1 to 2024 Q2. ### Experimental hedonic indices for existing apartments Model and estimation - RPPIs for existing apartments were calculated using the hedonic (time-dummy) method. - New apartments were excluded temporarily while the SSSU examined the unusual price movements for 2024 Q2. - A single index was calculated for each strata for existing apartments, with the same model applied for each. - Model specification: - Dependent variable: log of price (in local currency). - Independent variables: (i) floor area, (ii) number of rooms, and (iii) region. - Region variable was excluded from the Kyiv model due to lack of additional location information within Kyiv. Outcomes and data treatment - The authorities will require additional assistance to refine the hedonic models used for each stratum, and to implement the rolling-window approach. - Regional outcomes observed: - Regions in the west of the country are experiencing significant price increases. - Regions most impacted by the war in the east of the country have shown little growth or even falling prices. - Data exclusions: - Data exclude the territories which are temporarily occupied by the Russian Federation and part of territories where the military actions are/were conducted. - Outlier detection and treatment: - Outliers were removed at the level of the strata using Cook’s Distance. - National HPI result: - The national HPI for Ukraine is showing price change of about 15 percent over the six quarters from 2023 Q1 to 2024 Q2. Recommended actions (methodology and compilation) - Implement the time-dummy hedonic (TDH) method for index calculation. - Once familiar with the TDH method, extend the method by using a rolling window. - Use Cook’s Distance for outlier detection. - Use the OLX data for the expenditure weights. - Practice using the compilation spreadsheet for aggregating the stratum-level price indices. - Become familiar with the requirements under relevant EU statistical legislation. ### D. Dissemination — status and recommended next steps Findings and status - Several dissemination items remain contingent on the conclusion of the war. - Current dissemination items flagged as low priority (“L”) in the workplan until the conclusion of the war. Recommended actions (workplan) - L Review the revisions policy for the HPI after the conclusion of the war. — To be decided - L Update the structure of the published HPI data after the conclusion of the war. — To be decided - L Update the methods document after the conclusion of the war. — To be decided - L Publish the updated HPI on the authorities’ website after the conclusion of the war. — To be decided Procedural requirement - The SSSU are required to release methodological details of any new or updated publications as part of an annual plan of statistics. This annual plan is published in March every year. - The methods document will need to be released some time in advance of the publication of the data and should be included in planning for any update of the HPI for Ukraine. ### E. Capacity building and officials met Capacity building - Mission focused on increasing SSSU capacity: development of R scripts, practical training on hedonic methods, and provision of initial R scripts to staff. - Recommendation: SSSU staff should upskill on the R statistical package through identified supports (online or in-person training courses, internal IT support, or other modes). - SSSU staff ran the R scripts during the mission and were able to export clean data files to the network; staff ownership of scripts is crucial. Officials met - Nataliia Lysenko — Head of Division for Housing and Construction Price Statistics, Department for Price Statistics, SSSU - Victoriia Isaiuk — Senior Specialist of the Division for Construction and Housing Price Statistics, Department for Price Statistics, SSSU - Nadiia Profatska — Deputy Director-Head of Division for Consumer Price Statistics, Department for Price Statistic, SSSU - Hanna Buhaichuk — Deputy Head of the Division for Producer of industrial products Price Statistics, Department for Price Statistics, SSSU Appendix references - Appendix contains box plot analysis of price of new properties over time highlighting the 2024 Q2 median increase in at least one NUTS region. - Appendix figure referenced: FIGURE 3. Box Plots by Quarter for Log of Price, New Properties in Region UA1, 2023Q1–2024Q2. *IMF Technical Assistance Report – Ukraine RPPI Mission, Section I. Detailed Technical Assessment and Recommendations* ### Section I. Detailed Technical Assessment and Recommendations ...................................................... 4 ### Section I. Detailed Technical Assessment and Recommendations ### A. Introduction - TA mission conducted from July 15–19, 2024 to assist authorities in Ukraine to develop new processes and methods for the compilation of the House Price Index (HPI). This was the second mission of a project that commenced in April 2024. - Mission objectives: - Develop R scripts to clean listings data received from an online real estate platform. - Implement updated methods for index compilation. - Increase capacity of SSSU staff in methods and tools. - Mission location: Joint Vienna Institute (JVI) in Austria due to the ongoing war. - Risk note: Ukraine under martial law since early 2022; uncertainty over the war creates a significant risk to the workplan. ### B. Data Sources — findings and recommended actions Findings - Two data sources used for the HPI: - New apartments: survey data collected by SSSU regional offices (from real estate agents and property developers). - Existing apartments: online listings from OLX (OLX agreed to provide data directly to SSSU on a quarterly basis). - Current data processing: - OLX raw data are currently processed and filtered using MS Excel; each quarter this takes about two weeks. - List prices are almost exclusively denominated in USD; for 2024 Q2 only 1 percent of observations on OLX are denominated in a currency other than USD. Foreign currency prices are converted to Ukrainian hryvnia using the prevailing daily exchange rate; HPI calculated using local currency prices. - OLX data are filtered heavily; typical filters (Table 3) include removal of small towns/rural areas, apartments with >3 rooms, allowable floor areas by room count, allowable price range, removal of new apartments, removal of “construction in progress”, removal of single-story dwellings, and retention of observations in the middle month of the quarter. - Sample sizes (Table 4: Number New and Existing Apartments per Quarter, 2023 Q1 to 2024 Q2): - 2023 Q1: New apartments 893; Existing apartments 24,278 - 2023 Q2: New apartments 926; Existing apartments 30,070 - 2023 Q3: New apartments 925; Existing apartments 33,703 - 2023 Q4: New apartments 912; Existing apartments 33,529 - 2024 Q1: New apartments 923; Existing apartments 35,556 - 2024 Q2: New apartments 1873; Existing apartments 31,417 - On the OLX file for 2024 Q2, there are just over 10,000 existing apartments categorized as “construction in progress”. - SSSU expanded data collection for new apartments in 2024 Q2 (from about 900 observations per month to over 1,800), but unusual price movements were evident for the latest quarter (e.g., median prices significantly increased in one NUTS region). Recommended actions (explicit) - Identify appropriate IT supports for staff to upskill in the R statistical package. - Locate the raw data files received from OLX for each quarter from 2023 Q1. - Investigate the unusual price changes for 2024 Q2 for new apartments. - Implement the new R scripts for the regular processing and cleaning of the OLX data. - Gather more information on apartments categorized as “construction in progress” to decide how they should be treated during the data processing. - Relax the data filtering over time to expand the coverage of the HPI. - Experiment with using the OLX data, instead of survey data, for new apartments. - Extend the data coverage to include houses. - Collect more detailed information on the location of the apartments for sale. Operational notes on filtering (Table 3 filters preserved) - Filter 1: Apartments in small towns and rural areas are removed, leaving just larger cities. - Filter 2: Apartments with greater than three rooms are removed. - Filter 3: Allowable floor area depending on number of rooms: - If number of rooms is 1, then floor area should be between 30 and 100 square meters. - If number of rooms is 2, then floor area should be between 40 and 180 square meters. - If number of rooms is 3, then floor area should be between 50 and 250 square meters. - Filter 4: Allowable range for price in local currency is between 200,000 and 50,000,000. - Filter 5: New apartments are removed, leaving just existing apartments. - Filter 6: Apartments categorized as “construction in progress” are removed. - Filter 7: Single story dwellings are removed, leaving only apartments in buildings with more than one floor. - Filter 8: Observations in the middle month of the quarter are retained e.g. for Q2, May is retained and the other months are removed. - Duplicate properties removed within a quarter; duplicates not removed across quarters. Specific data investigations required - Determine treatment of apartments “transacted during the construction process or even before construction has begun” (common in Ukraine). - Consult Eurostat to clarify definition of new properties to meet relevant EU statistical legislation. - Organize meeting with OLX to: - Investigate extending data to include houses (current OLX file only includes apartments). - Collect more detailed information on the location of properties for sale. - Apply R scripts to historical OLX files once raw files for previous quarters are located. ### C. Compilation Methods — findings, experimental methods, and workplan actions Findings - In May 2024, SSSU published an improved HPI for reference quarter Q1 2024 with main changes: - New data source for existing apartments (OLX) with more detailed geographical coverage. - Use of list prices instead of transaction prices. - Improved compilation methods: quality adjustment carried out using regression-based (hedonic) methods instead of stratified means. - The current SSSU method is a non-standard approach; more optimal hedonic methods exist and should be applied to better follow international standards. - New stratification developed using broad NUTS regional classification; real estate market split into sixteen strata: eight NUTS regions (Kyiv City separated) further split into new and existing apartments. - Experimental quality-adjusted hedonic indices for the strata were compiled during the mission. - Regional results: western regions experiencing significant price increases; war-impacted eastern regions show little growth or falling prices. - A compilation spreadsheet in MS Excel was designed to aggregate stratum-level indices using expenditure weights (weights calculated as sum of expenditure for calendar year 2023). - National HPI for Ukraine shows price change of about 15 percent over six quarters from 2023 Q1 to 2024 Q2. Workplan and methodological milestones (from Table 2 — Priority Action/Milestone and Target Completion Date) Topic: Data sources for the RPPI - H Identify appropriate IT supports for staff to upskill in the R statistical package. — 30 Sep, 2024 - H Locate the raw data files received from OLX for each quarter from 2023 Q1. — 30 Sep, 2024 - H Investigate the unusual price changes for 2024 Q2 for new apartments. — 30 Sep, 2024 - H Implement the new R scripts for the regular processing and cleaning of the OLX data. — 31 Dec, 2024 - M Gather more information on apartments categorized as “construction in progress” to decide how they should be treated during the data processing. — 31 Dec, 2024 - M Relax the data filtering over time to expand the coverage of the HPI. — 30 Jun, 2025 - H Experiment with using the OLX data, instead of survey data, for new apartments. — 30 Jun, 2025 - M Extend the data coverage to include houses. — 31 Dec, 2025 - M Collect more detailed information on the location of the apartments for sale. — 31 Dec, 2025 Topic: Compilation methods - M Implement the time-dummy hedonic (TDH) method for index calculation. — 31 Mar, 2025 - M Once familiar with the TDH method, extend the method by using a rolling window. — 31 Mar, 2025 - M Use Cook’s Distance for outlier detection. — 31 Mar, 2025 - M Use the OLX data for the expenditure weights. — 31 Mar, 2025 - M Practice using the compilation spreadsheet for aggregating the stratum-level price indices. — 31 Mar, 2025 - M Become familiar with the requirements under relevant EU statistical legislation. — 31 Mar, 2025 Key methodological recommendations - Implement more optimal hedonic methods aligned with international standards (e.g., time-dummy hedonic method). - Use rolling-window extensions once staff are familiar with TDH. - Implement Cook’s Distance for outlier detection. - Use OLX data for constructing expenditure weights. - Ensure staff ownership of R scripts: SSSU staff to amend and update scripts after upskilling. Key statistics and data counts (Table 5 excerpt — Number of Observations per Strata for New and Existing Apartments for All Quarters, 2023 Q1 to 2024 Q2) - UA1 (Poltavska Oblast, Sumska Oblast, Kharkivska Oblast, Chernihivska Oblast): New apartments 935; Existing apartments 26,844 - UA3 (Dnipropetrovska Oblast, Zaporizka Oblast, Kirovohradska Oblast): New apartments 808; Existing apartments 22,627 - UA4 (Mykolaivska Oblast, Odeska Oblast, Khersonska Oblast): New apartments 698; Existing apartments 49,009 - UA5 (Vinnytska Oblast, Ternopilska Oblast, Khmelnytska Oblast): New apartments 540; Existing apartments 9,194 - UA6 (Kyivska Oblast, Cherkaska Oblast): New apartments 818; Existing apartments 14,515 - UA63 (M. Kyiv): New apartments 1,050; Existing apartments 48,595 - UA7 (Zakarpatska Oblast, Ivano-Frankivska Oblast, Lvivska Oblast, Chernivetska Oblast): New apartments 1,056; Existing apartments 12,822 - UA8 (Volynska Oblast, Zhytomyrska Oblast, Rivnenska Oblast): New apartments 547; Existing apartments 4,898 Aggregation and national index - Expenditure weights calculated as sum of expenditure (i.e., price in local currency) for calendar year 2023. - National HPI shows price change of about 15 percent over six quarters from 2023 Q1 to 2024 Q2. ### D. Dissemination — status and recommended next steps Findings and status - Several dissemination items remain contingent on the conclusion of the war. - Current dissemination items flagged as low priority (“L”) in the workplan until the conclusion of the war. Recommended actions (workplan) - L Review the revisions policy for the HPI after the conclusion of the war. — To be decided - L Update the structure of the published HPI data after the conclusion of the war. — To be decided - L Update the methods document after the conclusion of the war. — To be decided - L Publish the updated HPI on the authorities’ website after the conclusion of the war. — To be decided ### E. Capacity building and officials met Capacity building - Mission focused on increasing SSSU capacity: development of R scripts, practical training on hedonic methods, and provision of initial R scripts to staff. - Recommendation: SSSU staff should upskill on the R statistical package through identified supports (online or in-person training courses, internal IT support, or other modes). - SSSU staff ran the R scripts during the mission and were able to export clean data files to the network; staff ownership of scripts is crucial. Officials met - The mission worked closely with SSSU officials and regional offices; local staff at the JVI provided logistical support. - (List of specific officials met is documented in the source materials.) Table 1. Priority Recommendations (explicit) - 30 Sep, 2024 — Identify appropriate IT supports for staff to upskill in the R statistical package. — Responsible: SSSU - 30 Sep, 2024 — Locate the raw data files received from OLX for each quarter from 2023 Q1. — Responsible: SSSU - 30 Sep, 2024 — Investigate the unusual price changes for 2024 Q2 for new apartments. — Responsible: SSSU - 31 Dec, 2024 — Implement the new R scripts for the regular processing and cleaning of the OLX data. — Responsible: SSSU Appendix reference - Appendix contains box plot analysis of price of new properties over time highlighting the 2024 Q2 median increase in at least one NUTS region. *IMF Technical Assistance Report – Ukraine RPPI Mission, Section I. Detailed Technical Assessment and Recommendations* ### 25. Experimental quality-adjusted hedonic indices for the strata were compiled during the ### 25. Experimental quality-adjusted hedonic indices for the strata were compiled during the mission ### Experimental hedonic indices for existing apartments - RPPIs for existing apartments were calculated using the hedonic (time-dummy) method. - New apartments were excluded temporarily while the SSSU examined the unusual price movements for 2024 Q2. - A single index was calculated for each strata for existing apartments, with the same model applied for each. - Model specification: - Dependent variable: log of price (in local currency). - Independent variables: (i) floor area, (ii) number of rooms, and (iii) region. - Region variable was excluded from the Kyiv model due to lack of additional location information within Kyiv. - The authorities will require additional assistance to: - refine the hedonic models used for each stratum, and - implement the rolling-window approach. - Regional outcomes observed: - Regions in the west of the country are experiencing significant price increases. - Regions most impacted by the war in the east of the country have shown little growth or even falling prices. - Data exclusions: - Data exclude the territories which are temporarily occupied by the Russian Federation and part of territories where the military actions are/were conducted. - Figure referenced: FIGURE 1. Experimental HPIs for Existing Apartments by NUTS Region (2023Q1=100), 2023 Q1 to 2024 Q2. ### National HPI aggregation, weights, and outlier treatment - A compilation spreadsheet was designed in MS Excel to aggregate the price indices for the strata to a national HPI for Ukraine using a set of expenditure weights. - Weight calculation: - Weight information was calculated as the sum of expenditure (i.e. price in local currency) for the calendar year of 2023. - The same set of weights were applied to each of the six quarters from 2023 Q1 to 2024 Q2. - Outlier detection and treatment: - Outliers were removed at the level of the strata using Cook’s Distance. - The method was discussed in more detail in the previous TA report. - National HPI result: - The national HPI for Ukraine is showing price change of about 15 percent over the six quarters from 2023 Q1 to 2024 Q2. - Figure referenced: FIGURE 2. Experimental National HPI for Existing Apartments (2023Q1=100), 2023 Q1 to 2024 Q2. ### Recommended actions (methodology and compilation) - Implement the time-dummy hedonic (TDH) method for index calculation. - Once familiar with the TDH method, extend the method by using a rolling window. - Use Cook’s Distance for outlier detection. - Use the OLX data for the expenditure weights. - Practice using the compilation spreadsheet for aggregating the stratum-level price indices. - Become familiar with the requirements under relevant EU statistical legislation. ### Dissemination (procedural points and recommended actions) - Dissemination was not discussed in detail during this mission; recommendations from the previous TA report continue to apply. - Procedural requirement: - The SSSU are required to release methodological details of any new or updated publications as part of an annual plan of statistics. This annual plan is published in March every year. - The methods document will need to be released some time in advance of the publication of the data and should be included in planning for any update of the HPI for Ukraine. - Recommended actions: - Review the revisions policy for the HPI after the conclusion of the war. - Update the structure of the published HPI data after the conclusion of the war. - Update the methods document after the conclusion of the war. - Publish the updated HPI on the authorities’ website after the conclusion of the war. ### Officials met during the mission - Nataliia Lysenko — Head of Division for Housing and Construction Price Statistics, Department for Price Statistics, SSSU - Victoriia Isaiuk — Senior Specialist of the Division for Construction and Housing Price Statistics, Department for Price Statistics, SSSU - Nadiia Profatska — Deputy Director-Head of Division for Consumer Price Statistics, Department for Price Statistic, SSSU - Hanna Buhaichuk — Deputy Head of the Division for Producer of industrial products Price Statistics, Department for Price Statistics, SSSU ### Appendix - Appendix figure referenced: FIGURE 3. Box Plots by Quarter for Log of Price, New Properties in Region UA1, 2023Q1–2024Q2. *IMF | Technical Assistance Report – Ukraine RPPI Mission* --- _Source: https://www.imf.org/-/media/files/publications/tar/2024/english/tarea2024085-print-pdf.pdf_