## tarea2024050

## Source details

**Canonical URL:** [tarea2024050](https://www.imf.org/-/media/files/publications/tar/2024/english/tarea2024050.pdf)

## Other formats

- [Markdown version](/-/media/files/publications/tar/2024/english/tarea2024050.pdf.md)
- [Structured JSON version](/-/media/files/publications/tar/2024/english/tarea2024050.pdf.json)

---

### Section I. Introduction — Background, objectives, and key outcomes
- TA mission: August 29–September 8, 2023; response to request from the Central Statistics Office (CSO), Ministry of Finance, Dominica.
- Missions preceding: IMF Statistics Department mission in October 2022 and CARTAC mission in April 2023.
- Objectives:
  - Support improvements of the GDP series in line with the 2008 SNA standards, including re-referencing the volume estimates of GDP to 2018 prices (updated from the existing 2006 base year).
  - Identify data holdings to improve the quality of GDP estimates.
  - Develop a work program for improving economic statistics.
- Current compilation framework and staffing:
  - CSO compiles and releases current and constant 2006 price annual estimates of GDP by economic activities based on ISIC Revision 3.1 and broadly consistent with the 1993 SNA.
  - CSO has around 12 professional staff including three dedicated to the National Accounts (includes one new starter in November 2022).
  - Major data sources: Ministry of Agriculture, Ministry of Finance, financial statements of major companies, government agencies and regulatory authorities, ECCB data on insurance and finance, and the annual National Accounts Survey (NAS).
- Key data-quality issue:
  - NAS is a sample survey of around 500 businesses; recent response rates have been less than 10 percent.
  - Low NAS response required significant imputation, largely carrying forward previously reported values.
  - Mission replaced NAS reliance by incorporating alternative sources, notably:
    - Inland Revenue Department (IRD) data on turnover reported for the VAT system.
    - Department of Social Security (DSS) data on contributions to social security funds and numbers of employees.
  - Mission finalized systematic bases for processing and updating these datasets and for producing time series by economic activity.

### Section I. Introduction — Methodological and compilation improvements
- System and classification updates:
  - Developed a single ‘master’ workbook linked to a small number of ‘data’ workbooks; replaced numerous redundant workbooks and streamlined procedures.
  - Updated classification from ISIC Revision 3.1 to ISIC Revision 4.
  - Updated estimates for all years from 2010–2019; developed long-run time series from 2000–2019 in both current and constant prices.
- Specific estimation improvements:
  - Introduced a simple work in progress model for livestock production (cattle, sheep, goats, pigs) replacing previous stock-based method.
  - Used more detailed quantity data for fish caught to replace simple aggregation of tonnage across fish types.
  - Used VAT sales data to update output estimates for manufacturing, renting of machinery and equipment, and a number of business and personal services.
  - FISIM estimates aligned with 2008 SNA methodology, including a simple allocation of FISIM consumed as intermediate consumption of the business sector.
  - Volume of output for Private Households with Employed Persons updated using DSS employee-count data.
  - Method for estimating taxes on products and subsidies on products in constant prices updated to use the volume of total gross value added (GVA) for all economic activities as the indicator of change.
- Quality assurance and documentation:
  - Introduced basis for quality assuring estimates in current and constant prices based on contributions to the annual percentage change in total GDP from each economic activity.
  - Detailed business process documentation developed for the GDP system.

### Section I. Introduction — Revisions and selected numerical outcomes
- Current prices:
  - Level of GDP in current prices estimated as 18 percent higher than the previously published level; revision largely due to using VAT-based turnover data replacing NAS-based estimates.
- Constant prices — weight and share changes (2006 share → 2018 share):
  - Education services: 12 percent → 4 percent (mainly because of the closure of a major university at the end of 2017).
  - Business services: 1 percent → 11 percent (partly due to introduction of VAT turnover data replacing NAS).
  - Construction services: 4 percent → 11 percent (largely due to post-Hurricane Maria repair and construction).
  - Manufacturing: increased share due to improved VAT-based data sources.
- Selected revisions to annual percentage changes in GDP (constant prices) when moving from 2006 base to 2018 reference:
  - 2013: revised from -1.0 percent (2006 base) to 4.0 percent (2018 reference), a revision of +5.1 percentage points.
    - 1.8 percentage points from introducing VAT turnover for Business services.
    - 1.5 percentage points from changed methodology for estimating the volume of taxes on products.
  - 2018: revised from 3.5 percent (2006 base) to -3.7 percent (2018 reference), a revision of -7.2 percentage points.
    - Improved method for measuring the volume of taxes accounted for -6.6 percentage points of the total.
  - 2019: revised from 5.5 percent (2006 base) to 9.7 percent (2018 reference), a revision of +4.2 percentage points.
    - Improved method for measuring the volume of taxes accounted for 2.8 percentage points.
    - Revisions to manufacturing and business services (from VAT turnover) accounted collectively for 2.5 percentage points.
    - Agricultural output (livestock work in progress model) accounted for 1.3 percentage points.

### Section I. Introduction — Chainlinking and long-run series construction
- Long-run time series developed for 2000–2019 in both current and constant prices.
- Current price backcasting: GVA levels for each elementary activity backcast using previously published growth rates and aggregated.
- Constant price chainlinking: levels backcast based on previously published growth rates; chainlinking preserves previous base years before the latest ‘link’ year.
- Link year chosen: 2012.
  - For years 2000–2012, 2006 current price shares used as weights; for later years, 2018 shares used.
  - Chainlinking removes additivity for constant price GVA across activities in years before the link year (sum of constant price GVA by activity will not equal total constant price GDP for years before 2012).

### Section I. Introduction — Priority recommendations (summary with targets)
- Update the data in the GDP system to produce estimates for all years from 2010 to 2021, and publish the referenced series together with a note describing the revisions resulting from the rebasing.
  - Target Date: October 2023
  - Responsible Institution: CSO
- Acquire unit level annual data, including company names, from the Inland Revenue Department’s Corporate Income Tax, Personal Income Tax (for self-employed persons), and Value Added Tax.
  - Target Date: September 2023
  - Responsible Institution: CSO
- Code the IRD dataset to ISIC Revision 4 and link the records to the equivalent businesses in the DSS dataset to establish a business register.
  - Target Date: October 2023
  - Responsible Institution: CSO

*IMF | Technical Assistance Report – Dominica Improving Estimates of Gross Domestic Product | Section I. Introduction*

### Section III. Improving the Data Content of GDP and administrative data actions
- Administrative data sources identified as important: Customs (trade, cargo, shipping), ECCB (financial and insurance sector), DSS (employment and contributions), Ministry of Finance (government accounts), IRD (some company data), DDA (tourism), Fisheries Department (fishing).
- Current CSO use of administrative data:
  - CSO obtains aggregated VAT data from IRD and unit level employment information (excluding earnings) from the DSS.
  - Mission incorporated some VAT file information into estimates for manufacturing, business services, and telecommunications, but broader use limited by aggregated VAT categories and potential misclassification.
- Data that could considerably improve GDP quality if accessed at unit level:
  - Unit level VAT turnover information.
  - Unit level personal and corporate income tax schedules, including financial statements from IRD.
  - Insurable earnings from DSS (DSS can provide maximum insurable earnings but does not collect total compensation nor hours worked).
- Potential benefits of micro-level tax data:
  - Access to detailed micro level data, including company names, from CIT and PIT systems would enable significant increase in data content of GDP estimates.
  - Cross-referencing CIT/PIT with DSS employment and earnings can support a register for consistent industry classification and a frame for sample surveys.
- Action requested of MOF/IRD:
  - Facilitate CSO access to micro level data from the VAT system.

### Section III. Agriculture, DSS, Customs, Tourism, and other sectors — current issues and planned improvements
- Agriculture:
  - Current weakness: extrapolation of total supply by extrapolating historical benchmark for domestic consumption using the CPI for food and adding total agriculture exports.
  - Census of Agriculture expected to begin in the second quarter of 2024.
  - Plans to equip enumerators with computer assisted collection and auditing technology.
  - Chief Veterinarian provided livestock production cycle parameters and sample livestock production information for the PIM model.
- Dominica Social Security (DSS):
  - DSS data reasonably comprehensive and updated to ISIC Revision 4.
  - DSS regularly submits unit level employment and contributions data for over 4000 businesses.
  - DSS does not collect total salaries and wages nor number of hours worked; can provide maximum insurable earnings.
- Customs:
  - Import data collection very reliable due to tax resourcing.
  - Imports with value less than 150 EC enter duty free; customs officers do not classify these goods and they are not part of submission to CSO.
  - Mission requested an estimate of courier packages under 150 EC.
  - Export data challenges: non-standardized packaging for agriculture exports; insufficient officers to ensure accurate recording.
- Tourism:
  - DDA collects arrival information and traveler attributes from all ports of entry; limited data for tourists arriving by yacht.
  - Mission requested unit level information for each commercial accommodation type: maximum occupancy rate, length of stay, average price, and unit level data for commercial accommodation types.
  - Visitor Expenditure Survey (VES) was suspended during COVID and relaunched in May 2022; planned quarterly basis but interrupted by other priorities.
    - For 2023, VES will only be available for the 4th quarter.
- Other activities requiring improved sources/methods:
  - Land transportation: need for volume and price estimates for taxis, buses and trucks; administrative tax data plus sample information on distance travelled and fares needed.
  - Telecommunications: two main providers supply information but data inconsistent with VAT information; access to tax data would allow reconciliation.
  - Restaurants and meals outside the home: price GDP estimates currently extrapolated by bed nights and CPI for meals outside the home; administrative data on sales/costs of businesses needed.

### Section III. The way forward, benchmark data needs, and mission schedule
- CSO should continue accessing administrative data from government departments and public institutions; unit level data from CIT, PIT and VAT would greatly improve GDP estimates.
- IMF Real Sector goals:
  - Establish current benchmark estimates of economic activity through a supply and use framework.
  - Produce a comprehensive rebase GDP and improve price statistics (both consumer and producer price indexes).
- Data required for SUT benchmarks and rebasing CPI/GDP (listed):
  - Survey of Living Conditions data.
  - Data by SUT product group on trade and transport margins.
  - Data on construction industry costs and mark-ups by type of construction.
  - Data from Ministry of Agriculture on cultivated land and yields.
  - Data from DEXIA on agriculture export prices.
  - Detailed VAT Data.
  - Corporate and personal income tax data.
  - Comprehensive data for land transportation – taxis, buses and trucks.
  - Data for restaurants sales and intermediate consumption.
  - Data for livestock including lifecycle information.
  - Tourism expenditures.
- Mission schedule:
  - Workplan for a further 6-7 missions by the IMF Real Sector Division for the period up to December 2024 agreed with the CSO team; these missions aim to improve economic and price statistics and produce benchmark GDP statistics from a supply and use framework to rebase GDP for 2023. After these missions, the CARTAC real sector advisor will provide a peer review.

### Section III. Timeline items and key recommendations with targets
- Census of Agriculture expected to begin in the second quarter of 2024.
- VES relaunched in May 2022; for 2023 VES will only be available for the 4th quarter.
- Priority recommendations and target completion dates:
  - Update the data in the GDP system to produce estimates for all years from 2010 to 2021. — October 2023
  - Publish the referenced series together with a note describing the revisions resulting from rebasing to 2018 prices and the incorporation of more comprehensive data. — October 2023
  - Acquire unit level annual data, including company names, from the Inland Revenue Department’s Corporate Income Tax, Personal Income Tax (for self-employed persons), and Value Added Tax. — September 2023
  - Code the IRD dataset to ISIC Revision 4 and link the records to the equivalent businesses in the DSS dataset to establish a prototype business register. — October 2023
  - Where relevant, incorporate the CIT/PIT data from these sources into the annual estimation procedure, notably for manufacturing business services. — October 2023
  - Work with the MOA and obtain data on area under crop cultivation and yield rates to improve the output and IC estimates. — November 2023
  - Collaborate with the MOA to obtain key parameter information for the livestock model. — November 2023
  - Request an estimate of the courier packages under 150 EC that can be brought into Dominica free of custom duties. — November 2023
  - Obtain additional information on inbound tourism. — November 2023
  - Establish a regular quarterly visitor expenditure survey. — April 2024
  - Work with DEXIA to understand the agriculture export process. — December 2023

*Source: tarea2024050 - Section III. Improving the Data Content of GDP and a*

### Appendix II. GDP System Business Processing Documentation — system housekeeping and workflow
- System overview:
  - Set of integrated ‘industry’ workbooks and one ‘master’ workbook GDP_Master.xlsb.
  - Maintain linkages between individual workbooks and the master workbook; some industry workbooks also link to others.
  - Check links at the start and end of each GDP ‘round’ using Data-Edit Links -> Check Status in Excel; relink using ‘Change link’ if “Not found.”
  - Workbook naming is strict; folder containing workbooks should have fixed names (example: GDP_Live).
  - After a GDP ‘round’, copy entire folder and save as an archive version (example: GDP_2021).
  - CRITICAL ISSUE: DO NOT ADD ROWS OR COLUMNS IN ANY OF THE LINKED WORKBOOKS WITHOUT ALSO OPENING THE MASTER WORKBOOK FIRST, OTHERWISE THE LINKS WILL BE TO THE WTRONG ROWS/COLUMNS.
  - Workbooks are in .xlsb format.
- Purpose and structure of GDP_Master.xlsb:
  - Produces benchmark estimates of annual GDP in current and constant prices based on the selected base year (currently 2018).
  - Formulae in worksheets are extended to 2026; master workbook itself does not need updating until 2027, though linked workbooks do.
- GROUP 1: DATA — key worksheets and links:
  - Worksheets: ISIC, CPI, SS_Emp, SS_Conp, VAT_4dig, VAT_2Dig.
  - SS_Comp and SS_Emp link to Soc_Sec_Master.xlsb.
  - VAT_4dig and VAT_2Dig link to VAT_master.xlsb.
  - CPI links to CPI_A.xlsb.
  - Column A contains SUT_Activity codes; ISIC is ISIC Rev 4.
- GROUP 2: ECONOMIC ACTIVITIES — organization:
  - 18 worksheets, one for each economic activity; many contain yellow cells labeled "Enter data".
  - Activities listed verbatim include: Tax_Sub Manu Sea Real_Est; Bananas Elec Air Rent; O_Crops Water Aux_Trans Comp; Livestock Const FISIM Bus_Serv; Fishing Trade Banks_Data Pub_Sect; Forestry Accom_Data Banks Edu_Priv; Quarry Accom Aux_Fin Health_Priv; Manu_ICO Rest Ins Pers; Manu_Data Trans_Data Comms Priv_HH; Manu_Q Road Real_Est_Data.
  - Each dark-blue worksheet contains links to indicator data and to the Base worksheet containing benchmark levels of Output, IC and GVA for sub-activities.
  - Quarterly series from industry workbooks linked and used as indexes to extrapolate base year levels.
- GROUP 3: SUMMARY — contents:
  - Six worksheets: Summary, KP_GVA, KP_IC, KP_O, CP_GVA, CP_IC; summarized to published level in Group 7.
- GROUP 4: BASE:
  - Single worksheet Base contains benchmark levels of Output and IC or GVA for each economic activity.
- GROUP 5: PUBL — final publication tables and contributions:
  - Worksheets include: Tables, PUBL_KP, PUBL_KP_%_Ch, Cont_KP, PUBL_CP, Cont_CP, PUBL_CP_Share, PUBL_CP_O, PUBL_CP_IC.
  - PUBL_CP and PUBL_KP contain final tables in CP and KP (based on 2018 prices).
  - Cont_KP and Cont_CP contain contribution analysis for published series (contributions to annual percentage change in total GDP from each activity).
- Updating the Annual GDP Series — step-by-step:
  1. Update the ‘industry’ workbooks (see Annex 2).
  2. Update the CPI.xlsb workbook.
  3. Update all yellow cells indicated in Group 2 of the GDP_Master workbook.
  4. Update the Social Security and VAT databases: VAT_Master.xlsb and Soc_Sec_Master.xlsb.
  5. Review the KP_Contrib and CP_Contrib worksheets to establish reasons for observed % changes in total GDP in constant and current prices; investigate issues in individual industry workbooks if identified.
  - Calculations are automatic and workbook formulas extend to 2026.
- Annex 1: Worksheets in the master workbook (listed verbatim) and Annex 2: Workbooks linked to the Master workbook (listed verbatim) specify required files and workbook names (including A012_Other_Crops.xlsb, A014_Livestock_Poultry.xlsb, A014_Livestock_WiP_Model.xlsb, A03_Fishing.xlsb, C_Manufacturing_IC_O_Historic.xlsb, _VAT_Master.xlsb, _Soc_Sec_Master.xlsb, etc.).

*IMF | Technical Assistance Report – Dominica Improving Estimates of Gross Domestic Product | Appendix II. GDP System Business Processing Documentation, Author: Robin Youll, Last updated 9 September 2023*

### Section I. Introduction ................................................................................................

### Section I. Introduction

### Background and objectives
- A technical assistance (TA) mission took place between August 29 and September 8, 2023, responding to a request from the Central Statistics Office (CSO) in the Ministry of Finance in Dominica.
- The mission followed an IMF Statistics Department mission in October 2022 and a CARTAC mission in April 2023.
- Objectives of the mission:
  - supporting improvements of the GDP series in line with the 2008 SNA standards, including re-referencing the volume estimates of GDP to 2018 prices (updated from the existing 2006 base year);
  - identifying data holdings which can be used to improve the quality of the estimates of GDP;
  - developing of a work program for improving economic statistics.

### Current compilation framework and staffing
- The CSO compiles and releases current and constant 2006 price annual estimates of GDP by economic activities based on ISIC Revision 3.1 and broadly consistent with the 1993 SNA.
- The CSO has around 12 professional staff including three dedicated to the National Accounts (includes one new starter in November 2022).
- Major data sources used in compilation include: Ministry of Agriculture, Ministry of Finance, financial statements of major companies, government agencies and regulatory authorities, ECCB data on insurance and finance, and the annual National Accounts Survey (NAS).

### Key data-quality issue identified
- The NAS is a sample survey of around 500 businesses; recent response rates have been less than 10 percent.
- Because of the low NAS response rate, estimates derived from NAS required significant imputation, largely carrying forward previously reported values.
- Consistent with the prior CARTAC recommendation, the mission replaced NAS reliance by incorporating alternative sources, notably:
  - Inland Revenue Department (IRD) data on turnover reported for the VAT system;
  - Department of Social Security (DSS) data on contributions to social security funds and numbers of employees.
- The mission finalized systematic bases for processing and updating these datasets and for producing time series by economic activity (Appendix I provides details).

### Methodological and compilation improvements introduced
- Developed a single ‘master’ workbook linked to a small number of ‘data’ workbooks; replaced numerous redundant workbooks and streamlined procedures.
- Updated classification from ISIC Revision 3.1 to ISIC Revision 4.
- Updated estimates for all years from 2010–2019; developed long-run time series from 2000–2019 in both current and constant prices.
- Specific estimation improvements introduced:
  - A simple work in progress model for livestock production (cattle, sheep, goats, pigs) to replace previous stock-based method that implicitly treated all animals as slaughtered each year.
  - More detailed quantity data for fish caught to replace simple aggregation of tonnage across fish types.
  - Use of VAT sales data to update output estimates for manufacturing, renting of machinery and equipment, and a number of business and personal services.
  - FISIM estimates aligned with 2008 SNA methodology, including a simple allocation of FISIM consumed as intermediate consumption of the business sector.
  - Volume of output for Private Households with Employed Persons updated using DSS employee-count data.
  - Method for estimating taxes on products and subsidies on products in constant prices updated to use the volume of total gross value added (GVA) for all economic activities as the indicator of change (previous method deflated current price values using an aggregate deflator for GVA).
- Introduced a basis for quality assuring estimates in current and constant prices based on contributions to the annual percentage change in total GDP from each economic activity.
- Detailed business process documentation developed for the GDP system (Appendix II).

### Revisions and their sources (high-level outcomes)
- In current prices:
  - The level of GDP in current prices was estimated as 18 percent higher than the previously published level; revision largely due to using VAT-based turnover data replacing NAS-based estimates.
- In constant prices:
  - Weights used to aggregate volume series updated from published current price 2006 levels of value added to updated estimates for 2018.
  - Notable changes in shares (2006 share → 2018 share):
    - Education services: 12 percent → 4 percent (mainly because of the closure of a major university at the end of 2017).
    - Business services: 1 percent → 11 percent (partly due to introduction of VAT turnover data replacing NAS).
    - Construction services: 4 percent → 11 percent (largely due to post-Hurricane Maria repair and construction).
    - Manufacturing: increased share due to improved VAT-based data sources.
- Selected revisions to annual percentage changes in GDP (constant prices) when moving from 2006 base to 2018 reference:
  - 2013: revised from -1.0 percent (2006 base) to 4.0 percent (2018 reference), a revision of +5.1 percentage points.
    - Of this, 1.8 percentage points resulted from introducing VAT turnover for Business services; 1.5 percentage points from the changed methodology for estimating the volume of taxes on products.
  - 2018: revised from 3.5 percent (2006 base) to -3.7 percent (2018 reference), a revision of -7.2 percentage points.
    - The improved method for measuring the volume of taxes accounted for -6.6 percentage points of the total.
  - 2019: revised from 5.5 percent (2006 base) to 9.7 percent (2018 reference), a revision of +4.2 percentage points.
    - The improved method for measuring the volume of taxes accounted for 2.8 percentage points, revisions to manufacturing and business services (from VAT turnover) accounted collectively for 2.5 percentage points, and agricultural output (livestock work in progress model) accounted for 1.3 percentage points.

### Chainlinking and long-run series construction
- Long-run time series (2000–2019) were developed in both current and constant prices.
- Current price backcasting: levels of GVA for each elementary activity backcast using previously published growth rates and aggregated.
- Constant price chainlinking: levels (including aggregate GDP) backcast based on previously published growth rates; chainlinking used to preserve previous base years before the latest ‘link’ year.
- Link year chosen: 2012 (approximately midway between 2006 and 2018).
  - For years 2000–2012, 2006 current price shares used as weights; for later years, 2018 shares used.
- Note: chainlinking removes additivity for constant price GVA across activities in years before the link year (sum of constant price GVA by activity will not equal total constant price GDP for years before 2012).

### Priority recommendations (summary)
- Update the data in the GDP system to produce estimates for all years from 2010 to 2021, and publish the referenced series together with a note describing the revisions resulting from the rebasing.
  - Target Date: October 2023
  - Responsible Institution: CSO
- Acquire unit level annual data, including company names, from the Inland Revenue Department’s Corporate Income Tax, Personal Income Tax (for self-employed persons), and Value Added Tax.
  - Target Date: September 2023
  - Responsible Institution: CSO
- Code the IRD dataset to ISIC Revision 4 and link the records to the equivalent businesses in the DSS dataset to establish a business register.
  - Target Date: October 2023
  - Responsible Institution: CSO

*IMF | Technical Assistance Report – Dominica Improving Estimates of Gross Domestic Product | Section I. Introduction*

### Section III. Improving the Data Content of GDP and a

### Section III. Improving the Data Content of GDP and a

### Data sources and administrative data
- Administrative tax and regulatory data identified as important sources for national accounts compilation: trade, cargo, and shipping data from the Customs Department; financial and insurance sector data from the ECCB; employment and contributions data from the DSS; Government accounts data from the Ministry of Finance; some data from the Inland Revenue Department (for some individual companies); tourism data from DDA; fishing data from the Fisheries Department.
- Current CSO use of administrative data:
  - CSO currently obtains aggregated VAT data from IRD and unit level employment information (excluding earnings) from the DSS.
  - The mission incorporated some VAT file information into estimates of value added for manufacturing, business services, and telecommunications, but broader use limited by aggregated VAT categories and potential misclassification.
- Data that could considerably improve GDP quality if accessed at unit level:
  - Unit level VAT turnover information.
  - Unit level personal and corporate income tax schedules, including financial statements from IRD.
  - Insurable earnings from DSS (DSS can provide maximum insurable earnings but does not collect total compensation nor hours worked).
- Potential benefits of micro-level tax data:
  - Access to detailed micro level data, including company names, from CIT and PIT systems would enable significant increase in data content of GDP estimates.
  - Cross-referencing CIT/PIT with DSS employment and earnings can support a register for consistent industry classification and a frame for sample surveys.
- Action requested of MOF/IRD:
  - Facilitate CSO access to micro level data from the VAT system.

### Agriculture
- Current methodology weaknesses:
  - In absence of current information on cultivated land, crop yields and market prices for other agriculture crops, CSO extrapolates total supply of agriculture by extrapolating a historical benchmark for domestic consumption using the CPI for food and adds total agriculture exports.
  - Custom records may not provide accurate weights for certain export products; current methodology may underestimate agriculture importance.
- Planned improvements and activities:
  - Census of Agriculture expected to begin in the second quarter of 2024.
  - Plans to equip agricultural enumerators with computer assisted collection and auditing technology to enhance collection, editing, imputation and accessibility of agriculture data.
  - PS providing incentives to agricultural wholesalers, exporters and other traders to improve response rates.
  - Chief Veterinarian provided livestock production cycle parameters and sample livestock production information for the PIM model; current livestock data collection not consistently and regularly provided to MOA.

### Dominica Social Security (DSS)
- DSS data characteristics:
  - DSS data are reasonably comprehensive and have been updated to ISIC Revision 4.
  - DSS regularly submits to the CSO unit level employment and contributions data for over 4000 businesses.
  - DSS does not collect total salaries and wages nor number of hours worked; can provide maximum insurable earnings.

### Customs
- Import data:
  - Collection of import data is very reliable due to tax collection resourcing.
  - Imports with a value of less than 150 EC enter duty free; customs officers do not classify these goods and they are not part of submission to CSO.
  - Mission requested an estimate of the courier packages (under 150 EC).
- Export data challenges (particularly agriculture):
  - Exports of agriculture products lack standardized packaging, so weights may be under- or over-estimated.
  - Insufficient officers to ensure accurate recording of exports.

### Tourism
- Discovery Dominica Authority (DDA) responsibilities:
  - Collects arrival information and traveler attributes from all ports of entry, including place of stay and type of accommodation (limited data for tourists arriving by yacht).
  - Mission requested unit level information for each commercial accommodation type: maximum occupancy rate, length of stay, average price, and unit level data for commercial accommodation types.
- Visitor Expenditure Survey (VES):
  - VES was suspended during COVID and relaunched in May 2022; planned quarterly basis but interrupted by other priorities (CENSUS).
  - For 2023, VES will only be available for the 4th quarter.
- Template for data on stay-over visitors by place of stay provided (ACCOMMODATION TYPE; Year 2022 2021 2020; Maximum Occupancy Rate; Length of Stay; Average Price).

### Other activities requiring improved sources/methods
- Land transportation:
  - Volume and price estimates for taxis, buses and trucks need improvement.
  - Administrative tax data plus sample information on distance travelled and fares would strengthen price estimates.
  - Model could be improved by collecting information for licensed commercial trucks (CSO has total stock of trucks for personal and commercial use).
- Telecommunications:
  - Two main providers supply information to CSO, but data are inconsistent with VAT information.
  - Access to administrative tax data and unit level VAT or CIT would allow better reconciliation and adjustments.
- Restaurants and meals outside the home:
  - Current price GDP estimates derived by extrapolating bed nights and CPI for meals outside the home; significant household consumption not captured by current methodology.
  - Administrative data on sales and costs of businesses in this sector could provide essential information for the formal sector.

### The way forward and benchmark data needs
- CSO should continue accessing administrative data from government departments and public institutions; unit level data from CIT, PIT and VAT would greatly improve GDP estimates.
- IMF Real Sector goals:
  - Establish current benchmark estimates of economic activity through a supply and use framework.
  - Produce a comprehensive rebase GDP and improve price statistics (both consumer and producer price indexes).
- Data required for SUT benchmarks and rebasing CPI/GDP:
  - Survey of Living Conditions data.
  - Data by SUT product group on trade and transport margins.
  - Data on construction industry costs and mark-ups by type of construction.
  - Data from Ministry of Agriculture on cultivated land and yields.
  - Data from DEXIA on agriculture export prices.
  - Detailed VAT Data.
  - Corporate and personal income tax data.
  - Comprehensive data for land transportation – taxis, buses and trucks.
  - Data for restaurants sales and intermediate consumption.
  - Data for livestock including lifecycle information.
  - Tourism expenditures.
- Mission schedule:
  - A workplan for a further 6-7 missions by the IMF Real Sector Division for the period up to December 2024 was agreed with the CSO team; these missions aim to improve economic and price statistics and produce benchmark GDP statistics from a supply and use framework to rebase GDP for 2023. After these missions, the CARTAC real sector advisor will provide a peer review.

### Timeline items and specific targets (as presented)
- Census of Agriculture expected to begin in the second quarter of 2024.
- VES relaunched in May 2022; for 2023 VES will only be available for the 4th quarter.
- Workplan: further 6-7 IMF Real Sector missions up to December 2024.

### Key recommendations (as stated)
- Acquire unit level annual data, including company names, from the Inland Revenue Department’s Corporate Income Tax, Personal Income Tax (for self-employed persons), and Value Added Tax.
- Code the IRD dataset to ISIC Revision 4 and link the records to the equivalent businesses in the DSS dataset to establish a prototype business register.
- Where relevant, incorporate the CIT/PIT data from these sources into the annual estimation procedure, notably for manufacturing business services.
- Work with the Ministry of Agriculture (MOA) and obtain data on area under crop cultivation and yield rates to improve the output and intermediate consumption estimates.
- Collaborate with the MOA to obtain key parameter information for the livestock model.
- Request an estimate of the courier packages under 150 EC that can be brought into Dominica free of custom duties.
- Obtain additional information on inbound tourism.
- Establish a regular quarterly visitor expenditure survey.
- Work with DEXIA to understand the agriculture export process.
- Improve the sources and method for land transportation.

### Priority recommendations and target completion dates (as listed)
- Update the data in the GDP system to produce estimates for all years from 2010 to 2021. — October 2023
- Publish the referenced series together with a note describing the revisions resulting from rebasing to 2018 prices and the incorporation of more comprehensive data. — October 2023
- Acquire unit level annual data, including company names, from the Inland Revenue Department’s Corporate Income Tax, Personal Income Tax (for self-employed persons), and Value Added Tax. — September 2023
- Code the IRD dataset to ISIC Revision 4 and link the records to the equivalent businesses in the DSS dataset to establish a prototype business register. — October 2023
- Where relevant, incorporate the CIT/PIT data from these sources into the annual estimation procedure, notably for manufacturing business services. — October 2023
- Work with the MOA and obtain data on area under crop cultivation and yield rates to improve the output and IC estimates. — November 2023
- Collaborate with the MOA to obtain key parameter information for the livestock model. — November 2023
- Request an estimate of the courier packages under 150 EC that can be brought into Dominica free of custom duties. — November 2023
- Obtain additional information on inbound tourism. — November 2023
- Establish a regular quarterly visitor expenditure survey. — April 2024
- Work with DEXIA to understand the agriculture export process. — December 2023

*Source: tarea2024050 - Section III. Improving the Data Content of GDP and a*

### 5. Go the worksheet 'New' and see which companies are new (in the sense that they are not

### tarea2024050 - 5. Go the worksheet 'New' and see which companies are new (in the sense that they are not 

### System housekeeping
- The system consists of a set of integrated ‘industry’ workbooks and one ‘master’ workbook GDP_Master.xlsb.
- Maintain linkages between individual workbooks and the master workbook; some industry workbooks also link to others.
- Check links at the start and end of each GDP ‘round’ using Data-Edit Links -> Check Status in Excel.
- If links are “Not found,” relink using the ‘Change link’ procedure in the ‘Data->Edit Links’ menu.
- Workbook naming is strict: names (as listed in Annex 2) must not be changed (do not add initials or a date).
- The folder containing the workbooks should have fixed names, e.g. GDP_Live.
- After a GDP ‘round’, copy entire folder and save as an archive version (example: GDP_2021).
- CRITICAL ISSUE: DO NOT ADD ROWS OR COLUMNS IN ANY OF THE LINKED WORKBOOKS WITHOUT ALSO OPENING THE MASTER WORKBOOK FIRST, OTHERWISE THE LINKS WILL BE TO THE WTRONG ROWS/COLUMNS.
- Workbooks are in .xlsb format (binary) for space savings and faster calculation.

### Purpose and structure of GDP_Master.xlsb
- Produces benchmark estimates of annual GDP in current and constant prices based on the selected base year (currently 2018).
- Comprises broad ‘groups’ of worksheets (see Groups below).
- Formulae in worksheets are extended to 2026; the master workbook itself does not need updating until 2027, though linked workbooks do.

### GROUP 1: DATA — key worksheets and links
- Worksheets: ISIC, CPI, SS_Emp, SS_Conp, VAT_4dig, VAT_2Dig.
- SS_Comp and SS_Emp link to Soc_Sec_Master.xlsb (Social Security data in Earnings/employees).
- VAT_4dig and VAT_2Dig link to VAT_master.xlsb (VAT data as matched-pairs indexes).
- CPI links to CPI_A.xlsb (complete list of CPIs needed).
- Column A contains SUT_Activity codes used in the master workbook (based on the activity classification used in the GDP compilation system).
- ISIC is ISIC Rev 4.

### GROUP 2: ECONOMIC ACTIVITIES — organization and method
- There are 18 worksheets, one for each economic activity, including some light-blue ‘Data’ worksheets for preliminary calculations.
- Activities listed include (verbatim): Tax_Sub Manu Sea Real_Est; Bananas Elec Air Rent; O_Crops Water Aux_Trans Comp; Livestock Const FISIM Bus_Serv; Fishing Trade Banks_Data Pub_Sect; Forestry Accom_Data Banks Edu_Priv; Quarry Accom Aux_Fin Health_Priv; Manu_ICO Rest Ins Pers; Manu_Data Trans_Data Comms Priv_HH; Manu_Q Road Real_Est_Data.
- Some activities are linked to their own ‘industry’ workbook (see Annex 2).
- Each dark-blue worksheet broadly contains:
  - Links (light blue) to ‘indicator’ data from the GVA system.
  - Links to the Base worksheet (brown/yellow) containing benchmark levels of Output, IC and GVA for sub-activities.
- Quarterly series from each industry workbook are linked and used as indexes to extrapolate base year levels from the GDP published series.
- Many worksheets contain yellow cells labeled "Enter data" requiring direct data entry (examples: Real_Est, Aux_Trans).

### GROUP 3: SUMMARY — purpose and contents
- Comprises six worksheets: Summary, KP_GVA, KP_IC, KP_O, CP_GVA, CP_IC.
- Provide a summary of all worksheets in Group 1 and act as a basis for pulling together Group 1 data.
- Useful for more detailed investigation of specific industries; summarized to the published level in Group 7.

### GROUP 4: BASE
- Single worksheet: Base.
- Contains benchmark levels of Output and IC or GVA for each economic activity.

### GROUP 5: PUBL — final publication tables and contributions
- Consists of 8 worksheets: Tables, PUBL_KP, PUBL_KP_%_Ch, Cont_KP, PUBL_CP, Cont_CP, PUBL_CP_Share, PUBL_CP_O, PUBL_CP_IC.
- PUBL_CP and PUBL_KP contain final tables in CP and KP (based on 2018 prices).
- Cont_KP and Cont_CP contain contribution analysis for published series (show contributions to annual percentage change in total GDP from each activity). Example phrasing provided: ‘GDP increased by 5.2% in 2022 compared to 2021, with manufacturing accounting for 2.1 percentage points of the increase.’
- PUBL_CP_O and PUBL_CP_IC are the current price output and intermediate consumption tables.

### Updating the Annual GDP Series — step-by-step
- The new annual GDP-P system is built on-top of the existing GVA system to allow regular data capture to continue.
- Two other linked workbooks to update: VAT_Master.xlsb and Soc_Sec_Master.xlsb (see Instructions worksheets in those workbooks).
- Steps to update:
  1. Update the ‘industry’ workbooks (see Annex 2) as usual.
  2. Update the CPI.xlsb workbook.
  3. Update all yellow cells indicated in Group 2 of the GDP_Master workbook in this colour.
  4. Update the Social Security and VAT databases: VAT_Master.xlsb and Soc_Sec_Master.xlsb.
  5. Review the KP_Contrib and CP_Contrib worksheets to establish reasons for observed % changes in total GDP in constant and current prices; investigate issues in individual industry workbooks if identified.
- Calculations are automatic and workbook formulas extend to 2026.

### Annex 1: Worksheets in the master workbook (as listed)
- Tables: PUBL_KP_GDP, PUBL_KP_%_Ch, Cont_KP_GDP, PUBL_CP_GDP, Cont_CP_GDP, PUBL_CP_Share, PUBL_CP_O.
- Summaries: KP_GVA, KP_O, KP_IC, CP_GVA, CP_O, CP_IC.
- Base: Base.
- Activities/Data: Tax_Sub, Manu, Sea, Real_Est, ISIC, CPI, O_Crops, Water, Aux_Trans, Comp, SS_Emp, Livestock, Const, FISIM, Bus_Serv, SS_Conp, Fishing, Trade, Banks_Data, Pub_Sect, VAT_4dig, Forestry, Accom_Data, Banks, Edu_Priv, VAT_2Dig, Quarry, Accom, Aux_Fin, Health_Priv, Manu_ICO, Rest, Ins, Pers, Manu_Data, Trans_Data, Comms, Priv_HH, Manu_Q, Road, Real_Est_Data.

### Annex 2: Workbooks linked to the Master workbook (listed verbatim)
- ‘ISIC’ Workbooks requiring updating:
  - A012_Other_Crops.xlsb
  - A014_Livestock_Poultry.xlsb
  - A014_Livestock_WiP_Model.xlsb
  - A03_Fishing.xlsb
  - A03_Fishing_Data.xlsb
  - C_Manufacturing_IC_O_Historic.xlsb
  - D_Electricity.xlsb
  - E_Water.xlsb
  - G_Trade.xlsb
  - H49_Road_Trans.xlsb
  - H50_Sea_Transport.xlsb
  - H51_Air_Liat.xlsb
  - J_Communications.xlsb
  - K_FISIM.xlsb
  - K65_Insurance.xlsb
  - K66_Aux_Fin.xlsb
  - L_Real Estate.xlsb
  - OPQ_Public_Sector.xlsb
  - P_Education_Private.xlsb
  - Q_Health_and_Social_Work.xlsb
  - R_S_Personal_Services.xlsb
  - Taxes on production.xlsb
- Other workbooks:
  - _CPI.xlsb
  - _Population.xlsb
  - _VAT_Master.xlsb
  - _Soc_Sec_Master.xlsb

*IMF | Technical Assistance Report – Dominica Improving Estimates of Gross Domestic Product | Appendix II. GDP System Business Processing Documentation, Author: Robin Youll, Last updated 9 September 2023*

---


_Source: https://www.imf.org/-/media/files/publications/tar/2024/english/tarea2024050.pdf_
