## CABO VERDE Consumer Price Index — JULY 2023 (Section 1)

## Source details

**Canonical URL:** [CABO VERDE Consumer Price Index — JULY 2023 (Section 1)](https://www.imf.org/-/media/files/publications/cr/2023/english/1cpvea2023003.pdf)

## Other formats

- [Markdown version](/-/media/files/publications/cr/2023/english/1cpvea2023003.pdf.md)
- [Structured JSON version](/-/media/files/publications/cr/2023/english/1cpvea2023003.pdf.json)

---

### Mission outcome and immediate actions
- Mission dates: July 17–21, 2023.
- Primary objective: assist Instituto Nacional de Estatística Cabo Verde (INECV) with improving and updating the CPI.
- Key outcomes:
  - Review and assessment of CPI compilation methods.
  - Discussion of the January 2023 CPI calculation issue.
  - Development and agreement on a work plan to guide updating the CPI.
- Immediate data and timeline notes:
  - Current CPI weights are based on the 2015 HIES and are outdated.
  - A new HIES began in January 2023 and will continue through December 2023; an updated CPI is planned for introduction by January 2025.
  - Preliminary Q1 HIES expenditure data should be made available to CPI staff to begin item identification and COICOP 2018 classification, enabling early selection of varieties and collection of reference prices.
- Identified calculation issue:
  - INECV identified a coding error related to the process of updating the price reference period each December; INEPT corrected the issue. The calculation error did not affect the disseminated data for January 2023.

### Weights: current state and recommendations
- Current situation:
  - CPI weights are based on 2015 HIES; these data are now outdated and can introduce bias.
  - A new HIES (January 2023–December 2023) will be used to update CPI weights; updated CPI planned by January 2025.
- Recommendations:
  - Provide preliminary item-level HIES data to CPI staff to identify new items and classify them according to COICOP 2018.
  - Implement COICOP 2018 (INECV currently uses COICOP 1999); early access to item-level expenditure data is required because COICOP 2018 includes significant changes.
  - Build INECV capacity for updating CPI weights—historically INEPT processed and uploaded files; INECV requires TA to ensure sustainability and ownership.
  - Do not assign fixed expenditure weights at the airline-route variety level; airline routes should not be weighted as varieties because fixed weights prevent replacement of disappeared routes and inclusion of new routes.

### Item and variety samples; treatment of missing items
- Treatment of missing items:
  - When an item disappears, collectors must determine if missing permanently or temporarily.
  - Temporarily missing seasonal varieties should be imputed in each month the variety is missing (example: if mangoes are missing for eight months and available four months, collect four months of prices and impute for eight months).
- Thresholds and windows:
  - For non-seasonal varieties, INECV should define and adhere to a maximum number of months a variety can be temporarily missing. INECV currently has a three-month threshold that is not followed; mission recommends allowing items to be missing up to six months.
  - INECV uses seasonal windows for seasonal items; mission recommends eliminating seasonal windows and collecting prices as soon as items become available in sampled outlets.
- Sample maintenance:
  - The modified Young index (see Index Calculation) facilitates replacement of varieties, outlets, and items and should be used.

### Index calculation: preferred formula and implementation
- Recommended index family and rationale:
  - The 2020 CPI Manual gives preference to the Young index; in practice countries compile either a Lowe or a Young index.
  - The Young index is preferred because it accounts for substitution between the price and weight reference periods and avoids distortions arising from price-updating of expenditure values (Lowe index).
  - The modified Young (short-term) formula is recommended and reflects best practice: it chains together short-term price changes, is more flexible, and better facilitates replacement of varieties/outlets/items and treatment of missing prices and quality adjustment.
- Implementation notes:
  - CPI staff were not aware if expenditure values were price updated historically; going forward, the Young index should be used and the modified Young recommended for operational flexibility.

### Temporarily missing prices: policy change
- Current practice and issue:
  - INECV generally uses recommended imputation methods but also uses carry-forward (repeating the last available price).
  - The carry-forward method introduces a downward bias and should be discontinued except for controlled or regulated prices where government announces changes in advance.
- Recommendation:
  - Impute all temporarily missing prices, including seasonal prices. Only controlled or regulated prices may be carried forward under explicit government schedules.

### Price reference period: change in practice
- Current practice and issue:
  - The index calculation system updates the price reference period each December and links January rates of change onto the December index.
  - This annual update practice began with the last CPI update but is inappropriate for INECV because weights are updated infrequently (approximately every 5 years).
- Recommendation:
  - Discontinue annual updating of the price reference period. Update the price reference period only when weights are updated.
  - Implementing the Modified Young formula removes the need for frequent updates of the price reference period.

### Rent survey methodology and panel approach
- Current practice and issues:
  - INECV collects rent paid directly from a sample of rental units but loosely follows a quarterly panel approach that is not strictly followed.
  - Rent in Cabo Verde often remains the same for several years; new apartment construction requires timely sample incorporation.
- Recommended methodology:
  - Adopt a panel survey rotated annually. Divide the representative sample of rental units into panels (recommended: 6 panels).
  - Collect prices from one panel each month (example: panel 1 in January and July, panel 2 in February and August, etc.).
  - Rotate a defined percentage of the total sample annually (20 or 25 percent) to incorporate new rental units promptly.
- Calculation detail:
  - With a six-panel approach, comparisons produce six-month changes between the same panel observations. To convert a six-month change into a one-month change, calculate the sixth root of the change and multiply by the previous period index level.
  - INECV will confirm how the current system calculates the rent index when authorities visit INEPT in September 2023.

### Headline inflation: definition and recommended change
- Current practice:
  - INECV currently defines the rolling 12-month average as the headline or official index, effectively smoothing the data.
- Recommendation:
  - The headline index (official inflation rate) should not be adjusted or smoothed. INECV should define headline inflation as the annual (12-month) rate of change and adopt this as the official measure (target date: January 2024).

### Subsidized rent and inclusion in the index (Section 2 points)
- Current practice:
  - Currently, only market rents are collected; market rents represent most of the rent paid, but the government subsidizes rent for households that meet certain criteria.
- Recommendations:
  - Subsidized rent should be included in the rent survey.
  - Ideally, there should be weights for subsidized versus market rent, but these data are currently not available.
  - Alternative treatment suggested: prices collected would reflect the proportion of market versus subsidized rent (example: "if 85 percent of rent paid is market rent, then 85 percent of the prices collected would be market rent prices and 15 percent would be subsidized rent").
  - Implement the panel approach, with annual rotation, for the rent survey (more formalized than current practice).
  - Augment the sample of rental units – including both market and subsidized rent.
  - Review how rent is calculated by the index calculation system.

### Technical assistance needs and institutional context
- Capacity building:
  - TA is required to build INECV capacity for updating CPI weights and for implementing the modified Young index and COICOP 2018.
  - TA is recommended to support improvements in the rent index methodology and to guide the elimination of inappropriate practices (annual price reference updates, carry-forward imputation, seasonal windows).
- Operational steps to accelerate update:
  - Release preliminary item-level HIES data to CPI staff to identify new items and start price collection for reference prices prior to final weight update.
  - Confirm and correct calculation-system behaviors (example: rent index computation, price reference handling) with INEPT during planned visits.
- Institutional context:
  - INECV has relied heavily on support and assistance of INEPT; this support and assistance will end in 2024.
  - INECV requested a follow-up mission once final expenditure data are available to begin developing the weights.

### Explicit recommended actions (from the report)
- Lowest level of weight should be the item (elementary index level). Varieties should not have weights.
- Define the annual (12-month) change as the headline or official inflation rate for Cabo Verde.
- Return to using a fixed reference period and discontinue the practice of updating the price reference each December (annual chaining).
- Implement the Modified Young index calculation formula.
- Discontinue the practice of repeating the last collected price (carry-forward) for the treatment of seasonal and non-seasonal missing prices.
- Adjust the threshold for missing non-seasonal items.
- Seasonal products – discontinue the use of seasonal windows.
- Augment the sample of rental units – including both market and subsidized rent.
- Review how rent is calculated by the index calculation system.
- Implement the panel approach, with annual rotation, for the rent survey (more formalized than current practice).

### Key milestones and timeline (selected items)
- Variety sample updated and augmented: December 2023
- Using preliminary HIES data at item level, identify new items: December 2023
- Define headline inflation as the annual (12-month) rate of change: January 2024
- Begin collecting prices for new items identified: January 2024
- Begin classifying items into COICOP 2018: January 2024
- Review and adapt the index calculation system to reflect recommended index calculation methods: January 2024
- Finalize classification of items according to COICOP 2018: July 2024
- Preliminary national and regional weights and baskets available: September 2024
- Review and augment, as needed, the outlet sample: September 2024
- Finalize the variety sample: October 2024
- Finalize national and regional weights and baskets: November 2024
- Calculate reference prices (2024): December 2024
- Begin calculation of new index: January 2025
- Disseminate updated CPI Index: February 2025
- Update and disseminate detailed CPI compilation methods: February 2025
- Linking the old and the new CPI data series (2024=100): February 2025

*IMF Country Report No. 23/374 — Technical Assistance Report: Consumer Price Index, July 2023*

### Section 1

### CABO VERDE Consumer Price Index — JULY 2023 (Section 1)

### Summary of mission outcome and priority recommendations
- Mission dates: July 17–21, 2023.
- Primary objective: assist Instituto Nacional de Estatística Cabo Verde (INECV) with improving and updating the CPI.
- Key outcomes:
  - Review and assessment of CPI compilation methods.
  - Discussion of the January 2023 CPI calculation issue.
  - Development and agreement on a work plan to guide updating the CPI.
- Immediate data and timeline notes:
  - Current CPI weights are based on the 2015 HIES and are outdated.
  - A new HIES began in January 2023 and will continue through December 2023; an updated CPI is planned for introduction by January 2025.
  - Preliminary Q1 HIES expenditure data should be made available to CPI staff to begin item identification and COICOP 2018 classification, enabling early selection of varieties and collection of reference prices.
- Identified calculation issue:
  - INECV identified a coding error related to the process of updating the price reference period each December; INEPT corrected the issue. The calculation error did not affect the disseminated data for January 2023.
- Priority recommendations and target dates:
  - December 2023 — Review and confirm the existing variety sample. (Responsible: INECV)
  - December 2023 — Using preliminary HIES data at the item level, identify new items. (Responsible: INECV)
  - January 2024 — Define headline inflation as the annual (12-month) rate of change. (Responsible: INECV)

### Weights
- Current situation:
  - CPI weights are based on 2015 HIES; these data are now outdated and can introduce bias.
  - A new HIES (January 2023–December 2023) will be used to update CPI weights; updated CPI planned by January 2025.
- Recommendations:
  - Provide preliminary item-level HIES data to CPI staff to identify new items and classify them according to COICOP 2018.
  - Implement COICOP 2018 (INECV currently uses COICOP 1999); early access to item-level expenditure data is required because COICOP 2018 includes significant changes.
  - Build INECV capacity for updating CPI weights—historically INEPT processed and uploaded files; INECV requires TA to ensure sustainability and ownership.
  - Do not assign fixed expenditure weights at the airline-route variety level; airline routes should not be weighted as varieties because fixed weights prevent replacement of disappeared routes and inclusion of new routes.

### Item and variety samples
- Treatment of missing items:
  - When an item disappears, collectors must determine if missing permanently or temporarily.
  - Temporarily missing seasonal varieties should be imputed in each month the variety is missing (e.g., if mangoes are missing for eight months and available four months, collect four months of prices and impute for eight months).
- Thresholds and windows:
  - For non-seasonal varieties, INECV should define and adhere to a maximum number of months a variety can be temporarily missing. INECV currently has a three-month threshold that is not followed; mission recommends allowing items to be missing up to six months.
  - INECV uses seasonal windows for seasonal items; mission recommends eliminating seasonal windows and collecting prices as soon as items become available in sampled outlets.
- Sample maintenance:
  - The modified Young index (see Index Calculation) facilitates replacement of varieties, outlets, and items and should be used.

### Index calculation
- Recommended index family and rationale:
  - The 2020 CPI Manual gives preference to the Young index; in practice countries compile either a Lowe or a Young index.
  - The Young index is preferred because it accounts for substitution between the price and weight reference periods and avoids distortions arising from price-updating of expenditure values (Lowe index).
  - The modified Young (short-term) formula is recommended and reflects best practice: it chains together short-term price changes, is more flexible, and better facilitates replacement of varieties/outlets/items and treatment of missing prices and quality adjustment.
- Implementation notes:
  - CPI staff were not aware if expenditure values were price updated historically; going forward, the Young index should be used and the modified Young recommended for operational flexibility.

### Temporarily missing prices
- Current practice and issue:
  - INECV generally uses recommended imputation methods but also uses carry-forward (repeating the last available price).
  - The carry-forward method introduces a downward bias and should be discontinued except for controlled or regulated prices where government announces changes in advance.
- Recommendation:
  - Impute all temporarily missing prices, including seasonal prices. Only controlled or regulated prices may be carried forward under explicit government schedules.

### Price reference period
- Current practice and issue:
  - The index calculation system updates the price reference period each December and links January rates of change onto the December index.
  - This annual update practice began with the last CPI update but is inappropriate for INECV because weights are updated infrequently (approximately every 5 years).
- Recommendation:
  - Discontinue annual updating of the price reference period. Update the price reference period only when weights are updated.
  - Implementing the Modified Young formula removes the need for frequent updates of the price reference period.

### Rent survey
- Current practice and issues:
  - INECV collects rent paid directly from a sample of rental units but loosely follows a quarterly panel approach that is not strictly followed.
  - Rent in Cabo Verde often remains the same for several years; new apartment construction requires timely sample incorporation.
- Recommended methodology:
  - Adopt a panel survey rotated annually. Divide the representative sample of rental units into panels (recommended: 6 panels).
  - Collect prices from one panel each month (e.g., panel 1 in January and July, panel 2 in February and August, etc.).
  - Rotate a defined percentage of the total sample annually (20 or 25 percent) to incorporate new rental units promptly.
- Calculation detail:
  - With a six-panel approach, comparisons produce six-month changes between the same panel observations. To convert a six-month change into a one-month change, calculate the sixth root of the change and multiply by the previous period index level.
  - INECV will confirm how the current system calculates the rent index when authorities visit INEPT in September 2023.

### Headline inflation rate
- Current practice:
  - INECV currently defines the rolling 12-month average as the headline or official index, effectively smoothing the data.
- Recommendation:
  - The headline index (official inflation rate) should not be adjusted or smoothed. INECV should define headline inflation as the annual (12-month) rate of change and adopt this as the official measure (target date: January 2024).

### Technical assistance needs
- Capacity building:
  - TA is required to build INECV capacity for updating CPI weights and for implementing the modified Young index and COICOP 2018.
  - TA is recommended to support improvements in the rent index methodology and to guide the elimination of inappropriate practices (annual price reference updates, carry-forward imputation, seasonal windows).
- Operational steps to accelerate update:
  - Release preliminary item-level HIES data to CPI staff to identify new items and start price collection for reference prices prior to final weight update.
  - Confirm and correct calculation-system behaviors (e.g., rent index computation, price reference handling) with INEPT during planned visits.

*IMF Country Report No. 23/374 — Technical Assistance Report: Consumer Price Index, July 2023*

### Section 2

### 1cpvea2023003 - Section 2

### Rent survey, subsidized rent, and panel approach
- Currently, only market rents are collected; market rents represent most of the rent paid, but the government subsidizes rent for households that meet certain criteria.
- Subsidized rent should be included in the rent survey.
- Ideally, there should be weights for subsidized versus market rent, but these data are currently not available.
- Alternative treatment suggested: prices collected would reflect the proportion of market versus subsidized rent. Example given:
  - "if 85 percent of rent paid is market rent, then 85 percent of the prices collected would be market rent prices and 15 percent would be subsidized rent."
- Recommended implementation changes:
  - Implement the panel approach, with annual rotation, for the rent survey (more formalized than current practice).
  - Augment the sample of rental units – including both market and subsidized rent.
  - Review how rent is calculated by the index calculation system.

### Headline inflation measurement
- Current practice: headline (official) inflation rate for Cabo Verde is defined as the change in the 12-month rolling average index; each month a rolling average of the last 12 months is calculated and compared to the average across the same period in the previous year.
- Assessment:
  - The rolling average smooths the data.
  - "A headline index should not be smoothed or adjusted."
  - The 12-month rolling average index is considered an analytical index, not an official measure of inflation.
- Recommended change:
  - Define the headline inflation as the 12-month change (comparing the current month with the same month in the previous year).
  - "Monetary policy decisions, indexation, and the official measure of inflation should be based on the 12-month change."
  - "The practice of using the 12-month rolling average should be discontinued as soon as possible."

### Technical assistance needs and institutional context
- INECV has relied heavily on support and assistance of INEPT; this support and assistance will end in 2024.
- Senior management and staff noted the need to build capacity and assist with updating the CPI.
- INECV requested a follow-up mission once final expenditure data are available to begin developing the weights.

### Recommended actions (explicit list from the report)
- Lowest level of weight should be the item (elementary index level). Varieties should not have weights.
- Define the annual (12-month) change as the headline or official inflation rate for Cabo Verde.
- Return to using a fixed reference period and discontinue the practice of updating the price reference each December (annual chaining).
- Implement the Modified Young index calculation formula.
- Discontinue the practice of repeating the last collected price (carry-forward) for the treatment of seasonal and non-seasonal missing prices.
- Adjust the threshold for missing non-seasonal items.
- Seasonal products – discontinue the use of seasonal windows.
- Augment the sample of rental units – including both market and subsidized rent.
- Review how rent is calculated by the index calculation system.
- Implement the panel approach, with annual rotation, for the rent survey (more formalized than current practice).

### Key milestones and timeline from Annex II (selected items)
- Variety sample updated and augmented: December 2023
- Using preliminary HIES data at item level, identify new items: December 2023
- Define headline inflation as the annual (12-month) rate of change: January 2024
- Begin collecting prices for new items identified: January 2024
- Begin classifying items into COICOP 2018: January 2024
- Review and adapt the index calculation system to reflect recommended index calculation methods: January 2024
- Finalize classification of items according to COICOP 2018: July 2024
- Preliminary national and regional weights and baskets available: September 2024
- Review and augment, as needed, the outlet sample: September 2024
- Finalize the variety sample: October 2024
- Finalize national and regional weights and baskets: November 2024
- Calculate reference prices (2024): December 2024
- Begin calculation of new index: January 2025
- Disseminate updated CPI Index: February 2025
- Update and disseminate detailed CPI compilation methods: February 2025
- Linking the old and the new CPI data series (2024=100): February 2025

*Source: IMF | Technical Assistance Report – Cabo Verde, Consumer Price Index — 1cpvea2023003 - Section 2*

---


_Source: https://www.imf.org/-/media/files/publications/cr/2023/english/1cpvea2023003.pdf_
