## _cr13359

## Source details

**Canonical URL:** [_cr13359](https://www.imf.org/-/media/websites/imf/imported-full-text-pdf/external/pubs/ft/scr/2013/_cr13359.pdf)

## Other formats

- [Markdown version](/-/media/websites/imf/imported-full-text-pdf/external/pubs/ft/scr/2013/_cr13359.pdf.md)
- [Structured JSON version](/-/media/websites/imf/imported-full-text-pdf/external/pubs/ft/scr/2013/_cr13359.pdf.json)

---

### Executive summary and mission objectives
- Mission undertaken in response to a request from the Malawi authorities, in consultation with the African Department (AFR), during September 16–27, 2013; activity 14STJ6300 funded by East AFRITAC and part of project Price Statistics STA_MWI_2008_05.
- Main objectives:
  - Assist with improving prices statistics, including full conversion to direct pricing for import and export price indexes (MXPI).
  - Review the new producer price index (PPI).
  - Prepare detailed plans for quarterly import and export price surveys and the replacement of unit values with survey-based prices.
  - Provide training on data collection and index compilation.
  - Document methodology and computational procedures developed.
- Focus placed on the development of an import price index (MPI); documentation prepared is for MPI but applicable to XPI.

### Key findings on source data and feasibility
- A detailed analysis of 2010 and 2011 quarterly, country-specific import unit values selected for the experimental MPI found that:
  - Over 80 percent of the country unit values did not provide credible indicators of pure price change over time.
  - Less than 20 percent of the series provided credible indicators of pure price change over time.
- Result: compilation of the MPI could not proceed using Customs-based unit values; it is not feasible to proceed until quarterly transaction prices are available from import and export price surveys.
- Number of 8-digit HS items selected for pricing in the MPI: 411 out of a potential 6,000 plus HS items.
- Country-mix elimination improves unit-value data quality, but the existing unit values remain insufficiently credible for index compilation.
- Identified primary causes of non-credible unit values (even after eliminating country mix):
  - compositional mix of products; and
  - incorrect quantity data (misreporting and Customs not editing closely, switching of units, etc.).
- Conclusion: Serious problems with Customs quantity data; ongoing liaison between the MNSO and Customs will be required to improve the data.

### Recommended statistical prerequisites and immediate actions
- Initiate price surveys before any index compilation can be undertaken. Resources must be provided to fund:
  - A program of import and export price survey initialization interviews.
  - Subsequent ongoing quarterly price collections.
- An intensive program of field interviews is required to obtain necessary prices for all selected 8-digit HS items before index compilation can start.
- Initialization interviews should be undertaken within a relatively short time frame so all price series can commence in the same quarter and support a common index time reference period; retrospective quarterly prices are generally not feasible too long after the event.
- A funding proposal has been prepared (Appendix II).

### Methodology, concepts, and compilation decisions
- Concepts and definitions consistent with System of National Accounts 2008 (2008 SNA) and the Export and Import Price Index Manual were confirmed.
- Aggregation and formulae:
  - Elementary aggregate (EA) indexes at the 8-digit level of the HS.
  - Unweighted price relatives calculated using the Jevons geometric mean aggregation formula at EA level.
  - EA indexes aggregated to higher HS levels (4-digit, 2-digit, total) with fixed weights using the Laspeyres formula.
- Weighting:
  - Index weights will be smoothed by basing them on the average value of imports over the period 2007–2010; 2011 was not considered a “normal” year.
  - In the longer term, MXPI should be re-referenced to the same base year as the CPI and PPI.
- Reference period:
  - Initial plan was 2010 as time reference period, but because survey prices must be collected before compilation, a more recent period is required.
  - If initialization interviews can be completed before the end of 2013, request prices for March, June, September, and December quarters of 2013 and make 2013 the time reference period.
  - If interviews are delayed, a later time reference period will be selected and index compilation deferred.
- Processing systems:
  - The PPI processing system contains required functionality for MXPI; work identified to tailor the PPI system to MXPI requirements.

### Sampling and survey design
- A sample of importing establishments was selected for the import price survey using cut-off sampling techniques.
- MNSO to proceed to select a sample of exporters for the export price survey.
- Generic questionnaires were designed for a quarterly Survey of Import Prices and a Survey of Export Prices.
- Questionnaires designed for customization during initialization interviews to capture specific product specifications, units, country of origin/destination, and currency of transaction.

### Training, documentation, and capacity building
- Training provided on index concepts, compilation methodology, and price survey design and implementation.
- Emphasis on training for price survey initialization interviews and subsequent quarterly price collections.
- Detailed documentation provided on index methodology, compilation procedures, survey design, and conduct of initialization interviews.

### Implementation timetable and milestones (high level)
- Key milestones (from mission forward program):
  - Derive XPI structure, composition and weighting pattern (early October 2013).
  - Select the sample of exporters (early October 2013).
  - Conduct MPI and XPI initialization interviews (end December 2013—subject to funding).
  - Customize the PPI processing system for MPI and XPI (end December 2013).
  - Undertake ongoing quarterly price collection, editing and MXPI calculation.

### Assessment of import unit values (2010–2011)
- Total number of quarterly country unit value series for imports analyzed: 1,273.
- Series assessed as non-credible: 1,041 of the 1,273 series; i.e. 82 percent.
- Assessment approach:
  - Series with violent period-to-period fluctuations were categorized as non-credible.
  - Relatively smooth series with modest fluctuations were categorized as credible.
  - Borderline cases were given the benefit of the doubt and categorized as credible.
  - Assessment was subjective but undertaken by a group of staff; most decisions were clear-cut.

### Change in strategy: need for price surveys
- Because unit values are not suitable, price surveys must be initiated before any index compilation can proceed.
- Essential resources required to fund:
  - a program of import and export price survey initialization interviews; and
  - subsequent ongoing quarterly price collections.
- An intensive field interview program is needed to obtain prices for all selected 8-digit HS items before index compilation.
  - Interviews must be completed within a relatively short time frame so all price series commence in the same quarter and support a common annual time reference period.
  - Retrospective quarterly prices cannot feasibly be obtained too long after the event.

### Survey design and initialization materials
- Generic Import Price Survey and Export Price Survey questionnaires were developed for Malawi.
- Generic forms are to be customized per establishment with:
  - specific product specifications, units, countries of origin or destination, and currency of the transaction identified during initialization interviews.
- Principles for selecting sampled products per importer for quarterly price collection:
  - sampled products to be representative of a wider range of products in terms of price change over time;
  - usually the importer’s largest imports by value;
  - if possible, regularly imported;
  - rule of thumb: maximum of three products to represent an 8-digit HS item from each importer;
  - to price to constant quality and reflect pure price change, full specification of price-determining characteristics, unit of quantity, country of origin and currency of transaction are required.
  - Note: prices are to be converted in the office to Malawi Kwacha for index compilation.
- Supporting documentation developed in the November 2012 mission report:
  - Appendix V: “Field Interviewers’ Instructions”;
  - Appendix VI: “Introductory Letter to Importers”;
  - Appendix VII: “Interview Guide”.

### Steps to initialize Import/Export Price Surveys
- Main steps:
  - contact each importer to arrange an interview;
  - conduct initialization interviews, selecting up to three detailed representative product specifications per selected 8-digit HS code suitable for repeat pricing and fully specifying price-determining characteristics;
  - undertake follow-up action, including re-interviews, as necessary;
  - create a tailored collection form for each importer;
  - on an ongoing quarterly basis, deliver tailored collection forms, undertake collection control and edit and query prices;
  - load prices (after converting to Kwacha (K), as necessary) and compile the indexes.

### Sample selection and pruning for MPI
- Initial identification:
  - 411 8-digit HS codes selected for the MPI basket were identified on the Eurotrace system; 2012 shipment information for each importer transferred to Access.
  - Shipment values aggregated to annual totals for each importer.
  - Importers sorted by value within each 8-digit HS item; percentages and cumulative percentages calculated in Excel.
- Broad business rules for selecting importers:
  - 8-digit HS items categorized by 2012 value:
    - “large” (over 10 billion K),
    - “medium” (1–10 billion K),
    - “small” (less than 1 billion K).
  - For “large” items: three or more importers required and high coverage (70 percent plus if possible).
  - For “medium” items: two importers required and coverage of 50 percent plus if possible.
  - For “small” items: one importer would suffice if it dominated.
- Sample outcomes:
  - After netting duplicates, selected sample of importers: 204.
  - Pruning 4- and 8-digit HS items reduced HS items from 411 to 328 (deletion of 83 HS items, i.e. 20 percent) and importers from 204 to 165 (nearly 20 percent reduction).

### Export sample and staffing plans
- XPI sample considerations:
  - Smaller number of HS items and exporters than MPI.
  - For budgetary purposes, assumed exports sample size approximately half the imports sample: say a maximum of 75 exporters.
- Staffing for initialization interviews:
  - To conduct initialization interviews within 4–6 weeks and achieve a common commencement period, train about 12 MNSO staff from Trade and other areas to work intensively full-time for a short period.
  - Proposed initiation of both import and export price surveys before the end of 2013 so both indexes can have the same reference base period of 2013.
- Funding Proposal and resource estimates referenced in Appendix II.

### Periodicity, timeliness, and publication plans
- Timetable assuming funding and interviews completed before end of 2013:
  - Obtain prices for the four quarters of 2013 and calculate average annual 2013 prices for the reference period.
  - MPI and XPI could be calculated for each of the four quarters of 2013.
  - Ongoing quarterly collection would support calculation of quarterly MPI and XPI from March quarter 2014.
- Publication goals and dissemination policy:
  - Long-term goal: publish indexes at the total and 2-digit HS Chapter level for each index, subject to analysis of the data.
  - Annual and quarterly indexes down to the 8-digit level to be provided to National Accounts on an as-needs basis, with any necessary caveats.
  - Indexes should be fully rebased at least each five years; monitor import and export patterns annually to ensure weighting patterns remain broadly representative.
- Training:
  - Training on index methodology and price collection initialization provided to staff, emphasizing pricing principles, index aggregation, derivation of EA indexes, and application of base weights.

### Producer Price Index (PPI) and CPI follow-up
- PPI review and communications:
  - Review of new PPI addressed index series, compilation procedures, and communication/dissemination strategy.
  - Proposed modification to the application of the two-stage Jevons aggregation formula.
  - Communications strategy elements:
    - tailored briefings to key stakeholders;
    - media briefings; and
    - press releases prior to and at the time of release explaining PPI concepts and relationship to CPI.
  - Proposed publication format at 2-digit ISIC and All index PPI:
    - tables of quarterly index numbers and quarter-to-quarter percentage changes for the 7 quarters from March 2012 to September 2013, plus a graph of quarterly index numbers.
  - Explanatory notes should describe PPI basic concepts, including selling prices for output of manufacturing and utilities, ex-factory prices with valuation basis of “basic prices”, and relationship with CPI which measures “purchasers’ prices”.
  - Main Features should outline main movements in index numbers for latest quarter (September 2013) and main contributors to “All items”.
  - Website release and email distribution planning required.
  - Prepare internal comparison graphs of CPI and PPI (both reference period 2012) and their main components; exclude predominantly exported PPI components from domestic analysis where appropriate.
  - Consider expanding PPI scope to other major industries in the longer term.
- CPI linking and confidentiality:
  - New rebased CPI (COICOP) linked backwards to old series (CPC) for continuity: monthly CPI recompiled retrospectively for two years (2010 and 2011) to provide continuous monthly series from January 2010 to August 2013 for each new publication commodity group.
  - Recompilation earlier than 2010 not feasible due to severe resource constraints.
  - Detailed spreadsheets of monthly indexes of individual items and weights considered confidential under the Statistics Act; access requires contacting the Commissioner via the IMF Resident Representative.

### Project tasks, timetable, and funding estimates (Appendix I and II highlights)
- Key task status and completion timing (2013):
  - Finalize and submit the Funding Proposal: completed.
  - Calculate XPI structure, composition and weighting pattern: completed.
  - Select and analyze the export unit values: completed.
  - Decide whether experimental unit value-based XPI is feasible: completed.
  - Select sample of exporters: completed.
  - Finalize import and export price survey questionnaires: completed.
  - Conduct initialization interviews incl. follow-up: end Dec *.
  - Create tailored forms: end Dec *.
  - Ongoing quarterly price collections: ongoing.
  - Customize the PPI processing system to meet MXPI needs: end Dec.
  - Calculate 2013 average prices for the base period: end Dec *.
  - Enter prices on system: ongoing.
  - Ongoing quarterly compilation: ongoing.
  - * Note: The MNSO is currently under-funded. Any further development work would be subject to the funding proposal being approved and the MNSO budget being increased.
- Funding and sample size estimates (Appendix II):
  - Selected MPI sample of importers: 165.
  - Estimated XPI sample size: about 75.
  - Estimated total number of survey establishments: about 240.
  - Initial estimates for initialization interview costs (travel, allowances, printing, etc.): 5 million K.
  - Estimated annual cost for ongoing quarterly price collections: 10 million K per year.
  - A team of existing staff would undertake initialization interviews over a 3-week period, ideally before the end of 2013.

*Prepared by David Collins, Prices Expert, East AFRITAC. October 25, 2013.*

### 2013. The views expressed in this document are those of the staff team and do not necessarily

### MALAWI — REPORT ON THE PRICE STATISTICS MISSION September 16–27, 2013

### Executive summary and mission objectives
- Mission undertaken in response to a request from the Malawi authorities, in consultation with the African Department (AFR), during September 16–27, 2013; activity 14STJ6300 funded by East AFRITAC and part of project Price Statistics STA_MWI_2008_05.
- Main objectives:
  - Assist with improving prices statistics, including full conversion to direct pricing for import and export price indexes (MXPI).
  - Review the new producer price index (PPI).
  - Prepare detailed plans for quarterly import and export price surveys and the replacement of unit values with survey-based prices.
  - Provide training on data collection and index compilation.
  - Document methodology and computational procedures developed.
- Focus placed on the development of an import price index (MPI); documentation prepared is for MPI but applicable to XPI.

### Key findings on source data and feasibility
- A detailed analysis of 2010 and 2011 quarterly, country-specific import unit values selected for the experimental MPI found that over 80 percent of the country unit values did not provide credible indicators of pure price change over time.
- Less than 20 percent of the series provided credible indicators of pure price change over time.
- Result: compilation of the MPI could not proceed using Customs-based unit values; it is not feasible to proceed until quarterly transaction prices are available from import and export price surveys.
- Number of 8-digit HS items selected for pricing in the MPI: 411 out of a potential 6,000 plus HS items.
- Country-mix elimination improves unit-value data quality, but the existing unit values remain insufficiently credible for index compilation.

### Recommended statistical prerequisites and immediate actions
- Initiate price surveys before any index compilation can be undertaken. Resources must be provided to fund:
  - A program of import and export price survey initialization interviews.
  - Subsequent ongoing quarterly price collections.
- An intensive program of field interviews is required to obtain necessary prices for all selected 8-digit HS items before index compilation can start.
- Initialization interviews should be undertaken within a relatively short time frame so all price series can commence in the same quarter and support a common index time reference period; retrospective quarterly prices are generally not feasible too long after the event.
- A funding proposal has been prepared (Appendix II).

### Methodology, concepts, and compilation decisions
- Concepts and definitions consistent with System of National Accounts 2008 (2008 SNA) and the Export and Import Price Index Manual were confirmed.
- Aggregation and formulae:
  - Elementary aggregate (EA) indexes at the 8-digit level of the HS.
  - Unweighted price relatives calculated using the Jevons geometric mean aggregation formula at EA level.
  - EA indexes aggregated to higher HS levels (4-digit, 2-digit, total) with fixed weights using the Laspeyres formula.
- Weighting:
  - Index weights will be smoothed by basing them on the average value of imports over the period 2007–2010; 2011 was not considered a “normal” year.
  - In the longer term, MXPI should be re-referenced to the same base year as the CPI and PPI.
- Reference period:
  - Initial plan was 2010 as time reference period, but because survey prices must be collected before compilation, a more recent period is required.
  - If initialization interviews can be completed before the end of 2013, request prices for March, June, September, and December quarters of 2013 and make 2013 the time reference period.
  - If interviews are delayed, a later time reference period will be selected and index compilation deferred.
- Processing systems:
  - The PPI processing system contains required functionality for MXPI; work identified to tailor the PPI system to MXPI requirements.

### Sampling and survey design
- A sample of importing establishments was selected for the import price survey using cut-off sampling techniques.
- MNSO to proceed to select a sample of exporters for the export price survey.
- Generic questionnaires were designed for a quarterly Survey of Import Prices and a Survey of Export Prices.
- Questionnaires designed for customization during initialization interviews to capture specific product specifications, units, country of origin/destination, and currency of transaction.

### Training, documentation, and capacity building
- Training provided on index concepts, compilation methodology, and price survey design and implementation.
- Emphasis on training for price survey initialization interviews and subsequent quarterly price collections.
- Detailed documentation provided on index methodology, compilation procedures, survey design, and conduct of initialization interviews.

### Implementation timetable and milestones (high level)
- Key milestones (from mission forward program; see Appendix I):
  - Derive XPI structure, composition and weighting pattern (early October 2013).
  - Select the sample of exporters (early October 2013).
  - Conduct MPI and XPI initialization interviews (end December 2013—subject to funding).
  - Customize the PPI processing system for MPI and XPI (end December 2013).
  - Undertake ongoing quarterly price collection, editing and MXPI calculation.

### PPI and CPI issues addressed
- New PPI reviewed; a broad communication and dissemination strategy developed, including publication content.
- CPI linking: issue of linking the new, rebased CPI series with the old series backwards at a component level was addressed.

### Policy recommendations and resource implications
- Essential that resources be provided to fund import and export price survey initialization interviews and ongoing quarterly collections.
- Prioritize completion of initialization interviews so a common 2013 time reference period can be used (assuming completion before end-2013).
- Customize and deploy the PPI processing system for MXPI compilation as planned.
- Apply the MPI development procedures and survey design also to XPI, noting that XPI development is generally more straightforward than MPI.

*Prepared by David Collins, Prices Expert, East AFRITAC. October 25, 2013.*

### 33.      The current mission undertook a thorough analysis of the quarterly time series of the

### _cr13359 - 33.      The current mission undertook a thorough analysis of the quarterly time series of the

### Assessment of import unit values (2010–2011)
- Total number of quarterly country unit value series for imports analyzed: 1,273.
- Series assessed as non-credible: 1,041 of the 1,273 series; i.e. 82 percent.
- Assessment approach:
  - Series with violent period-to-period fluctuations were categorized as non-credible.
  - Relatively smooth series with modest fluctuations were categorized as credible.
  - Borderline cases were given the benefit of the doubt and categorized as credible.
  - Assessment was subjective but undertaken by a group of staff; most decisions were clear-cut.
- Identified primary causes of non-credible unit values (even after eliminating country mix):
  - compositional mix of products; and
  - incorrect quantity data (misreporting and Customs not editing closely, switching of units, etc.).
- Conclusion: Serious problems with Customs quantity data; ongoing liaison between the MNSO and Customs will be required to improve the data.

### Change in strategy: need for price surveys
- Because unit values are not suitable, price surveys must be initiated before any index compilation can proceed.
- Essential resources required to fund:
  - a program of import and export price survey initialization interviews; and
  - subsequent ongoing quarterly price collections.
- An intensive field interview program is needed to obtain prices for all selected 8-digit HS items before index compilation.
  - Interviews must be completed within a relatively short time frame so all price series commence in the same quarter and support a common annual time reference period.
  - Retrospective quarterly prices cannot feasibly be obtained too long after the event.

### Survey design and initialization materials
- Generic Import Price Survey and Export Price Survey questionnaires were developed for Malawi.
- Generic forms are to be customized per establishment with:
  - specific product specifications, units, countries of origin or destination, and currency of the transaction identified during initialization interviews.
- Principles for selecting sampled products per importer for quarterly price collection:
  - sampled products to be representative of a wider range of products in terms of price change over time;
  - usually the importer’s largest imports by value;
  - if possible, regularly imported;
  - rule of thumb: maximum of three products to represent an 8-digit HS item from each importer;
  - to price to constant quality and reflect pure price change, full specification of price-determining characteristics, unit of quantity, country of origin and currency of transaction are required.
  - Note: prices are to be converted in the office to Malawi Kwacha for index compilation.
- Supporting documentation developed in the November 2012 mission report:
  - Appendix V: “Field Interviewers’ Instructions”;
  - Appendix VI: “Introductory Letter to Importers”;
  - Appendix VII: “Interview Guide”.

### Steps to initialize Import/Export Price Surveys
- Main steps:
  - contact each importer to arrange an interview;
  - conduct initialization interviews, selecting up to three detailed representative product specifications per selected 8-digit HS code suitable for repeat pricing and fully specifying price-determining characteristics;
  - undertake follow-up action, including re-interviews, as necessary;
  - create a tailored collection form for each importer;
  - on an ongoing quarterly basis, deliver tailored collection forms, undertake collection control and edit and query prices;
  - load prices (after converting to Kwacha (K), as necessary) and compile the indexes.

### Sample selection and pruning for MPI
- Initial identification:
  - 411 8-digit HS codes selected for the MPI basket were identified on the Eurotrace system; 2012 shipment information for each importer transferred to Access.
  - Shipment values aggregated to annual totals for each importer.
  - Importers sorted by value within each 8-digit HS item; percentages and cumulative percentages calculated in Excel.
- Broad business rules for selecting importers:
  - 8-digit HS items categorized by 2012 value:
    - “large” (over 10 billion K),
    - “medium” (1–10 billion K),
    - “small” (less than 1 billion K).
  - For “large” items: three or more importers required and high coverage (70 percent plus if possible).
  - For “medium” items: two importers required and coverage of 50 percent plus if possible.
  - For “small” items: one importer would suffice if it dominated.
- Sample outcomes:
  - After netting duplicates, selected sample of importers: 204.
  - Pruning 4- and 8-digit HS items reduced HS items from 411 to 328 (deletion of 83 HS items, i.e. 20 percent) and importers from 204 to 165 (nearly 20 percent reduction).

### Export sample and staffing plans
- XPI sample considerations:
  - Smaller number of HS items and exporters than MPI.
  - For budgetary purposes, assumed exports sample size approximately half the imports sample: say a maximum of 75 exporters.
- Staffing for initialization interviews:
  - To conduct initialization interviews within 4–6 weeks and achieve a common commencement period, train about 12 MNSO staff from Trade and other areas to work intensively full-time for a short period.
  - Proposed initiation of both import and export price surveys before the end of 2013 so both indexes can have the same reference base period of 2013.
- Funding Proposal and resource estimates referenced in Appendix II.

### Periodicity, timeliness, and publication plans
- Timetable assuming funding and interviews completed before end of 2013:
  - Obtain prices for the four quarters of 2013 and calculate average annual 2013 prices for the reference period.
  - MPI and XPI could be calculated for each of the four quarters of 2013.
  - Ongoing quarterly collection would support calculation of quarterly MPI and XPI from March quarter 2014.
- Publication goals and dissemination policy:
  - Long-term goal: publish indexes at the total and 2-digit HS Chapter level for each index, subject to analysis of the data.
  - Annual and quarterly indexes down to the 8-digit level to be provided to National Accounts on an as-needs basis, with any necessary caveats.
  - Indexes should be fully rebased at least each five years; monitor import and export patterns annually to ensure weighting patterns remain broadly representative.
- Training:
  - Training on index methodology and price collection initialization provided to staff, emphasizing pricing principles, index aggregation, derivation of EA indexes, and application of base weights. See Appendices III and IV.
- Work program:
  - A work program for development of the MXPI is provided in Appendix I.

### Producer Price Index (PPI) and CPI follow-up
- PPI review and communications:
  - Review of new PPI addressed index series, compilation procedures, and communication/dissemination strategy.
  - Proposed modification to the application of the two-stage Jevons aggregation formula.
  - Communications strategy elements:
    - tailored briefings to key stakeholders;
    - media briefings; and
    - press releases prior to and at the time of release explaining PPI concepts and relationship to CPI.
  - Proposed publication format at 2-digit ISIC and All index PPI:
    - tables of quarterly index numbers and quarter-to-quarter percentage changes for the 7 quarters from March 2012 to September 2013, plus a graph of quarterly index numbers.
  - Explanatory notes should describe PPI basic concepts, including selling prices for output of manufacturing and utilities, ex-factory prices with valuation basis of “basic prices”, and relationship with CPI which measures “purchasers’ prices”.
  - Main Features should outline main movements in index numbers for latest quarter (September 2013) and main contributors to “All items”.
  - Website release and email distribution planning required.
  - Prepare internal comparison graphs of CPI and PPI (both reference period 2012) and their main components; exclude predominantly exported PPI components from domestic analysis where appropriate.
  - Consider expanding PPI scope to other major industries in the longer term.
- CPI linking and confidentiality:
  - New rebased CPI (COICOP) linked backwards to old series (CPC) for continuity: monthly CPI recompiled retrospectively for two years (2010 and 2011) to provide continuous monthly series from January 2010 to August 2013 for each new publication commodity group.
  - Recompilation earlier than 2010 not feasible due to severe resource constraints.
  - Detailed spreadsheets of monthly indexes of individual items and weights considered confidential under the Statistics Act; access requires contacting the Commissioner via the IMF Resident Representative.

### Project tasks, timetable, and funding estimates (Appendix I and II highlights)
- Key task status and completion timing (2013):
  - Finalize and submit the Funding Proposal: completed.
  - Calculate XPI structure, composition and weighting pattern: completed.
  - Select and analyze the export unit values: completed.
  - Decide whether experimental unit value-based XPI is feasible: completed.
  - Select sample of exporters: completed.
  - Finalize import and export price survey questionnaires: completed.
  - Conduct initialization interviews incl. follow-up: end Dec *.
  - Create tailored forms: end Dec *.
  - Ongoing quarterly price collections: ongoing.
  - Customize the PPI processing system to meet MXPI needs: end Dec.
  - Calculate 2013 average prices for the base period: end Dec *.
  - Enter prices on system: ongoing.
  - Ongoing quarterly compilation: ongoing.
  - * Note: The MNSO is currently under-funded. Any further development work would be subject to the funding proposal being approved and the MNSO budget being increased.
- Funding and sample size estimates (Appendix II):
  - Selected MPI sample of importers: 165.
  - Estimated XPI sample size: about 75.
  - Estimated total number of survey establishments: about 240.
  - Initial estimates for initialization interview costs (travel, allowances, printing, etc.): 5 million K.
  - Estimated annual cost for ongoing quarterly price collections: 10 million K per year.
  - A team of existing staff would undertake initialization interviews over a 3-week period, ideally before the end of 2013.

*Source: _cr13359 - 33.      The current mission undertook a thorough analysis of the quarterly time series of the*

---


_Source: https://www.imf.org/-/media/websites/imf/imported-full-text-pdf/external/pubs/ft/scr/2013/_cr13359.pdf_