## cr17295 - Executive Summary

## Source details

**Canonical URL:** [cr17295 - Executive Summary](https://www.imf.org/-/media/files/publications/cr/2017/cr17295.pdf)

## Other formats

- [Markdown version](/-/media/files/publications/cr/2017/cr17295.pdf.md)
- [Structured JSON version](/-/media/files/publications/cr/2017/cr17295.pdf.json)

---

### Mission context and objectives
- TA mission to Kampala during June 30–July 18, 2014 to assist with finalizing the rebased Producer Price Index (PPI) and developing Export and Import Price Indices (XMPI).
- Activity: 15ST053:00; project: Prices Statistics STA_UGA_2013_05.
- Principal tasks:
  - Review rebased PPI compilation methodology.
  - Assess PPI series and prices data and develop a dissemination strategy.
  - Develop XMPI compilation methodology and determine data sources.
  - Select base period, basket and weights.
  - Train staff on methodology and surveys.
  - Prepare methodological and computational documentation.

### PPI and XMPI scope, pricing points, and valuation
- PPI scope and pricing:
  - Extended beyond manufacturing to include mining and quarrying and electricity and water industries.
  - Pricing point: ex-factory, ex-mine, etc.; valuation basis: basic prices (exclusive of trade and transport margins and Value Added Tax (VAT)).
  - Index relates to both domestic usage and exports; separate prices to be collected for exports and domestic usage as appropriate.
  - Short–medium term: extend scope to include agriculture; longer-term: consider major business service activities.
- XMPI scope and pricing:
  - Will relate to actual transaction prices for exports and imports of merchandise trade.
  - Pricing points: XPI = Free on Board (FOB); MPI = Cost Insurance Freight (CIF).

### Reference periods, baskets, and weights
- PPI:
  - Reference base period (time reference period for the new index): June quarter 2013.
  - Weighting base period: 2009–10 (reference period for the most recent Supply Use Tables (SUT)).
  - Proposed rebasing frequency: fully rebased on a 5-yearly basis using the latest finalized SUT.
  - First level aggregation: 4-digit ISIC class level.
  - Twenty-eight 4-digit ISIC class industries selected as the index “basket.”
- MPI:
  - First level aggregation: 8-digit Harmonized System (HS) level.
  - 151 8-digit HS items selected as the MPI index basket.
  - MPI weights derived from value of imports smoothed over the three years 2011 to 2013 using a top-down approach.
- XMPI reference base consideration:
  - 2012 tentatively selected as the time reference period for the XMPI, subject to change given resource and initialization interview timeframes.
  - Common reference period required for XPI and MPI to facilitate Terms of Trade (TOT) calculation.

### Aggregation formulas and processing system
- Elementary Aggregates (EA):
  - EA indices calculated using the Jevons geometric mean price index formula (unweighted geometric mean of price ratios).
  - Unweighted EA formula chosen because maintaining lower level product and establishment weights continuously is not feasible.
- Higher-level aggregation:
  - Progressive aggregation to 3-digit, 2-digit, 1-digit ISIC and all-groups levels using the Laspeyres formula and derived weighting system.
- Processing system:
  - Proposed use of a modified and enhanced version of the current PPI Excel-based processing system.
  - EA formula in the Excel aggregation module will need to be replaced with the Jevons formula.
  - “Access” module (data capture and sample management) reviewed.

### Data collection, sampling, and timing — PPI (Producer Price Survey)
- Periodicity and reference:
  - Monthly periodicity; pricing point at the 15th of the month (or the closest previous trading day).
  - Annual indices calculated as the simple average of the monthly indices.
  - Retrospective monthly prices collected from March 2013 to March 2014; planned initial monthly indices compilation and publication will commence with April 2013, linked as far as possible to the current index.
- PPS questionnaire and implementation:
  - Replace complex multi-purpose PPS questionnaire with the streamlined PPS mid-month prices form (already commenced).
  - New PPS collects monthly mid-month product prices using a streamlined questionnaire.
  - Monthly prices are being collected retrospectively from March 2013 during PPS initialization interviews; reference period set to June quarter 2013.
- Manufacturing sample:
  - Original selection: sample size approximately 250 manufacturing establishments (4-digit ISIC selection with a general cut-off of 90 percent).
  - Revised selection:
    - Regimen of 28 ISIC classes categorized into 6 large classes (5 percent or more of the total weight), 17 medium classes (between 1 and 5 percent), and 5 small classes (less than 1 percent).
    - New business-rule cut-offs: 90 percent for the large industries; 70 percent for both the medium and small industries.
    - Resulting sample size: 176 manufacturing establishments (included supplementation by reference to the URA “largest taxpayers” and local knowledge).
  - Recommendation: annual review of the sample of establishments using the Uganda Business Register to add growing businesses and remove contracted ones.
- Initialization status:
  - Original schedule: completion by end-September 2013.
  - Actual start: mid-June 2014 due to corporate funding cuts.
  - To date about one third of the interviews have been conducted; program needs rescheduling to relate to the revised sample.
  - Field observations: scope to streamline procedures; some establishments could not be located; some major “manufacturers” were actually wholesalers and incorrectly classified.
- Publication goals:
  - Minimum publication requirement: separate indices for total sales and total sales excluding exports.
  - Publish total sales indices at least at the 2-digit Division level of ISIC.
  - If user demand confirms, publish export and local sales indices by 2-digit ISIC Division and higher-level aggregations.
  - Dissemination objective: publish the new index on a very timely basis (ideally prior to the release of the CPI).

### Data collection, sampling, and timing — MPI and MPS
- MPI indicators and MPS design:
  - Quarterly MPI price indicators will comprise:
    - Customs unit values for relatively homogeneous commodities (sometimes for individual countries or with outliers removed).
    - Survey-based prices from an Import Price Survey (MPS) for heterogeneous items.
  - Quarterly unit value trends for 2011–2013 for the 151 selected 8-digit HS commodities were analyzed and categorized as either suitable as price indicators or unsuitable (survey-based prices required).
- Selection and division of 151 HS items:
  - Of the 151 selected 8-digit HS items:
    - 45 were designated as having credible price indicators (Customs unit values suitable).
    - 106 required survey-based price indicators (MPS).
- MPS sampling rules and product selection:
  - Broad business rules for importer sampling:
    - (i) at least two importers required for each commodity, if possible;
    - (ii) ideally achieve significant coverage (greater than 50 percent if feasible), particularly for large commodities;
    - (iii) if very low concentration dominated by many small importers, select the largest importer then apply a skip interval to select a representative sample of small importers;
    - (iv) priority to achieve good coverage and representation of high value commodities;
    - (v) for relatively small commodities with low concentration, delete the 8-digit item from the regimen and pro-rate its weight amongst the remaining 8-digit items in the 4-digit Heading.
  - Principles for selecting products (paragraph 81):
    - Sampled products must be representative of a wider range of products in terms of price change over time.
    - Selected products will usually be the importers’ largest imports by value.
    - For workload reasons, restrict to a manageable number of products (say, restricted to 3 or 4, if possible) per importer per 8-digit HS item.
    - Full specification of price-determining characteristics required: physical product characteristics, unit of quantity, transaction details such as country of origin and currency of the transaction.
    - Prices are to be converted in the office to Ugandan Shillings for index compilation.
- Initialization and ongoing MPS operations (paragraphs 82–83):
  - Documentation: Appendix III (“MPS Field Interviewers’ Instructions”), Appendix IV (“Introductory Letter to Importers”), Appendix V (“MPS Interview Guide”), Appendix VI (“Example of a Generic MPS Questionnaire”).
  - Main operational steps:
    - Contact each importer to arrange interview.
    - Conduct initialization interviews; select up to three or four representative product specifications per 8-digit code.
    - Create tailored collection form for each importer.
    - On an ongoing quarterly basis deliver tailored forms, undertake collection control, edit and query prices.
    - Regularly confirm and update representativeness; load prices (after converting to Shillings) and compile indices.

### MPI index structure, weight derivation, and selection process
- Weighting base and smoothing:
  - Time series analyzed: annual import patterns from 2009 to 2013.
  - Smoothing decision: use averages from more than one year.
  - Considered distributions: 3-year average (2011 to 2013) and 5-year average (2009 to 2013).
  - Decision: the 3-year average (2011 to 2013) judged most credible and broadly representative.
  - Control aggregate: total value of imports over 2011 to 2013.
- Top-down selection steps:
  - Step 1 (2-digit HS Chapters):
    - Aim: around 80 percent or more direct coverage; selected Chapters contributing 1.0 percent or more initially.
    - Result: selection of 19 HS Chapters out of 96, accounting for some 86 percent of total imports over 2011 to 2013.
    - Selected 2-digit items and cumulative percentages over 2011–2013:
      - 27 Mineral fuels, etc. — 23.0 — 23.0
      - 84 Boilers, machinery, etc. — 9.3 — 32.3
      - 87 Vehicles — 9.0 — 41.3
      - 85 Electrical machinery — 8.1 — 49.5
      - 30 Pharmaceuticals — 4.9 — 54.4
      - 72 Iron and steel — 4.3 — 58.7
      - 15 Fats and oils — 4.2 — 62.9
      - 39 Plastics — 4.0 — 66.9
      - 10 Cereals — 2.5 — 69.4
      - 17 Sugars — 2.5 — 71.9
      - 25 Salt, cement, etc. — 2.2 — 74.1
      - 48 Paper — 2.1 — 76.2
      - 90 Precision instruments — 1.8 — 78.0
      - 63 Other textiles, etc. — 1.6 — 79.6
      - 38 Miscellaneous chemicals — 1.4 — 81.0
      - 22 Beverages — 1.2 — 82.5
      - 73 Articles of iron and steel — 1.2 — 83.7
      - 40 Rubber — 1.1 — 84.8
      - 33 Essential oils — 1.1 — 85.9
  - Step 2 (4-digit Headings):
    - Selected major 4-digit items within each selected 2-digit item aiming for some 80 percent coverage.
    - Resulted in selection of 63 items at the 4-digit level.
    - Adjusted 2-digit values prorated among selected 4-digit items.
  - Step 3 (8-digit Commodities):
    - Selected dominant 8-digit items within each selected 4-digit item aiming at high coverage.
    - Resulted in selection of 151 8-digit items.
    - Adjusted 4-digit values prorated among selected 8-digit items.
- Outcome:
  - Structure, composition and weighting pattern for the MPI derived using smoothed import values over 2011 to 2013.
  - Weights to be presented in percentage terms.
  - As a check, sum of 151 adjusted 8-digit HS items equaled the total value of imports over the period; no values left “floating.”
- XPI:
  - UBOS to repeat same process; expect a much smaller sample of items for XPI because of trading patterns.

### Accuracy, reliability, conceptual decisions, and validation
- Concepts and definitions reviewed and ratified; consistent with the 2008 SNA and the international PPI Manual.
- Separate indices/tables recommended for domestic sales, exports and total output at successive levels of the index industrial hierarchy and total manufacturing (current practice).
- Minimum analytical requirement: PPI for total sales and PPI for total output excluding exports.
- Methodology devised to compute separate exports and domestic sales structures and weighting patterns using establishment-reported percentages from the “Monthly Survey, Initial Survey 2014” form.
- Weights credibility:
  - Weights need credibility checks against local knowledge and Ugandan Revenue Authority (URA) information on “largest taxpayers.”
  - Substantial shifts from present weights (based on the 2000–01 Ugandan Business Inquiry (UBI)) must be justified.
  - Detailed methodology documented in June 2013 prices mission report.
- Customs unit values quality control:
  - Eliminating country mix from 8-digit unit values (using individual country unit values or removing outlier countries) can improve credibility.
  - Analysis repeated at individual country of origin unit values; in some cases country-specific selection or outlier removal made Customs data suitable and reduced dependence on survey data.

### Operational findings and immediate recommendations
- Replace the complex multi-purpose PPS questionnaire with the streamlined PPS mid-month prices form (already commenced).
- Modify the Excel-based aggregation module to implement Jevons EA formula and retain Laspeyres for higher levels.
- Complete and accelerate initialization interviews program given retrospective price collection to March 2013 and the June quarter 2013 reference period.
- Proceed with selection and initialization of MPS samples for the 151 8-digit HS items classified as requiring survey-based prices.
- Prepare and submit XMPI work program and budget to senior management to secure resources.
- Apply credibility checks to derived weights using URA and local knowledge; justify shifts from prior UBI-based weights.
- Review the sample of importers in light of general industry knowledge and undertake an annual review to add new/expanding importers and remove contracted importers.

### Serviceability, periodicity, rebasing, dissemination, training, and resources
- Periodicity and span:
  - Proposed that 2012 based XMPI be compiled quarterly from March 2012 and subsequently annually.
- Publication goals:
  - Publish indices at the 2-digit HS Chapter level together with total indices, subject to analysis of the data.
  - Provide annual and quarterly indexes down to the 8-digit level to National Accounts on an as-needs basis.
- Rebasing:
  - Indices should be fully rebased at least each five years.
  - Monitor annual patterns to assess whether rebasing should occur sooner than five years.
- Training:
  - Training provided on price index concepts, methodology and price collection initialization and compilation (see Appendix VII).
  - Emphasis on pricing principles (constant quality, transaction prices, product sampling), EA index derivation, and applying base weights.
- Work program and resource constraints:
  - A work program provided in Appendix II; target completion dates contingent upon availability of resources.
  - Further progress on the MPI project delayed until September 2014 because of staff commitment to the Population Census.

### Key tasks and selected timetables (selected milestones and exact deadlines)
- PPI Project (Appendix I) selected milestones:
  - Consult further with major users on need for separate exports and local usage tables by industry: end-July.
  - Finalize the new sample of establishments: end-July.
  - Review the interviewing strategy: end-July.
  - Re-schedule and prioritize the interview program: end-July.
  - Commence interviewing program: beg-Aug.
  - Amend, upgrade and test the PPI processing system: end-Aug*.
  - Finalize the new index structure and weights, incl. credibility checks and rationalization of major shifts: end-Sep.
  - Complete the interview program for the new PPS sample and create tailored collection forms: end-Oct*.
  - Compile the new monthly PPI and commence publication for the period April 2013 to February 2015: end-Mar.
  - Note: * Benchmark Actions prior to any further TA on the topic (Appendix I).
- XMPI Project (Appendix II) selected milestones:
  - 2014:
    - Review MPI structure and weights: end-Sep.
    - Review processing system: end-Sep.
    - Compile Customs-based MPI: end-Sep*.
    - Finalize MPS commodities: end-Oct.
    - Select MPS sample of importers: end-Oct.
    - Finalize survey strategy including collection form: end-Oct.
    - Prepare budget and work plan: end-Oct.
    - Train field staff: end-Nov.
    - Pilot survey: end-Nov.
  - 2015:
    - Conduct program of field interviews: end-Jan*.
    - Replace Customs unit values with survey prices: end-Feb*.
    - Recompile MPI: end-Feb.
    - Progressively repeat above for XPI: end-Mar.
    - Analyze results and develop dissemination strategy: end-Apr.
    - Compile TOT: end-Apr.
    - Ongoing compilation: ongoing.
  - Note: *Benchmark Actions prior to any further TA on the topic (Appendix II).

### Field operations, interview instructions, and questionnaire highlights
- Initial contact and telephone script (Appendix III):
  - Send an introductory letter to each selected importer and enter identification and 8-digit HS codes on an Interview Guide.
  - Telephone contact script explains NBS import price survey purpose, confidentiality, legal authority (Statistics Act 2002), and requests appointment within the next seven days to initialize quarterly mail collection.
- Interview content and product selection (Appendix III and V):
  - Up to three specific product descriptions per 8-digit code, specifying country of origin, unit, currency, and CIF price.
  - CIF prices required in currency of transaction; office will convert to TZS.
  - Target pricing dates: 15th of the middle month each quarter (i.e., 15th of February, May, August, and November). If 15th is not a trading day, report shipment closest and prior to that date.
- Product pricing rules and logistics (Appendix III, V, VI):
  - Prices recorded must be actual transaction prices, valued on a CIF basis.
  - Preferably prices as at the 15th of the middle month of the quarter, or nearest previous trading day.
  - For infrequent shipments select a shipment as close as possible but prior to the 15th of the middle month of the quarter.
- Generic questionnaire (Appendix VI) fields:
  - HS Code, full description, unit of quantity, country of origin, currency of transaction, CIF import price per unit as at 15th of the month.
  - Ask whether prices changed since previous quarter and, if YES, provide reasons.

### Training program highlights (Appendix VII)
- Core topics covered:
  - Mission report; Ten Steps in PP/TPI (XPI and MPI) Development; PPI Technical Notes; Long Term System of Price Indices.
  - International XMPI and PPI Manual.
  - What a price index measures: price movements over time, not levels; temporal vs. spatial indexes.
  - Confidentiality and legal framework: Statistics law; index nos. not prices; publication rules.
  - Main elements of a price index: time reference period (= 100 or 100.0); prices; samples; market transaction prices; index structure; weighting base period and weights.
  - System relationships: PPI, CPI, MPI, XPI; pricing points.
  - Analytical role of MPI and XPI: upstream signals for downstream household inflation; consumer goods vs. capital goods; leakage and entry of imports/exports.
  - Pricing principles: constant quality, quality adjustment, specification pricing, transaction pricing, including discounts.
  - Model pricing of unique products and sampling principles.
  - Editing, analysis, and publication goals.

*IMF mission report content as provided (cr17295 - Executive Summary).*

### Executive Summary ......................................................................................................

### cr17295 - Executive Summary

### Mission context and objectives
- TA mission to Kampala during June 30–July 18, 2014 to assist with finalizing the rebased Producer Price Index (PPI) and developing Export and Import Price Indices (XMPI).
- Activity: 15ST053:00; project: Prices Statistics STA_UGA_2013_05.
- Principal tasks:
  - Review rebased PPI compilation methodology.
  - Assess PPI series and prices data and develop a dissemination strategy.
  - Develop XMPI compilation methodology and determine data sources.
  - Select base period, basket and weights.
  - Train staff on methodology and surveys.
  - Prepare methodological and computational documentation.

### Key determinations and conceptual bases
- PPI scope:
  - Extended beyond manufacturing to include mining and quarrying and electricity and water industries.
  - Pricing point: ex-factory, ex-mine, etc.; valuation basis: basic prices (exclusive of trade and transport margins and Value Added Tax (VAT)).
  - Index relates to both domestic usage and exports; separate prices to be collected for exports and domestic usage as appropriate.
  - Short–medium term: extend scope to include agriculture; longer-term: consider major business service activities.
- XMPI scope:
  - Will relate to actual transaction prices for exports and imports of merchandise trade.
  - Pricing points: Export Price Index (XPI) = Free on Board (FOB); Import Price Index (MPI) = Cost Insurance Freight (CIF).

### Reference periods, baskets, and weights
- PPI reference base period: June quarter 2013 (time reference period for the new index).
- Weighting base period for PPI: 2009–10 (reference period for the most recent Supply Use Tables (SUT)).
- Proposed rebasing frequency for PPI: fully rebased on a 5-yearly basis using the latest finalized SUT.
- PPI index composition:
  - First level aggregation: 4-digit ISIC class level.
  - Twenty-eight 4-digit ISIC class industries selected as the index “basket.”
- XMPI (MPI) composition:
  - First level aggregation: 8-digit Harmonized System (HS) level.
  - 151 8-digit HS items selected as the MPI index basket.
  - MPI weights derived from value of imports smoothed over the three years 2011 to 2013 using a top-down approach.

### Aggregation and formulas
- Elementary Aggregate (EA) indices:
  - Calculated using the Jevons geometric mean price index formula (unweighted geometric mean of price ratios).
  - An unweighted EA formula selected because maintaining lower level product and establishment weights continuously is not feasible for reasons of data availability and cost.
- Higher-level aggregation:
  - Progressive aggregation to 3-digit, 2-digit, 1-digit ISIC and all-groups levels using the Laspeyres formula and derived weighting system.
- Current processing system:
  - Proposed use of a modified and enhanced version of the current PPI Excel-based processing system.
  - EA formula in the Excel aggregation module will need to be replaced with the Jevons formula.
  - “Access” module (data capture and sample management) reviewed.

### Data collection, sampling, and timing
- Producer Price Survey (PPS):
  - New PPS collects monthly mid-month product prices using a streamlined questionnaire.
  - Monthly prices are being collected retrospectively from March 2013 during PPS initialization interviews; reference period set to June quarter 2013.
  - Planned publication: monthly indices at the ISIC Division level and above, separately for total sales, exports and local sales.
  - Sample review and re-selection reduced the manufacturing sample size from 250 to 176.
  - Initialization interviews delayed by nearly a year due to corporate funding cuts; interview program rescheduled and streamlined where possible.
  - Initialization interviews undertaken by experienced UBOS Statistical Assistant field staff trained on PPS methodology.
- Import Price Survey (MPS) and MPI indicators:
  - Quarterly MPI price indicators will comprise:
    - Customs unit values for relatively homogeneous commodities (sometimes for individual countries or with outliers removed).
    - Survey-based prices from an Import Price Survey (MPS) for heterogeneous items.
  - Quarterly unit value trends for 2011–2013 for the 151 selected 8-digit HS commodities were analyzed and categorized as either suitable as price indicators or unsuitable (survey-based prices required).
  - For the MPS, a sample of importers needs to be selected for each 8-digit item and a program of initialization interviews conducted to commence the quarterly MPS; business rules for sample selection were developed.
  - Business rules to support selection of sample of importers for the MPS (and subsequently exporters for the XPS) were developed.

### Accuracy, reliability, and methodological decisions
- Concepts and definitions:
  - Reviewed and ratified; consistent with the 2008 SNA and the international PPI Manual.
- Price collection and market breakdown:
  - Separate indices/tables recommended for domestic sales, exports and total output at successive levels of the index industrial hierarchy and total manufacturing (current practice); minimum analytical requirement: PPI for total sales and PPI for total output excluding exports.
  - Methodology devised to compute separate exports and domestic sales structures and weighting patterns using establishment-reported percentages from the “Monthly Survey, Initial Survey 2014” form.
- Credibility and validation:
  - Weights need credibility checks against local knowledge and Ugandan Revenue Authority (URA) information on “largest taxpayers.” Substantial shifts from the present weights (based on the 2000–01 Ugandan Business Inquiry (UBI)) must be justified.
  - Detailed methodology for derivation of index structure and weights documented in June 2013 prices mission report.

### Implementation planning, training, and resource needs
- Work programs prepared:
  - For further development of the new PPI reflecting delays in PPS field interviews.
  - For the XMPI, including programs of survey initialization interviews.
- Resource planning:
  - A work program and budget need to be prepared and submitted to senior management to support the XMPI project.
- Training:
  - Provided to PPI and XMPI work groups on price index concepts, compilation methodology (including splicing and linking techniques), and price survey procedures (see Appendix VII).
- Regional coordination:
  - Several East African countries are developing PPIs and XMPIs; periodic structured meetings to share experiences would be valuable.

### Operational findings and immediate recommendations
- Replace complex multi-purpose PPS questionnaire with the streamlined PPS mid-month prices form (already commenced).
- Modify the Excel-based aggregation module to implement Jevons EA formula and retain Laspeyres for higher levels.
- Complete and accelerate initialization interviews program given retrospective price collection to March 2013 and the June quarter 2013 reference period.
- Proceed with selection and initialization of MPS samples for the 151 8-digit HS items classified as requiring survey-based prices.
- Prepare and submit XMPI work program and budget to senior management to secure resources.
- Apply credibility checks to derived weights using URA and local knowledge; justify shifts from prior UBI-based weights.

*cr17295 - Executive Summary*

### 30. The decision has also been made to collect mid-month prices on a monthly rather

### cr17295 - 30. The decision has also been made to collect mid-month prices on a monthly rather

### Periodicity, pricing point, and index span
- Both the periodicity and the frequency of the new PPS will be monthly.
- The pricing point will be at the 15th of the month (or the closest previous day on which there was a transaction).
- Annual indices will be calculated as the simple average of the monthly indices.
- Retrospective monthly prices were collected from March 2013 to March 2014; price collection for interviewed establishments must be updated and more recent prices sought for those yet to be interviewed.
- Planned initial monthly indices compilation and publication will commence with April 2013, linked as far as possible to the current index.
- After a time series is established, indices should be re-referenced to a calendar year and ultimately coordinated across PPI, CPI and national accounts on a common reference base period.

### Sampling of manufacturing establishments (PPS)
- Initial sample selection:
  - Selection undertaken at the 4-digit ISIC class level using a combination of cut-off and judgmental sampling.
  - A general cut-off of 90 percent was applied, resulting in a sample size of some 250 manufacturing establishments.
- Revised sample selection:
  - The regimen of 28 ISIC classes was categorized into:
    - 6 large classes (5 percent or more of the total weight)
    - 17 medium classes (between 1 and 5 percent)
    - 5 small classes (less than 1 percent)
  - New business-rule cut-offs applied:
    - 90 percent for the large industries
    - 70 percent for both the medium and small industries
  - Resulting sample size: 176 manufacturing establishments (included supplementation by reference to the URA “largest taxpayers” and local knowledge).
- Recommendations on sample maintenance:
  - An annual review of the sample of establishments should be undertaken to introduce new businesses or growing ones and remove establishments that have contracted, using the Uganda Business Register.

### Product representativeness and form changes
- A question should be added to the monthly collection form, after the question about reasons for any price change, confirming representativeness of products (example wording: “Are these products still representative volume sellers?”).
- This question should be asked during each month’s field collection and replacement products selected as necessary for splicing into the index compilation system.

### Price collection initialization: status and actions
- Original schedule: initialization interviewing program to be completed by end-September 2013.
- Actual start: resources allocated to commence the initialization project in mid-June 2014 due to cuts in corporate funding.
- To date, about one third of the interviews have been conducted; the program needs rescheduling to relate to the revised sample of establishments.
- Field experience observations:
  - Scope exists to streamline interviewing procedures.
  - Some establishments could not be located (likely address changes and outdated UBOS records).
  - Some major “manufacturers” were actually wholesalers and incorrectly classified.
- Action: develop a revised program of initialization interviews accounting for interviewing experience, the revised sample, and a review of interview conduct.
- Assessment: sample of product specification descriptions obtained to date was judged adequate to support pricing to constant quality.
- Reference: Paragraphs 58 to 62 and Appendices IV–VII of the June 2013 mission report describe steps in price collection initialization and pricing principles.

### Publication goals, dissemination, and training
- Minimum publication requirement: separate indices for total sales, and total sales excluding exports.
- At minimum, publish total sales indices at least at the 2-digit Division level of ISIC.
- If user demand confirms, publish export and local sales indices by 2-digit ISIC Division and higher-level aggregations.
- Dissemination objective: publish the new index on a very timely basis (ideally prior to the release of the CPI) to support its role as a forward-looking inflation indicator and serve National Accounts needs.
- Training delivered:
  - Comprehensive training on price index concepts, compilation methodology, price survey initialization and ongoing collection techniques.
  - Particular emphasis on “splicing” and linking of index series on different time reference periods.

### XMPI (Export and Import Price Indices) strategy and design
- Programming strategy: modify and improve current PPI spreadsheet processing to support XMPI and the new PPI.
- XMPI design: hybrid indices combining Customs unit values for more homogeneous commodities and survey-based prices for other commodities obtained through XMPS.
- Ideal XMPS implementation: intensive initialization interviews over a short period (say 1–2 months) including collection of retrospective prices to the selected time reference period to supplement Customs unit values and enable hybrid indices as soon as possible.
- If intensive funding is not available, progressive implementation over an extended period may be necessary, with the disadvantage of limiting feasible retrospective price collection (perhaps no more than 4 or 5 quarters) and potentially forcing a very recent reference period.
- Focus during the mission: development of an MPI (documentation relates to imports); same techniques apply to XPI. MPI development and maintenance expected to be more complex than XPI due to broader commodity range, heterogeneity, and “lumpy” import patterns.

### Accuracy, reliability, and conceptual basis for XMPI
- Pricing point:
  - XPI: Free on Board (FOB)
  - MPI: Cost Insurance and Freight (CIF)
- Reference base period:
  - Feasibility of obtaining annual average prices for a recent year considered.
  - 2012 tentatively selected as the time reference period for the XMPI, subject to change given resource and initialization interview timeframes.
  - Note: common reference period required for XPI and MPI to facilitate Terms of Trade (TOT) calculation.

### Aggregation and compilation methodology
- EA indices at the 8-digit level of the HS.
- Unweighted indices: Jevons price index formula (ratio of unweighted price indicators which are either survey-based prices or Customs unit values).
- Rationale: an unweighted EA formula selected because it is not feasible to maintain lower level product and establishment weights on a continuous basis.
- Aggregation: EA indices aggregated with fixed weights using the Laspeyres formula.
- Weight calculation methodology described in paragraphs 63–69.

### Weighting base period and rebasing decisions
- Time series analyzed: annual import patterns from 2009 to 2013.
- Observed: percentage distributions of import values between 2-digit HS Chapters fluctuated year-to-year; no clear trend.
- Smoothing decision: use averages from more than one year.
- Considered distributions:
  - 3-year average (2011 to 2013)
  - 5-year average (2009 to 2013)
- Decision: the 3-year average (2011 to 2013) judged most credible and broadly representative of import trading patterns over the next few years.
- Structure, composition and weighting pattern for the MPI derived using smoothed import values over 2011 to 2013.

### Methodology for deriving index structure, composition, and weights
- Control aggregate: total value of imports over 2011 to 2013.
- Top-down approach aimed to maximize indirect representation of un-priced items via chain of representation.
- Step 1 (2-digit HS Chapters):
  - Aim: around 80 percent or more direct coverage (in practice all Chapters contributing 1.0 percent or more to total imports were initially selected).
  - Result: selection of 19 HS Chapters out of 96.
  - These 19 Chapters accounted for some 86 percent of total imports over 2011 to 2013.
  - Selected 2-digit items and cumulative percentages over 2011–2013:
    - 27 Mineral fuels, etc. — 23.0 — 23.0
    - 84 Boilers, machinery, etc. — 9.3 — 32.3
    - 87 Vehicles — 9.0 — 41.3
    - 85 Electrical machinery — 8.1 — 49.5
    - 30 Pharmaceuticals — 4.9 — 54.4
    - 72 Iron and steel — 4.3 — 58.7
    - 15 Fats and oils — 4.2 — 62.9
    - 39 Plastics — 4.0 — 66.9
    - 10 Cereals — 2.5 — 69.4
    - 17 Sugars — 2.5 — 71.9
    - 25 Salt, cement, etc. — 2.2 — 74.1
    - 48 Paper — 2.1 — 76.2
    - 90 Precision instruments — 1.8 — 78.0
    - 63 Other textiles, etc. — 1.6 — 79.6
    - 38 Miscellaneous chemicals — 1.4 — 81.0
    - 22 Beverages — 1.2 — 82.5
    - 73 Articles of iron and steel — 1.2 — 83.7
    - 40 Rubber — 1.1 — 84.8
    - 33 Essential oils — 1.1 — 85.9
- Step 2 (4-digit Headings):
  - Selected major 4-digit items within each selected 2-digit item, aiming for some 80 percent coverage.
  - Resulted in selection of 63 items at the 4-digit level.
  - Adjusted 2-digit values prorated among selected 4-digit items.
- Step 3 (8-digit Commodities):
  - Selected dominant 8-digit items within each selected 4-digit item, aiming at high coverage.
  - Resulted in selection of 151 8-digit items.
  - Adjusted 4-digit values prorated among selected 8-digit items.
- Outcome:
  - Structure and composition (the basket) of the MPI and relative values within 2-digit, 4-digit and 8-digit hierarchy form the fixed weighting pattern.
  - Weights to be presented in percentage terms.
  - As a check, sum of 151 adjusted 8-digit HS items equaled the total value of imports over the period; no values left “floating.”
- For the XPI:
  - The same process will be repeated by UBOS.
  - Because of trading patterns, a much smaller sample of items will be selected at each HS level for the XPI.

### Source data and selection of price indicators
- Plan to use Customs unit values for relatively homogeneous items where quarter-to-quarter unit value variations credibly measure pure price change, and survey-based prices for non-homogeneous items via a quarterly MPS.
- Data quality note: eliminating country mix from 8-digit unit values (using individual country unit values or removing outlier countries) can improve credibility.
- Analysis of quarterly unit values for the 151 selected HS items over 2011 to 2013:
  - Initial assessment of suitability for each unit value series; smoothing of a few quarters’ outlier values sometimes improved credibility.
  - Assessment repeated at individual country of origin unit values; in some cases country-specific selection or outlier removal made Customs data suitable and reduced dependence on price survey data.
- Result:
  - Of the 151 selected 8-digit HS items:
    - 45 were designated as having credible price indicators (Customs unit values suitable).
    - 106 required survey-based price indicators (MPS).
- Commodity concentration observation:
  - Dominance of sophisticated machinery and equipment among imports, apart from Mineral fuels (Chapter 27).
  - Chapter 84, Chapter 87 and Chapter 85 contained 80 of the 106 8-digit commodities selected for the MPS.
- Importer sampling for MPS (for each of the 106 HS commodities):
  - Sample selection from Customs data using a combination of cut-off and judgmental sampling.
  - Broad business rules:
    - (i) at least two importers required for each commodity, if possible;
    - (ii) ideally achieve significant coverage (greater than 50 percent if feasible), particularly for large commodities;
    - (iii) if very low concentration dominated by many small importers, select the largest importer then apply a skip interval to select a representative sample of small importers;
    - (iv) priority to achieve good coverage and representation of high value commodities;
    - (v) for relatively small commodities with low concentration, delete the 8-digit item from the regimen and pro-rate its weight amongst the remaining 8-digit items in the 4-digit Heading.

*IMF mission report content as provided.*

### 79. This sample of importers should now be reviewed in the light of general industry

### Price collection initialization/review (cr17295 - 79)

### Sample selection and review
- The sample of importers should be reviewed against general industry knowledge to ensure a comprehensive and representative sample of current trading activity (paragraph 79).
- An annual review of the sample of importers should be undertaken to:
  - enable the introduction of new importers or existing ones that have expanded, and
  - remove importers who have contracted (paragraph 80).

### Principles for selecting products for direct price collection (paragraph 81)
- Sampled products must be representative of a wider range of products in terms of price change over time.
- Selected products will usually be the importers’ largest imports by value.
- For workload reasons, a manageable number of products (say, restricted to 3 or 4, if possible) should be selected to represent an 8-digit HS item from each importer, especially if an importer has been selected for a large number of items.
- To price to constant quality over time and reflect pure price change, full specification of all price-determining characteristics is required, including:
  - physical product characteristics,
  - unit of quantity,
  - transaction details such as country of origin and currency of the transaction.
- Prices are to be converted in the office to Ugandan Shillings for index compilation (paragraph 81).

### Documentation to support initialization (paragraph 82)
- Appendix III: “MPS Field Interviewers’ Instructions.”
- Appendix IV: “Introductory Letter to Importers.”
- Appendix V: “MPS Interview Guide.”
- Appendix VI: “Example of a Generic MPS Questionnaire” to be modified to UBOS standards.

### Main operational steps for initialization and ongoing collection (paragraph 83)
- Contact each importer to arrange an interview.
- Conduct interviews to initialize ongoing collection; select for each importer some three or four detailed representative product specifications for each 8-digit code suitable for repeat pricing; ensure full specification of 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 the tailored collection forms, undertake collection control and editing and query prices.
- Regularly confirm and update the representativeness of the sample of products with each importer.
- Load the prices (after converting to Shillings) and compile the indices.

---

### Serviceability: periodicity, dissemination, rebasing, training, and work program

### Periodicity and span (paragraph 84)
- It is proposed that 2012 based XMPI be compiled quarterly from March 2012 and subsequently annually.

### Publication goals and dissemination policy (paragraphs 85–86)
- Goal: publish indices at the 2-digit HS Chapter level together with total indices, subject to analysis of the data.
- Annual and quarterly indexes down to the 8-digit level should be provided to National Accounts on an as-needs basis, with any necessary caveats.

### Rebasing (paragraph 87)
- Indices should be fully rebased at least each five years.
- Monitor annual patterns of imports and exports to assess whether significant changes in trend indicate rebasing should occur sooner than five years.

### Training (paragraph 88)
- Training on price index concepts and methodology and price collection initialization and compilation was provided to staff (see Appendix VII).
- Emphasis included pricing principles (constant quality, transaction prices, product sampling), index aggregation processes such as derivation of EA indexes, and the application of base weights to calculate indices at higher aggregation levels.

### Work program and resource constraints (paragraph 89)
- A work program is provided in Appendix II.
- Target completion dates are contingent upon availability of resources.
- Further progress on the MPI project will be delayed until September 2014 because of staff commitment to the Population Census.

---

### Key tasks and timetables (Appendix I and II excerpts)

- Appendix I: PPI Project tasks and timetable (2014–2015) — selected milestones and exact deadlines:
  - Continue the ongoing PPS collection and PPI compilation and publication: ongoing.
  - Consult further with major users on the need for separate exports and local usage tables by industry: end-July.
  - Finalize the new sample of establishments (including rationalizing with the old sample): end-July.
  - Review the interviewing strategy: end-July.
  - Re-schedule and prioritize the interview program: end-July.
  - Commence interviewing program: beg-Aug.
  - Amend, upgrade and test the PPI processing system: end-Aug*.
  - Commence liaison with Agriculture Directorate: end-Aug.
  - Finalize the new index structure and weights, incl. credibility checks and rationalization of major shifts from the old weighting pattern: end-Sep.
  - If separate exports and local sales tables are needed:
    - 10.1 derive separate weights, and
    - 10.2 amend processing system: end-Sep*.
  - Complete the interview program for the new PPS sample and create tailored collection forms: end-Oct*.
  - Conduct the monthly PPS ongoing: 2015.
  - Compile the new monthly PPI and commence publication for the period April 2013 to February 2015: end-Mar.
  - Note: * Benchmark Actions prior to any further TA on the topic (Appendix I).

- Appendix II: XMPI project tasks and timetable (2014–2015) — selected milestones and exact deadlines:
  - 2014:
    - Review MPI structure and weights: end-Sep.
    - Review processing system: end-Sep.
    - Compile Customs-based MPI: end-Sep*.
    - Finalize MPS commodities: end-Oct.
    - Select MPS sample of importers: end-Oct.
    - Finalize survey strategy including collection form: end-Oct.
    - Prepare budget and work plan: end-Oct.
    - Train field staff: end-Nov.
    - Pilot survey: end-Nov.
  - 2015:
    - Conduct program of field interviews: end-Jan*.
    - Replace Customs unit values with survey prices: end-Feb*.
    - Recompile MPI: end-Feb.
    - Progressively repeat above for XPI: end-Mar.
    - Analyze results and develop dissemination strategy: end-Apr.
    - Compile TOT: end-Apr.
    - Ongoing compilation: ongoing.
  - Note: *Benchmark Actions prior to any further TA on the topic (Appendix II).

---

### Field operations: interviewers’ instructions, interview guide, and questionnaire

### Initial contact and telephone script (Appendix III)
- Send an introductory letter to each selected importer and enter identification and 8-digit HS codes on an Interview Guide.
- Telephone contact: explain the Import Price Survey with the exact script provided, including:
  - NBS conducting an import price survey to collect selected product price data from a sample of importers in Tanzania.
  - Workload kept to a minimum; prices for a small number of representative specific products will be requested each quarter.
  - Uses: build MPIs, support inflation management, compile national accounts volume statistics for Tanzania, derive Terms of Trade estimates.
  - Confidentiality: all individual information kept strictly confidential as required by law; only consolidated price index data released.
  - Authority: survey conducted under the Statistics Act 2002 and NBS seeks cooperation.
- Make an appointment within the next seven days to explain data collection, agree on products, and initialize quarterly mail collection.

### Interview content and product selection (Appendix III and V)
- Explain purpose, scope, sampling approach (up to three for each 8-digit HS code), and need for tightly specified product characteristics (make, model, size, quality, technical specifications).
- Specify unit of quantity, size lot, country/countries of origin, and currency of the transaction to ensure comparability and measure pure price change.
- Products should usually be the company’s largest imports in value terms.

### Product pricing rules and logistics (Appendix III, V, VI)
- Prices recorded must be actual transaction prices, valued on a CIF basis (inclusive of overseas freight and insurance) (paragraph 7.1).
- Preferably prices should be as at the 15th of the middle month of the quarter, or the nearest previous trading day (if a holiday) (paragraph 7.2).
- If shipments are infrequent select a shipment for a day as close as possible, but prior to, the 15th of the middle month of the quarter (paragraph 7.3).
- Seek to obtain prices for the selected products as at the 15th of the middle month of the previous quarter during the interview (paragraph 7.4).
- Logistics: confirm contact details, agree person responsible for filling form and handling queries, undertake follow-up/re-interviewing as necessary, create tailored forms, send form for final clearance, and deliver the form each quarter for price collection (paragraphs 7.5–7.9).

### Interview Guide fields (Appendix V)
- Collect identification: company name, code, address, contact name, phone, email.
- Record Harmonized System codes selected for the company.
- For each 8-digit code select up to three specific product descriptions, specifying country of origin, unit, currency, and CIF price.
- CIF prices required in currency of transaction; office will convert to TZS.
- Target pricing dates: 15th of the middle month each quarter (i.e., 15th of February, May, August, and November) (Appendix V, Part E).
- If 15th is not a trading day, report the shipment closest and prior to that date.

### Generic questionnaire instructions (Appendix VI)
- Record import prices (CIF) per unit in terms of the currency of the transaction.
- Import prices should be the actual prices paid for a shipment on the 15th of the middle month of the Quarter; if not feasible, use the closest prior shipment.
- Include table fields: HS Code, full description, unit of quantity, country of origin, currency of transaction, CIF import price per unit as at 15th of the month.
- Ask whether prices changed since previous quarter and, if YES, provide reasons.

---

### Training program highlights (Appendix VII)
- Core topics covered:
  - Documentation: Mission report; Ten Steps in PP/TPI (XPI and MPI) Development; PPI Technical Notes; Long Term System of Price Indices.
  - International XMPI and PPI Manual.
  - What a price index measures: price movements over time, not levels; temporal vs. spatial indexes.
  - Confidentiality and legal framework: Statistics law; index nos. not prices; publication rules.
  - Main elements of a price index: time reference period (= 100 or 100.0); prices; samples; market transaction prices; index structure; weighting base period and weights.
  - System relationships: PPI, CPI, MPI, XPI; pricing points (output indices, input indices); valuation bases.
  - Analytical role of MPI and XPI: upstream signals for downstream household inflation; consumer goods vs. capital goods; leakage and entry of imports/exports.
  - Pricing principles: constant quality, quality adjustment, specification pricing, transaction pricing, including discounts.
  - Model pricing of unique products and sampling principles.
  - Editing, analysis, and publication goals.

*Source: cr17295 - 79. This sample of importers should now be reviewed in the light of general industry (cr17295 - 79. This sample of importers should now be reviewed in the light of general industry).*

---


_Source: https://www.imf.org/-/media/files/publications/cr/2017/cr17295.pdf_