This Technical Assistance Report paper on Georgia highlights various aspects of producer price indexes (PPI) mission. The purpose of the mission was to assist the National Statistics Office of Georgia (Geostat) with expanding the coverage of the PPI to additional services activities. Expanded PPI coverage will facilitate more accurate measurement of real output generated from services activities in Georgia. Since services activities comprised 58.4 percent of 2018 gross domestic product on a value-added basis, improved measurement in this area will significantly enhance the accuracy of total measures of the size, growth, and productivity of the Georgian economy. Improved PPIs will also provide policymakers with a clearer view of price development at the producer level, which may signal future changes in the cost of living for Georgian consumers. Geostat analysis indicated that prices for all carriers have historically changed at similar rates. This makes the consumer price index for air transportation an adequate estimator for producer prices, which should include domestic airlines only. However, PPI prices should ideally include sales to all buyer types, not just resident consumers. This makes unit value prices an attractive option. These can be calculated as total fare revenue generated from all passengers that flew in a specified class of service on all flights for a specified route, divided by this same number of passengers.

Abstract

This Technical Assistance Report paper on Georgia highlights various aspects of producer price indexes (PPI) mission. The purpose of the mission was to assist the National Statistics Office of Georgia (Geostat) with expanding the coverage of the PPI to additional services activities. Expanded PPI coverage will facilitate more accurate measurement of real output generated from services activities in Georgia. Since services activities comprised 58.4 percent of 2018 gross domestic product on a value-added basis, improved measurement in this area will significantly enhance the accuracy of total measures of the size, growth, and productivity of the Georgian economy. Improved PPIs will also provide policymakers with a clearer view of price development at the producer level, which may signal future changes in the cost of living for Georgian consumers. Geostat analysis indicated that prices for all carriers have historically changed at similar rates. This makes the consumer price index for air transportation an adequate estimator for producer prices, which should include domestic airlines only. However, PPI prices should ideally include sales to all buyer types, not just resident consumers. This makes unit value prices an attractive option. These can be calculated as total fare revenue generated from all passengers that flew in a specified class of service on all flights for a specified route, divided by this same number of passengers.

Summary of Mission Outcomes and Priority Recommendations

1. The purpose of the mission was to assist the National Statistics Office of Georgia (Geostat) with expanding the coverage of the producer price index (PPI) to additional services activities. This was the fourth price statistics mission to Georgia conducted under the auspices of the three-year Project to Improve National Accounts and Price Statistics in Eastern and Southeastern Europe. This project is funded by the Government of the Netherlands.

2. Expanded PPI coverage will facilitate more accurate measurement of real output generated from services activities in Georgia. Since services activities comprised 58.4 percent of 2018 GDP on a value-added basis, improved measurement in this area will significantly enhance the accuracy of total measures of the size, growth, and productivity of the Georgian economy. Improved PPIs will also provide policymakers with a clearer view of price development at the producer level, which may signal future changes in the cost of living for Georgian consumers.

3. Geostat has made exceptional progress since the first project mission in 2018 and is now ready to disseminate new PPIs for transportation and storage services and telecommunications services in reference period January 2020. This achievement is the result of strong commitment and resourceful workplan execution from Geostat staff and management. This included successful collection of new surveys for warehousing and telecommunications producer prices in 2019. The new indexes will be published along with existing series for freight transportation as part of a separate services PPI release. These new indexes are well aligned with measuring Georgia’s increasing role as a regional transportation and logistics hub.

4. Published data for reference period January 2020 release will also feature an update of the classification structure from the Classification of Products by Activity (CPA) 1996 to the CPA 2008. This update significantly improves the coherence of PPIs with national accounts and business statistics data.

5. Producer prices for travel agency and tour operator services will be directly collected beginning in January 2020, with a goal of publishing results in reference period January 2021. Adding coverage for these activities will provide new detailed data on market developments in the growing tourism sector in Georgia. During the mission, Geostat finalized new data collection procedures and questionnaires for these service activities. Lists of companies that reported sales of these services in the business statistics program were reviewed, and corrections to classifications were made. Geostat is now ready to train fieldworkers and start data collection.

6. Future PPI expansion should continue with a focus on better data for the key transportation and tourism sectors. New indexes for support services for transportation and accommodation services would support this outcome. Geostat should begin collection and calculation of these indexes in reference period January 2021, with a focus on coverage of cargo handling services and airport operations. Port operations services may also be considered for future coverage, while producer prices for food and beverage services may be estimated based on consumer price index (CPI) data.

7. To offset the costs of expanding PPI coverage, Geostat should consider implementing efficiencies in sampling respondent businesses that would have minimal impact on the quality of aggregate indexes. This includes selecting samples at a less detailed classification level, and truncating coverage of detailed product areas that have minimal weight in their relevant aggregate group. Geostat should also evaluate whether annual re-selection of new samples of businesses and products generate benefits in terms of index quality that justify the considerable staff resource and respondent burden costs.

8. To support progress in the above work areas, the mission recommended a detailed action plan with the following priority recommendations to improve the PPI:

Table 1.

Priority Recommendations

article image

Detailed Technical Assessment and Recommendations

A. Classification and Coverage Updates—January 2020

9. With the release of the January 2020 reference period data in February 2020, Geostat will update PPIs to the classification CPA 2008. The classification change will result in some additional indexes for new detailed categories within the industrial sector. These new industrial PPIs will be published beginning in the January 2020 reference period with the annual average of 2010 equal to 100. Indexes for CPA 2008 categories were calculated back to 2010, but Geostat is not planning to re-publish this historical data. These data were, however, provided to national accounts compilers.

10. Geostat will also introduce new PPIs for telecommunications and transportation services in reference period January 2020. This will mark the successful on time completion of phases one and two of the project to expand PPI coverage to new services activities that began in 2018. The new data series will be introduced in a distinct press release for services PPIs. This release should be organized based on the CPA 2008 classification structure, with detail to the three-digit product category level. Indexes will be published with the annual average of 2019 equal to 100.

11. Geostat has produced indexes for freight transportation services since 2008. With the transition to CPA 2008, these freight transportation services will now appear in multiple index categories, depending on whether they are offered by rail, road, or air. It will be useful to maintain the historical time series for freight transportation services as a special index outside of the CPA 2008 structure, with the annual average of 2010 equal to 100. This could be paired with a new aggregated index for all passenger transportation services across all travel modes.

12. The new aggregate index for CPA H, Transportation and storage services, will include less than 50 percent coverage of product output. This should be noted in the data release. Coverage is not yet available for CPA 52.2, Support services for transportation, which comprises more than half of total output generated in section H. This product category, which includes cargo handling services and the operation of ports and airports, is a suitable target for future coverage expansion.

13. The new transportation and storage services index will include coverage for CPA 49, Land transport; CPA 51, Air transport; and CPA 52.1, Warehousing and storage. Freight transportation activities and warehousing and storage services will be covered by directly collected producer prices. Since prices for warehousing and storage services have been successfully collected since January 2019. the annual average of 2019 can also be used as the reference period for this detailed index. Producer prices for passenger transportation services will be estimated based on the relevant CPIs. Adjustments will be applied when tax rates change to ensure they do not affect the PPI. CPA 50, Water transport is not included since there is very little Georgian production in this area.

14. Geostat should consider directly collecting producer prices for passenger air transportation in the future. Passenger fares on all domestic and foreign airlines are eligible for inclusion in the CPI. Geostat analysis indicated that prices for all carriers have historically changed at similar rates. This makes the CPI for air transportation an adequate estimator for producer prices, which should include domestic airlines only. However, PPI prices should ideally include sales to all buyer types, not just resident consumers. This makes unit value prices an attractive option. These can be calculated as total fare revenue generated from all passengers that flew in a specified class of service on all flights for a specified route, divided by this same number of passengers.

15. The new index for CPA 61, Telecommunications activities will consist of detailed indexes for wired, wireless, satellite, and other telecommunications activities. This corresponds with the three-digit level of CPA detail. The category for satellite telecommunications will not be published to avoid data disclosure for a dominant company but will be included in the aggregate calculations.

Recommendations:

  • Organize the new January 2020 services PPI release based on the CPA 2008 publication structure with detail to the 3-digit product category level, and the annual average of 2019 equal to 100.

  • Indicate in the release that the aggregate index for CPA H, Transportation and storage services, will include less than 50 percent coverage of product output.

  • Maintain existing time series for freight transportation services as a special index outside of the CPA 2008 structure, with annual average of 2010 equal to 100.

  • Consider direct collection of unit value producer prices for passenger air transportation services in the future.

B. Coverage Expansion to Travel Agency and Tour Operator Services

16. Data collection for a new index covering CPA 79.1, Travel agency and tour operator services will begin in January 2020. The mission assisted Geostat staff with drafting two separate questionnaires for travel agency and tour operator services, respectively. Separate questionnaires were developed because these two services are transacted differently. Tour operators take responsibility for virtually all aspects of a trip, often providing local transportation and guide services directly. Tour operators in Georgia generate a substantial majority of their sales from nonresident visitors. Travel agents, however, do not take operational responsibility for trip execution. They simply act as middlemen connecting travel service providers to interested travelers. Travel agents in Georgia generate most of their sales from assisting Georgian residents with travelling abroad.

17. The sampling frame from the business statistics survey was reviewed and many tour operator companies were found to be misclassified as travel agencies, and vice versa. To ensure companies are correctly classified, fieldworkers should contact each selected company and clarify their activities before the relevant questionnaire is administered. Corrected classification information should be shared with the business statistics program when available.

18. The price collected for tour operators should represent the gross fees received for operating the tour. Tour operators receive bundled fees encompassing charges for local transportation, lodging, tour guide, food, and attraction admission charges. Tour operators often provide local transportation and guide services directly, while lodging and food services are most commonly provided by hotels and restaurants. In general, producer prices should be collected on a net basis if the company does not provide the services themselves or does not take capacity risk if services are unsold. However, since tour operators bundle each of these services together and are not able to report prices as a bundle of net and gross fees, it is acceptable to collect the gross price of the tour as the producer price. Since tour operators frequently offer discounts without updating list prices, unit value prices are a good option if available. This can be calculated as the total revenue generated from all customers that took a specific tour in a month divided by the number of these customers. If tour operators are unable to provide a unit value price, a transaction price should be collected reflecting the tour price for a single customer in a double occupied room. Specifications should detail the services included, the tour destinations, class of accommodation if included, and the quality of the local transportation and guide services (e.g. air-conditioned coach, English speaking guide, etc.).

19. Net commission and/or flat fee transaction prices should be collected for travel agent services. Each transaction should represent the fee received for booking a specific plane ticket, hotel room, car rental, or other travel service. If commissions are charged, the price should be calculated as the current price for the travel services multiplied by the commission rate. Each period respondents should update both the commission rate and the travel service price that would be charged if a similar traveler were to book the same itinerary. Specifications should detail the booked travel, include the travel provider (e.g. airline, hotel), destination, class of service.

20. New index series for travel agency and tour operator services should be introduced in reference period January 2021, with the 2020 annual average index equal to 100, pending results of pilot data collection. This provides enough time to execute the survey on a pilot basis and modify the procedures as necessary based on results. If the pilot survey results are found to be inadequate, adjustments can be made prior to January 2021, and December 2020 can be used as the index reference period instead of the 2020 annual average.

Recommendations:

  • A new sample of travel agent and tour operators should be fielded in January 2020.

  • Fieldworkers should contact each selected company and clarify classification of their activities before the relevant questionnaire is administered.

  • Gross prices inclusive of the full cost of travel should be collected for tour operators. Unit value prices should be collected if available.

  • Commissions or fees net of the full cost of travel should be collected for travel agents. Both commission rate percentages and the cost of travel should be updated in future periods to obtain the monetary value of commissions.

  • The new index series should be disseminated in reference period January 2021 with annual average of 2020 equal to 100, pending results of the pilot data collection.

C. Additional Future Coverage Expansion

21. Future PPI expansion should continue to CPA product category 52.2, Support services for transportation. These services contribute more than half of total transportation and storage output. This product area is very diverse, including many detailed six-digit product categories each with different price determining characteristics. Coverage should focus on the largest products only, such as airport operation and cargo handling services. As an area of potential growth, port operations should also be considered. Since there are a small number of large port and airport operators, data collections costs should be low. It will, however, be critically important to thoroughly cover the operations of these few large producers, either with unit value prices or more than four selected products. A few separate questionnaires will likely be needed to accurately collect the diverse products in this group. Research and questionnaire preparation should start in 2020, with collection and calculation of the indexes in reference period January 2021.

22. Geostat should also expand coverage of tourism activities by creating new PPIs for CPA product category 55.1, Hotel and similar accommodation services. This category is currently covered in the CPI, but that index is based only on resident consumer expenditures. The PPI would cover all domestic production, including sales to nonresident consumers and businesses. The outlets where Georgian consumers purchase accommodation services are likely to vary substantially from the outlets where nonresident tourists and business travelers stay. While it may be possible to obtain some useful prices from the CPI, a distinct sample of hotels for the PPI is recommended. CPI data are, however, acceptable to use as a proxy of producer prices for food and beverage serving services.

Recommendations:

  • Begin research for future expansion to coverage of support services for transportation and hotel and similar accommodation services in 2020.

  • Collect and calculate new PPIs for these activities in reference period January 2021.

D. Sampling

23. Geostat identifies businesses to be surveyed for the domestic PPI and export index by selecting a subsample of business statistics respondents by six-digit CPA. The subsample is selected using a cutoff of 80 percent of weighted production value. The weights are taken from the business statistics program, where they are used to represent non-selected businesses. However, the business statistics weights are based on a sample design that includes geographic detail that are not relevant to the PPI and that covers activities at less detailed product levels. In the services sector, business survey weights may represent non-selected respondents that primarily offer products not covered by the PPI.

24. As a future improvement, Geostat should instead select cutoff samples using unweighted production value directly from the business register. Turnover may be used if production values are not available in the business register. Since cutoff sampling is not a probability proportionate-to-size or stratified probability selection method, sampling weights are not needed. The companies that are not selected because their cumulative contribution to product output is below the cutoff level are truncated from coverage. Moving to a cutoff sample of unweighted output from the business register is likely to result in larger sample sizes compared to current methods.

25. Geostat can achieve more efficient allocation of sample units by allowing for more flexibility in the level of CPA detail at which samples are selected. Many six-digit product categories include only a few Georgian producers and contribute small shares of output at the aggregated two-digit level. But the current sample design selects producers from each of these categories. For transportation and telecommunications services, sample selection at the three-digit level of CPA detail (e.g. wireless telecommunications, freight rail transportation services, etc.) will achieve adequate coverage of sector production. This change is likely to result in smaller sample sizes compared to current methods.

26. A cutoff approach for including three-digit product categories can also be used to prevent allocating data collection resources to activities that represent a small share of output. For example, activities can be included until 80 percent coverage of the two-digit product aggregate is achieved. Those activities that contribute small amounts of production would therefore be input based on the prices for products that generate more output.

27. Geostat’s resampling of all PPI producers and products every year may not produce benefits commensurate with the data collection and respondent burden costs. This is because the producers in most product categories are stable year-to-year. Since the cutoff sampling method is used, the next largest producer can be added to the sample whenever an existing producer stops or changes production. Geostat can also periodically review the results of the business survey and add new producers of significant size. During the mission technical assistance was provided on changing the composition of items in elementary aggregates without impacting price trends. Geostat should evaluate the costs and benefits of full annual resampling in comparison to this alternative approach. Annual updates of index activity weights should continue regardless of the methods used to select producers and products.

Recommendations:

  • Geostat should transition to selecting cutoff product samples based on unweighted production values taken directly from the business register.

  • To improve efficiency of data collection, Geostat should allow for more flexibility in the level of CPA detail that is used for selecting cutoff samples of producers.

  • Geostat should consider applying a cutoff method to determine coverage of detailed activities within two-digit aggregates.

  • Geostat should evaluate the costs and benefits of annual PPI resampling in comparison to alternative approaches.

E. Data Review and Dissemination

28. Geostat has developed outstanding tools to identify impactful price changes and outlier price observations in all price indexes, including PPIs. These tools employ data visualization to guide staff to observations that may require further evaluation. This is a significant improvement that facilitates efficient and accurate review of collected prices prior to index publication.

29. Currently, separate press releases are disseminated on the same day for the total PPI (domestic plus export), domestic PPI, export index, and import index. The domestic PPI refers to production made and sold in the domestic market or stored as inventory. These releases can be consolidated to improve data accessibility, promote contextual analysis across all series, and make the dissemination process more efficient.

30. Geostat should consolidate these five releases down to the following two: one for industrial activities and another for services activities. The industrial activities release would highlight two aggregations: the total PPI (domestic plus export), and a new domestic supply index (domestic plus import). The individual components would then be described in detail. The services activities release would not include detailed indexes for exports or imports, since they are not available. The services indexes should be noted as inclusive of all production of domestic producers, including sales to residents and to the export market. Indexes in this release will have variable index reference periods, as coverage is expanded to new services activities in future years.

31. The PPI press releases adequately summarize relevant information about aggregate monthly and annual index changes, and the most impactful detailed product contributors. The releases could be improved by incorporating visualizations similar to those currently used in the CPI release. These visualizations clearly present key index changes and contributions in a user-friendly summary. Geostat should also modify the measures of monthly and annual change in the releases from index values to percentage points. For example, if a PPI increased from an index value of 150 to 200 in a given month, the increase would currently be reflected as one month change of 133.3, rather than 33.3 percent. The latter presentation would be clearer.

32. Geostat has set a laudable goal of expanding the user base for PPI statistics. One effective approach is to highlight the value of the PPI indexes as tools for contract adjustment by Georgian businesses. A description of how PPIs could be used for this purpose can be posted on the Geostat website, highlighted with a press release, and communicated to key business organizations. Links to similar notices prepared by other national statistical offices are provided in the appendix of this report.

Recommendations:

  • Consolidate PPI and related index releases to two: one for industrial activities and one for services activities.

  • Allow for variable index reference periods in the services activities release.

  • Incorporate visualizations similar to those currently used for the CPI in the PPI releases.

  • Modify presentation of monthly and annual changes to percentage points.

  • Prepare a document highlighting how PPIs can be used for contract adjustment and share it with potential users in the business community.

F. Recommendation Action Plan

Table 2.

Recommendation Action Plan

article image

G. Officials Met During the Mission

article image

Appendix I. Links to Examples of Documents Explaining the Use of PPIs for Contract Adjustment

Australia

https://www.abs.gov.au/websitedbs/D3310114.nsf/home/Inflation+and+Price+Indexes+-+Use+of+Price+Indexes+in+Contracts

France

https://www.insee.fr/en/information/2491441

Netherlands

https://business.gov.nl/running-your-business/products-and-services/sales-and-advertising/adjust-your-prices-with-a-price-index/

United States

https://www.bls.gov/ppi/ppiescalation.htm

Appendix II. Monthly Questionnaire for Travel Agency Services

Statistical Survey of Service Producer Prices for the Activities of Travel Agencies Questionnaire N 06.2.10.4

(monthly)

article image

The table contains the following fields:

  • Name of Service– in all types of travel agency’s services the top four type of services with the highest share (e.g airline ticket reservation, hotel reservation, car rental, full package tour, etc.), indicated in the beginning of the survey;

  • Service Characteristics – the main characteristics of the indicated service that affect the price formation:

    • In case of booking airline tickets – airline, destination, class;

    • In case of booking hotel – name of hotel, class (the number of stars), type of room in the hotel;

    • In case of car rental – company, location, type of car;

    • In case of full package tour – name of travel agency, duration of tour, destination;

  • Measurement Unit – measurement unit of the type of travel service (e.g. one package tour, one airline ticket, etc.);

  • Fixed Fee for Service – in case the price is set as the fixed fee for service – the indicated amount of money for the service in reporting month;

  • Commission Rate (percent)- in case the price is set as the percentage commission on service provided- commission rate in reporting month;

  • Price of Travel– in case the price is set as the percentage commission on service provided – total value of indicated service in reporting month (e.g. price of airline ticket, value of hotel, value of tour), GEL;

  • Comment – other additional information.

Appendix III. Monthly Questionnaire for Tour Operator Services

Statistical Survey of Service Producer Prices for Tour Operator Activities Questionnaire N 06.2.9.4

(monthly)

article image

The table contains the following fields:

  • Tour Name – in all types of services titles of the top four tours with the highest share (indicated by the company in the beginning of the survey);

  • Duration of Tour – duration of the indicated tour, day-night;

  • Destination(s) – destination(s) that is/are included in tour;

  • Accommodation – accommodation type and class (number of stars);

  • Other Characteristics – other important characteristics (service of guide, transportation at the airport, etc.)

  • Price of Tour:

    • o Current month - the average price of tour per person (the total revenue generated from all tours provided in reporting month divided by the number of people who used the tour in reporting month) or the actual value per person staying in a room for 2 persons in reporting month (GEL);

  • Comment – indicate if price is average price for all people on the tour or transaction price for single person staying in room for 2 persons. In case of price change the reasons of the change or other additional information.

Appendix IV. Example of Substituting Items in Index Calculated with Item Weights and Long-Term Relatives

article image
Note that the aggregate in the prior month and aggregate N-l are the same, unless the items in the index change in the current month The substitution from Company 3 to the smaller Company 4 does not affect the index change in March