Producer Price Index for Industrial Products and Import Price Index — Technical Manual

📚

References

International methodological handbooks

The presented technical manual is elaborated by the National Statistics Office of Georgia according to the internationally acclaimed methods and practice and is based on the following handbooks:

“Producer Price Index Manual: Theory and Practice”, International Monetary Fund, 2004.
Responsible organizations: International Labor Organization (ILO), International Monetary Fund (IMF), Organization for Economic Co-operation and Development (OECD), United Nations Economic Commission for Europe (UNECE), and World Bank.
https://www.imf.org/…
“Handbook on industrial producer price indices (PPI)”, Eurostat, 2012.
Responsible organizations: European Statistical Office (Eurostat) and European Commission.
https://ec.europa.eu/eurostat/…
“Export and Import Price Index Manual: Theory and Practice”, International Monetary Fund, 2009.
Responsible organizations: International Labor Organization (ILO), International Monetary Fund (IMF), Organization for Economic Co-operation and Development (OECD), United Nations Economic Commission for Europe (UNECE), and World Bank.
https://www.imf.org/external/…

📋

1. Introduction

1.1 Producer Price Index for Industrial Products and Import Price Index and their use

Producer Price Index (PPI) for Industrial Products measures the average price level of the goods produced by producers compared to the reference period. Import Price Index (MPI) measures the average price level of products produced abroad and imported to the country, compared to the reference period.

The structure of PPI follows the Statistical Classification of Products by Activity (CPA 2008) and covers the following sections: mining and quarrying (B); manufactured products (C); electricity, gas, steam and air conditioning (D); water supply, sewerage, waste management and remediation services (E).

Producer Price Index for Industrial Products and Import Price Index are used for the following purposes:

The indices have an important role in deflating different economic indicators;
The indices are used for indexation of contracts in both public and private sectors;
The indices are analytical instruments for researchers and representatives of the business sector.

1.2 Structure of Producer Price Index

Producer Price Index comprises two sub-indices: Domestic Producer Price Index (DPPI) for Industrial Products and Export Price Index (XPI). The former measures the average price level of industrial goods produced in the country and sold on the domestic market compared to the reference period. Export Price Index measures the average price level of industrial goods produced for export purposes compared to the reference period.

Import Price Index measures the average price level of products produced abroad and imported to the country. Domestic Supply Producer Price Index (DSPPI) is obtained by combining MPI and DPPI for industrial products. It measures the average price level of industrial goods produced in the country and sold on the domestic market, as well as the price level of products produced abroad and imported to the country, compared to the reference period.

1.3 Coverage of PPI and MPI and the observable prices

The prices used for calculating Domestic Producer Price Index are those at the factory gate and do not include VAT, excise, and transport expenses. The prices are collected for the output of domestic enterprises across the country.

Regarding Export Price Index, the observable price is the free-on-board (F.O.B.) price set by producers in the reporting period. The F.O.B. price comprises the price at the factory gate, net taxes on products (taxes minus subsidies), transportation, and other expenses incurred in bringing the product to the point of departure from the economic territory of the producing country.

In the case of Import Price Index, the observable price is the cost, insurance and freight (C.I.F.) price of the imported product. The imported product price includes: the transaction value of the product and the cost of delivering it to the border of the importing country. The use of C.I.F. prices is recommended by the United Nations.

2

Sampling of products and organizations

DPPI For price registration of industrial products produced for the domestic market, products are selected according to their shares in the volume of total domestic industrial production. Product sampling is conducted according to CPA 2008. The statistical data of enterprises by kind of industrial products in value terms are used for the sampling.

XPI For exported products, sampling is performed according to the shares of products in the total export value (re-exports excluded). External trade statistics data are used for this purpose.

MPI The selection of observable imported goods is also based on external trade statistics data, presented by the class of CPA.

A survey of producer/exporter/importer organizations is conducted based on the sampled products. On the following stage, sampled enterprises are surveyed in order to define product specifications.

During the product selection process, detailed specifications are determined. Following the specifications is the most important part of price registration, since the monthly recorded price difference should be caused by the pure price change of a product, rather than changes in its characteristics. Relying on the obtained survey data, the prices for sampled products are recorded throughout the year. Product selection is updated annually.

3

Price collection fieldworks

Prices for domestic and imported industrial production are collected by price enumerators. Price collection fieldworks are conducted from the 1st to the 8th of the month following the reporting period, through electronic questionnaires on the Geostat website. In these questionnaires enterprises indicate the following information about four selected products: measurement unit, prices in the reference, previous and current months; for XPI — the country of destination; and for MPI — the country of origin.

The information about product prices provided by enterprises is confidential and is protected by the “General Administrative Code of Georgia” and article 28 of the “Law of Georgia on Official Statistics”.

Unless otherwise provided for by the legislation of Georgia, persons registered in the Register of Entrepreneurial and Non-entrepreneurial (Non-commercial) Legal Entities are obliged, upon a written or electronic request from the National Statistics Office of Georgia, to present information available to them in material or electronic form, including confidential information.

4

Validation procedures

PPI and MPI validation procedures are conducted in two stages:

At the first stage, validation takes place simultaneously with the price registration fieldworks. In the case of a price change, the person responsible for filling in the questionnaire is required to define by a comment the reason for the change. After the data is sent to the central office, a responsible employee conducts analysis and logical control of the data.

At the second stage, the accuracy of prices that deviate significantly from the previous month is checked after the indices have been calculated.

5

Weights

Weights for individual products in PPI and MPI are updated annually based on the production structure defined by the National Accounts System and external trade statistics data, reflecting the latest information on industrial production output and imported products across the country. The obtained weights represent the share of the product’s value in the overall value of products produced or imported in the country. Weights for the reporting period t are calculated based on t‑2 period information. The list of industrial products included in the index may also be changed when weights are updated.

6

Price imputation techniques

If the price for a product is not indicated by an enterprise in the reporting period, the price imputation method is applied.

For example, if there is no price recorded in April for brand A, the imputed index for this product will be the index of the group that includes this product. The group index is calculated using the actual price indices of products in the group.

Table 1.

Product	Weight, %	Base price	March price	April price	Ratio (Mar./Dec.)	Ratio (Apr./Dec.)
Brand A	0.051	4.55	4.50	—	4.50/4.55≈0.99	1.15*
Brand B	0.032	5.20	5.20	5.50	5.20/5.20=1.00	5.50/5.20≈1.06
Brand C	0.067	5.00	4.50	5.50	4.50/5.00=0.90	5.50/5.00=1.10

* Imputed index

In April, the imputed index for brand A is calculated in the following steps:

Imputed index for brand A (April)

Group LT index April = 1.06 × [0.032/(0.032+0.067)] + 1.10 × [0.067/(0.032+0.067)] ≈ 0.34 + 0.74 = 1.08

Group LT index March = 1.00 × [0.032/(0.032+0.067)] + 0.90 × [0.067/(0.032+0.067)] ≈ 0.32 + 0.61 = 0.93

Group ST index April = 1.08 / 0.93 ≈ 1.16

Imputed LT index for Brand A = 1.16 × 0.99 ≈ 1.15

If in the reporting month no price is recorded for products in a group, the imputed index will be calculated using the upper level group’s index, according to the CPA 2008 structure. If in the reporting month no price is recorded up to the third level of the CPA structure (e.g. 10.1 “preserved meat and meat products”), the imputed index will be calculated using the price repeating (carry-forward) method, rather than the upper level group index.

7

Quality adjustment

It is possible that an enterprise may no longer produce or import a product of the same quality for which prices have been observed. In order to ensure the comparability of prices for old and new products, a quality adjustment method should be used, for which a conditional base price is calculated using the following methods:

Method 1 — Qualitative difference is known

If in the reporting month a replacement product is qualitatively different from the product in the previous month, and the value of the difference is evaluated, the base price for the replacement product is calculated using the previous month’s price and the qualitative difference defined by the person responsible for filling the questionnaire.

Table 2.

Product	Base price	March price	April price	Qual. diff.	Ratio (Apr./Dec.)
Brand A	4.55	4.50	—	—	—
Qualitatively different — Brand B	5.86*	—	8.50	1.30	8.50/5.86≈1.45

* Imputed base price

Imputed base price for Brand B

Base price = (4.50 + 1.30) / (4.50 / 4.55) ≈ 5.86

Method 2 — Previous month’s price is available

If in the reporting month it is possible to determine the previous month’s price for the replacement product, the base price for the replacement product is calculated using this price and the index of the previous month.

Table 3.

Product	Base price	March price	April price	Ratio (Mar./Dec.)	Ratio (Apr./Dec.)
Brand A	4.55	4.50	—	4.50/4.55≈0.99	—
Replacement — Brand B	5.26*	5.20	5.50	—	5.50/5.26≈1.05

* Imputed base price

Imputed base price for Brand B

Base price = 5.20 / (4.50 / 4.55) ≈ 5.26

Method 3 — Group index method

If in the reporting period the price enumerator discovers that brand A will no longer be sold starting from the reporting month, and it is impossible to obtain the previous month’s price and the value of the qualitative difference for the replacement brand B, the difference between the current month’s price of brand B and the previous month’s price of brand A will be treated entirely as a qualitative difference. The imputed base price of the replacement product is calculated based on the current month’s index and the price of brand B.

Table 4.

Product	Weight, %	Base price	March	April	Ratio (Mar./Dec.)	Ratio (Apr./Dec.)
Brand A	0.051	4.55	4.50	—	4.50/4.55≈0.99	—
Replacement — Brand B	0.051	6.09**	—	7.00	—	7.00/6.09≈1.15*
Brand C	0.032	5.20	5.20	5.50	5.20/5.20=1.00	5.50/5.20≈1.06
Brand D	0.067	5.00	4.50	5.50	4.50/5.00=0.90	5.50/5.00=1.10

* Imputed long-term index ** Imputed base price

Imputed base price for Brand B (group index method)

Group LT index April = 1.06 × [0.032/(0.032+0.067)] + 1.10 × [0.067/(0.032+0.067)] ≈ 1.08

Group LT index March = 1.00 × [0.032/(0.032+0.067)] + 0.90 × [0.067/(0.032+0.067)] ≈ 0.93

Group ST index April = 1.08 / 0.93 ≈ 1.16

Imputed LT index for Brand A = 1.16 × 0.99 ≈ 1.15

Imputed base price for Brand B = 7.00 / 1.15 ≈ 6.09

🧮

8. Calculation of PPI and MPI on different levels

8.1 Calculation of the lowest level PPI and MPI

An index calculated for each product produced or imported by an enterprise is the lowest level index for PPI and MPI. Graph #1 shows the structure of the industrial sector, where the price indices for Product A, Product B and other individual products are elementary indices. The lowest level index, compared to the price reference period, is obtained from the ratio of reporting (t) and reference period product prices:

Lowest level index

Iᵢ^t/0 = pᵢ_t / pᵢ₀

i — product produced by an enterprise for which a comparable price is registered

Iᵢ^t/0 — lowest level index for product i in the reporting period t, compared to the index reference period

pᵢ_t — price of product i in period t

pᵢ₀ — price of product i in the price reference period

Graph #1. Structure of the industrial sector

8.2 PPI and MPI for separate groups and the whole industry

The long-term PPI and MPI for the whole industry compared to the price reference period is calculated using the following Laspeyres-type formula:

Laspeyres-type aggregate index

I^t/0 = Σⁿ_i=1 Iᵢ^t/0 × sᵢ^b

Iᵢ^t/0 — lowest level long-term index for product i compared to the price reference period

sᵢ^b = pᵢ^bqᵢ^b / Σpᵢ^bqᵢ^b — weight of product i in the weight reference period, representing the share of produced/imported product i in the whole production/import, where Σsᵢ^b=1

pᵢ_t — price of product i produced/imported by the sampled enterprise in the weight reference period (b)

pᵢ₀ — quantity of product i produced/imported in the weight reference period (b)

The same formula is used for calculating all upper level indices. For example, a section index is calculated by weighting the long-term indices of the products belonging to the section, where the sum of the weights of indices in the section equals 100%.

A short-term index compared to the previous month is obtained from the ratio of long-term indices in the reporting and previous months, calculated compared to the price reference period.

8.3 Chain index

During the annual update of samples of industrial products and enterprises, or the specifications of products in December (the update period), prices are collected for products in both the old and new samples. This enables chain-linking of indices calculated for two different samples. Chaining enables calculating indices with a long-term reference period despite changes in weights.

For example, before December 2016, compared to December 2015, the overall index was calculated using wᵢ weights, whereas the 2017 index is calculated compared to December 2016, using kᵢ weights (see Table 5).

Table 5.

12.2015=100	12.2016=100
12.2016: I^12.16/12.15 = ΣIᵢ^12.16/12.15×wᵢ = 106	12.2016: I^12.16/12.16 = ΣIᵢ^12.16/12.16×kᵢ = 100
X₁	01.2017: I^01.17/12.16 = ΣIᵢ^01.17/12.16×kᵢ = 102

Chain-linking formula

106 / X₁ = 100 / 102 ⇒ X₁ = (106 × 102) / 100 ≈ 108

I^{12.2016/12.2015} × I^{01.2017/12.2016} = 106 × 102/100 ≈ 108

📢

9. Data dissemination

9.1 Press release

Press releases for the following indices are published through the Geostat website (www.geostat.ge) on a monthly basis: DPPI, XPI, PPI for Industrial Products, MPI and DSPPI. They contain information about monthly and annual index rates, as well as contributions of groups to the index formation. Press releases also include time series graphs in various breakdowns.

Along with the press release, different time series are published on the Geostat website every month:

The indices compared to the previous month;
The indices compared to the long-term base (Average of 2010=100);
The indices compared to the same month of the previous year;
The 12-month average over the previous 12-month average.

The published indices are rounded to four digits and are final.

Time series data are published on the website along with the corresponding graphs.

The data are also available via PC-Axis — a data dissemination software created by Statistics Sweden. It is a comprehensive data dissemination system that gives users the opportunity to obtain different types of statistical information in various formats (text, tables, graphs, etc.) from the Geostat website (www.geostat.ge).

9.1.1 Contributions of product groups to the overall index percentage change

Calculating the contributions of certain product groups to changes in the overall index provides a powerful analytical tool for analysing PPI and MPI. The contribution of a group to the change in the overall index is defined as the percentage change of the overall index caused by the change of the given group index only, with all other group indices held constant.

The contribution of a product produced by an enterprise to the change of the overall index is calculated using the following formula:

Monthly contribution formula

Contribution_i (monthly) = (Iᵢ_t / Iᵢ_t‑1 − 1) × 100 × (Iᵢ_t‑1 / I^a_t‑1) × wᵢ_t

Iᵢ_t‑1 — index for product i in period t

I^a_t‑1 — index for product i in period t‑1

wᵢ_t — PPI/MPI for the whole industry in period t‑1

wᵢ_t — weight of product i in period t

The contribution of a group to the monthly index is the sum of contributions of all products within that group.

In the case of a weight change, the contribution of group i to the annual index is calculated using the following formula:

Annual contribution formula (with weight change)

Contribution_i (annual) = [(Iᵢ_L − Iᵢ_t‑12) / I^a_t‑12] × wᵢ_t‑12 × 100 + [(Iᵢ_t − 100) / I^a_t‑12] × I^a_L × wᵢ_t

Iᵢ_L — index for group i in the weight change period

Iᵢ_t‑12 — index of group i in period t‑12 (previous reference period=100)

I^a_t‑12 — PPI/MPI for the whole industry in period t‑12

wᵢ_t‑12 — weight of group i in the production volume of period t‑12

Iᵢ_t — index for group i in period t

I^a_L — PPI/MPI for the whole industry in the weight change period

wᵢ_t — weight of group i in the production volume of period t

📐

Example: The contribution of price change for the section of products from mining and quarrying to the annual PPI of October 2018, considering the weights of 2017 and 2018 (0.35 and 0.28, respectively) — see Table 6.

Table 6. Indices compared to December of the previous year

	Dec. 2016	Oct. 2017	Dec. 2017	Oct. 2018
Mining and quarrying	100.0	101.2	101.7	102.2
Overall index	100.0	101.6	103.2	101.8

Worked example

[(101.7 − 101.2) / 101.6] × 0.35 × 100 + [(102.2 − 100) / 101.6] × 0.28 × 103.2 = 0.8%

Thus, the contribution of the section of products from mining and quarrying to the annual index in October 2018 amounted to 0.8 percentage points.

Graph #2 represents the stages of PPI and MPI calculation and their periodicity.

Graph #2. Stages of PPI and MPI calculation — Annual activities

Graph #2. Stages of PPI and MPI calculation — Monthly activities