# Statistics

Home > Statistics > Outline of Statistics and Statistical Release Schedule > Explanations of Statistics > Explanation of Price Index > Outline of Corporate Services Price Index (CSPI, 2005 base)

# Outline of Corporate Services Price Index (CSPI, 2005 base)

October 2009
Bank of Japan
Research and Statistics Department

## 1. Purpose and Application

The Corporate Services Price Index (CSPI) is an index compiled monthly to measure the changes in price of a range of service products provided by businesses to other businesses and to local and central government in Japan. The CSPI began to be developed in 1985.

The main purpose of the index is to provide timely and detailed information about the price developments of individual services. Also, the index helps one to capture the overall supply and demand conditions of service industries and provides useful information for monetary policy making. The indexes of lowest level within an index structure (e.g., Item) not only have a public role as deflators that transform nominal output values into real quantities, but have been used as useful references for the pricing of individual transactions.

## 2. Coverage

The CSPI covers service products that are provided by businesses to other businesses and to local and central government in Japan. Service products provided to individuals are beyond the scope of the CSPI. However, some service products (i.e., postal services, telecommunications services, etc.) provided not only to individuals but also to businesses are covered.

Coverage of the CSPI has been gradually expanded. Currently, the 2005 CSPI is estimated to cover one half of all service transactions among businesses in Japan. However, there are remaining important service products not covered in the 2005 CSPI, for example financial intermediation, wholesale and retail trade, whose prices are difficult to survey reliably, and so efforts to increase the coverage are continuing.

## 3. Structure, Classification and Weights

The CSPI is composed of the basic grouping index and the reference index.

### Basic grouping index

The basic grouping index (BGI) covers domestic and imported service products traded among businesses. The BGI is subdivided into 7 Major groups, 20 Groups, 49 Subgroups, and 137 Items. Item corresponds to the elementary aggregates. Weights assigned to each Item are based on the transaction values of service products during the base year 2005 within the "Total of intermediate sectors," "Gross domestic fixed capital formation," and "Consumption expenditures outside households" of the "Input-Output Tables (I-O Tables)," published by the Ministry of Internal Affairs and Communications. The BGI includes consumption tax, and transactions contracted in foreign currencies are converted into yen basis. As for the references, the BOJ publishes "All items (excluding International transportation)" and "Indexes of contract currency basis" which are compiled and released in the original currency basis. See Component Ratio by Contract Currency in the CSPI for the relative weight of the Sample prices contracted in various currencies.

### Reference index

The reference indexes, which are not included in the BGI or in which adjustments are made to the BGI, are compiled and published to satisfy users' needs. These indexes include the "Corporate Services Price Index excluding Consumption Tax," "Lease rate," "Ocean freight transportation," and "International air freight transportation" in the Export CSPI. The structure, classification and weight of "Corporate Services Price Index excluding Consumption Tax" are identical to those of the BGI.

## 4. Criteria for the Selection of Subgroups and Items

Basic sectors in the "I-O Table" whose transaction values for business and governments exceed 500 billion yen are in principle selected as Subgroups. Then, Item in each Subgroup is selected, if sufficient information for calculating index weights is available and Sample prices can be collected continuously with the quality of products being held at constant.

## 5. Price Survey Procedure for Sample Prices

Surveyed prices and product information, which are referred to as Sample prices, are used to compile the Item indexes. At least three Sample prices from plural respondents are surveyed for each Item. In the middle of every month, respondents are requested to report prices of the previous month, specifically prices observed at the time of delivery to the customer. When the transactions are based on foreign currencies, reported prices are converted into their yen equivalents using the monthly average spot exchange rate.

Two important principles for the survey of Sample prices are i) to collect actual transaction prices reflecting the market conditions and ii) to control the quality of product, contract terms and transaction partners so that the changes in reported prices accurately reflect pure price changes. The standard price survey method satisfies these principles when there are many producers and their service products are fairly homogeneous. However, there exist many service products for which prices are discriminated. In addition, some custom-made products are not traded repeatedly. For these service products, alternative price survey methods are applied.

(a) Unit value method
Average prices, i.e., the total value of transactions divided by the total quantity of transactions for the group of some products, are surveyed, if the veraging does not significantly compromise the principles stated above.

(b) Model pricing method
The model pricing method is used to survey the prices of service products with price discrimination and custom-made service products which are not traded repeatedly. For products with price discrimination, buyers are assumed to be offered different price menus depending on their types, and the weighted average of the set of lowest prices by type of buyer is surveyed. For custom-made products which are not traded repeatedly, normalized service products in terms of quality, contract conditions and transaction partners are assumed and the prices of such products are surveyed.

(c) Pricing based on working time method
The prices of services per unit of labor input are surveyed in the case where the quality of the product is proportional to the quantity of labor input.

(d) Percentage fee method
Where surveyed prices take the form of commission rates such as financial services or leasing services, the gross growth rate of the prices of products, defined as the multiple of gross growth rate of commission rates and gross growth rate of some appropriate price indexes which are called inflators, are surveyed.

(e) List price method
List prices, such as standard prices or recommended prices with fixed services or contract conditions, are surveyed, if such prices can be regarded to reflect actual transaction prices in the market.

#### Treatment of missing prices

Missing prices may occur if no price report has been received or if no contract is made for the sample product within the month. In these cases, the prices of the previous month are simply imputed. When the prices are contracted in foreign currencies, the prices of the previous month in their original currencies are used and converted into a yen basis, which therefore reflect the exchange rate movement from the previous month.

When contract terms last for several months, and when the formal and final transaction prices are determined after the contract has been fulfilled, the index is tentatively composed using "pre-transaction prices" i.e., the interim prices used for the actual transactions, "expected prices" or latest final transaction prices.

The missing prices, except the case where no contract is made, are revised later on.

#### Adoption of prices from external statistics and databases

For some Items such as "Marine freight transportation," "Civil engineering and architectural services" and "International passenger transportation," the prices collected from external statistics and databases which are reliably accurate are adopted as Sample prices.

#### Replacement of Sample prices and quality adjustment method

The Sample prices should be replaced without delay in cases where: (1) they lose their representativeness in the market; (2) there is a change in contract terms; (3) it is necessary to replace the respondent. When replacing a Sample price, one of the five quality adjustment methods stated below is applied in order to ensure that the changes in reported price reflect pure price changes.

 Direct comparison The reported price differences between the old and the new products are reflecting pure price changes. Unit price comparison The quality of products is assumed to be proportional to the quantity of products, and prices are measured per unit of size, weight, or number. Overlapping The price differences between the old and the new products are assumed to represent pure quality differences when both services are available simultaneously during a certain period of time and when the relative prices of the two products are stable. Production cost The changes in production costs are assumed to reflect quality differences. In this method, the data which contains information on the difference in production cost of the old and the new products are asked to be reported to make appropriate quality adjustments. Hedonic regression The functional relationships between the characteristics of the products and prices of products are specified by estimating the Hedonic regression. The estimated relationship is used to model the hypothetical price for new products and the difference between the old price and the hypothetical price is regarded to represent pure price changes.

## 6. Index Formula

The index formula is the fixed-weight Laspeyres formula. The prices of each service are converted into index values at every period and a weighted arithmetic mean based on the fixed value-based weights for the base year is adopted for the index calculation:

The index for Item, which is the lowest level within an index structure, is compiled by the weighted arithmetic mean of the relative Sample price index, i.e., the ratio of the current period price to the base period price, using relative weight within the Item. Similarly, the index of upper classification levels (Subgroup, Group, Major group, All items) is compiled, like the Item index, using Laspeyres formula and appropriate relative weight.

## 7. Publication

In principle, the preliminary figures are released at 8:50 a.m. on the eighteenth working day of the month following the survey month. The release date may be brought forward a few working days, when the number of working days in the month is limited. The revised figures are released in line with the next month's preliminary figure indexes; revised data are marked with the symbol "r".

In some cases the number of Sample prices for a certain Item becomes less than three or the number of respondents becomes less than two. In such cases, unless approval is obtained before the release from the respondents, the index for that Item and for other corresponding Items in the same Subgroup will not be released and will be marked with the symbol "x" in order to protect the individual price information of these respondents.

Some figures are revised in order to incorporate late reports and corrections from respondents. The released indexes are periodically revised twice a year, in March and September, along with the release of the preliminary figures for the February and August indexes. When contract terms last for several months, and when prices are determined after the contract has been fulfilled, the final prices are not determined. The missing prices are replaced with the final prices upon making the periodic revisions.

Specifically, indexes for "Cellular phone services and PHS services," "Custom software," "Certified public accountant services" and "Architectural design services" are susceptible to such revisions since the improved price survey method for such indexes requires data that are only available at low frequencies. Periodic revisions cover, in principle, the most recent 18 months of data. In addition to the periodic revisions, occasional revisions are conducted when a change in the figures that has a non-negligible impact on the overall index is revealed after the release; the figures are then revised as soon as possible.

## 8. The 2005 Base Linked Indexes

The 2005 Base Linked Index offers a retroactively compiled time-series of indexes for all levels of aggregation in the BGI, going back as far as January 1985. Linked indexes are calculated for each index series on a monthly basis, using the following formula:

Note that in applying the link formula, the structure and classification of Items are rearranged to conform with the BGI used in the 2005 base index. Items selected and their weights, however, remain as in their original base indexes.