Statistics NZ Home > Browse for statistics > Economic indicators > Prices indexes > International PPP Comparison

International practices for business price index sample selection

1. Introduction
2. International practice

3. Sample selection

4. References

1. Introduction

Statistics New Zealand is progressively redeveloping a suite of business price indexes (BPIs) comprising the Producers Price Index (PPI), the Farm Expenses Price Index (FEPI) and the Capital Goods Price Index (CGPI).

As part of the redevelopment, New Zealand practice for various aspects of BPI compilation has been compared with the practices of five other Organisation for Economic Co-operation and Development (OECD) countries.

This article looks at how BPI samples are selected in New Zealand and in five other countries.

top

2. International practice

For the purposes of international comparison, Statistics NZ selected five OECD countries representing Europe, North America, Asia and Australasia: the United Kingdom, Australia, Canada, the United States and Japan. They were selected in large part based on the reputation of their economic statistics, and the availability of comprehensive documentation in English. The sources used (see the references section) are publicly available and may not reflect the most current developments in those countries.

It is worth noting that each of these nations has a much larger economy than New Zealand, making possible certain sampling procedures that may not be possible in a nation of four million people. It may be a worthwhile future endeavour to make a similar comparison with a second tier of smaller OECD nations, such as Ireland.

Statistics NZ tried to make the closest possible match with New Zealand’s BPIs in terms of data coverage. Each nation arranges their indexes slightly differently; for ease of comparison, the indexes under consideration are included below.

The UK Office for National Statistics PPI suite includes: home PPIs (domestic), export price indexes and import price indexes. Capital goods are included in their coverage.

The Australian Bureau of Statistics (ABS) produces a suite of PPIs, an Import Price Index, and an Export Price Index. The PPIs include two services: transport (freight) and storage division, and the property and business services division.

Statistics Canada produces an Industrial Product Price Index, a Raw Materials Price Index and four service PPIs (accounting services, informatics professional services, consulting engineering services and traveller accommodation services). Canada produces a Machinery and Equipment Price Index that parallels New Zealand’s CGPI, and a Farm Input Price Index that parallels New Zealand’s FEPI.

The U.S. Bureau of Labor Statistics (BLS) produces a PPI that includes capital goods. The PPI covers almost the entire output of domestic goods-producing sectors, and many service industries. Import and export prices are collected and indexed separately.

The Bank of Japan produces a Corporate Goods Price Index composed of the Domestic Corporate Goods Price Index, Export Price Index, and Import Price Index; capital goods are included in its coverage. They produce a separate Corporate Service Price Index.

The New Zealand suite of PPIs comprehensively covers both goods-producing and service industries.

top

3. Sample selection

Statistics New Zealand

The BPIs measure price changes, over time, for broad sectors of the New Zealand economy. Transactions for each sector may cover purchases and sales of thousands of different commodities at a wide variety of prices. The sheer volume and complexity of these transactions means that it is impossible to collect prices for every commodity or to take into account every price at which commodities are sold. Consequently, it is necessary to adopt a sampling approach – that is, to price a sample of specifications for a sample of commodities from a sample of respondents.

Purposive sampling

Statistics NZ’s BPIs are based on ‘purposive’ or judgement samples, where the sample is selected on the basis of the knowledge and judgement of staff compiling the index. The alternative of using probability (or statistical) sampling would be far more difficult and expensive to use. In particular:

Factors other than sales volume are important when selecting items and respondents. These include availability of prices on an ongoing basis, degree of price dispersion and the pricing behaviour of respondents.
Judgement sampling is more practical in the day-to-day operations of price collection, where specifications, commodities and respondents regularly need to be replaced.
The size of New Zealand’s economy limits the sampling options.

The effectiveness of this sample approach depends on the representativeness at each level of an index. A 'chain of representativeness' approach is adopted in which:

Key commodities are selected to adequately represent the price movements of all the commodities that come within the scope of the particular price index.
Respondents are selected to adequately represent all the suppliers/users of the selected commodities.
Specifications are selected from each respondent to adequately represent the whole product range within the selected commodity description.
Transaction prices that best represent the price movements of all transactions for each selected specification are obtained.

The use of judgement sampling has implications for the selection of replacement specifications or respondents. Generally, those chosen represent a range of specifications or respondents, not just themselves. For example, a respondent may represent medium-sized firms or a specification may represent a broader product grouping. Therefore the selection of replacements takes into account these characteristics and, as far as possible, ensures that they are still covered by the new specification or respondent.

Respondent selection

Business directories, market reports, economic census data and other information all help to form the basis for the selection and ongoing maintenance of samples of respondents and specifications for pricing. This information is essential in developing a comprehensive understanding of the market, and for making appropriate judgements in the sample selection process. Sampling of firms is on a cut-off basis, meaning that the largest firms in each industry are selected until 80 percent of the industry’s Annual Enterprise Survey unweighted sales volume is represented.

When a respondent ceases operating or no longer carries the commodity in question in between index developments, a new respondent is located using the same judgement considerations, and is contacted by telephone or mail.

Selecting specifications to price

The number and description of specifications priced in any area is set in the redevelopment of each index, and is dependent on the importance (that is, weight) of that broad category in relation to total expenditure or income. Typically, specifications priced are representative of a broad category or group of commodities where small weights have been added together.

Price indexes rely on the samples of prices on which they are based. Consequently, in selecting the samples for index commodities with large weights in a published series, the aim is to cover those businesses which account for a high proportion of sales or purchases of the products making up the commodities. For less significant commodities, coverage may be relatively low, but care is taken to ensure that the selected businesses are representative of all the businesses trading in the commodities.

Price indexes aim to reflect changes in market transaction prices. Thus, the specifications identify the precise terms under which transactions are typically made. Prices collected take account of discounts, surcharges, conditions of sale, order sizes, type of customer, etc. List prices are only used where most transactions for a specification occur at the list price or no alternative is available.

When selecting specifications to be priced the following factors are considered:

The relative importance of the commodity in an index; more important commodities are given greater attention.
The degree of homogeneity within the index commodity; more homogeneous commodities require smaller samples.
Influences which can cause prices of some specifications to move differently from others within the commodity; it is important to reflect all the significant influences in the sample.
The extent to which the specification can be expected to be continuously available for pricing; preference is given to specifications for which transaction prices are available on a continuous basis.
Whether the specification can be clearly described in terms of quantity and quality, so those prices can be collected over time to constant quality.

Office for National Statistics (UK)

A new sampling scheme for the home PPIs was introduced in 2001, in which the sample is chosen as contributor-product pairs from those companies selected for Products of the European Community (Prodcom) enquiry. To ensure that the PPI sample gives the best results possible within its limit of 9,000 prices a month, an optimal allocation has been statistically calculated. This increased the number of contributors from 3,700 to 7,000 so that each contributor was providing fewer prices; 70 percent of the sample provide just one price quote.

A new sample is selected annually from Prodcom, which allows for rotation of the smaller contributors. Sample rotation will pick up new products and contributors, pick up existing companies making a product for the first time, update company structures annually, fill in gaps in the sample caused by contributors dropping out, and spread the burden of completing forms between smaller companies.

Respondents for the import and export indexes are selected from the largest firms.

Australian Bureau of Statistics

ABS price indexes are based on judgement samples, where the sample is selected on the basis of the knowledge and judgement of staff compiling the index. Price reporting by businesses is initiated by a personal visit by an ABS officer. The selection of a particular specification for pricing is made in consultation with the respondent. The aim is to identify particular products that will best reflect the price movements of all the products sold or purchased by the business which are within the scope of the relevant index item.

Commodities directly represented in the import and export indexes were selected on the basis of the significance of their import and export values in the period 2005-06 for imports and 2004-05 and 2005-06 for exports. All significant commodities were selected for pricing. The weights for minor commodities that are not directly priced are included with those of comparable, directly priced commodities whose prices are likely to move in a similar way.

Statistics Canada

The PPIs are sampled by formal probability sampling from manufacturing establishments in Canada, as identified through the Annual Survey of Manufactures. Important producers are 'must take' respondents. A random selection is taken from the smaller producers. These samples are redrawn on a regular cycle as dictated by changes in the market. A directly surveyed elemental index is usually supported by 12 to 15 price quotes. An index may be derived from as few as eight quotes or as many as 30 quotes.

Commodity specialists discuss products and terms of sale with each respondent, and use that information to draw up a detailed price specification. In most cases, the specific quotation chosen pertains to the sale of a major product line to a major type of customer. However, if the same manufacturer is being asked for more than one quotation on the same product line, a procedure is followed which gives some chance of secondary lines or secondary markets being included.

The Raw Materials Price Index is sampled judgementally by selecting a range of specific price quotations to represent purchases of the commodity as a whole.

The Machinery and Equipment Price Indexes are sampled from the Annual Survey of Manufactures.

The Farm Inputs Price Index sample is selected judgementally from the distributors of all inputs (goods and services) going into the farming sector. This includes distributors of farm machinery, petroleum products, crop inputs (for example, seeds and fertilisers), veterinary services, etc. In the absence of a formal sampling frame of farm suppliers, various sources of information are used (such as telephone directories and association lists) to identify and select respondents.

Service PPIs are sampled from all establishments primarily engaged in the designated services with annual revenue exceeding a specified threshold, as identified on Statistics Canada's Central Frame Database. Respondents for most of these surveys are selected through a probability sample survey proportional to the significance of their operating income and based on the region in which they are located. The sample size for these surveys ranges from 400 to 500. In contrast, the 1,000-unit informatics services sample is a cut-off sample.

In each case, the entire survey sample is usually reviewed in total once every five years. In the interim, re-sampling occurs primarily for commodity groups subject to relatively rapid development in firm structure or product development.

Bureau of Labor Statistics (USA)

For any given industry, producers are selected for the survey via a systematic sampling from a stratified listing of all firms that file with the Unemployment Insurance System. Typically, a firm's probability of selection is based on its employment size. If there is evidence that some strata have more heterogeneity in price change, they are oversampled. After a firm is selected and agrees to participate in the survey, a probability sampling technique called disaggregation is used to determine which specific products or services will be in the PPI.

Disaggregation is a process in which iterative steps are taken to select items based on their proportionate value to the manufacturer's overall revenue. First a reporter breaks down into categories the type of items shipped. Next, these categories are broken down further by price-determining characteristics, for example, options, colour, size. Further break downs may be necessary to differentiate between types of buyers or discounts. Disaggregation continues until a specific product sold to a specific buyer is selected. To minimise respondent burden, the entire process is usually completed within a two-hour initiation interview.

The collected data are reviewed first by the BLS regional office staff to ensure consistency and completeness. They are subject to final review by BLS national office, where a survey is prepared for that respondent, listing all price-determining variables and terms of sale for each selected product. From this point forward, the establishment reports prices for the selected products, usually on a monthly basis, on a form provided by BLS. Establishments continue to report until a new sample is selected – after seven years, on average, for an industry.

Bank of Japan

Representative products are selected using judgement for each commodity category: 910 items for the Domestic Corporate Goods Price Index, 222 items for the Export Price Index, and 293 for the Import Price Index. Producers of those products are then purposively selected to supply quotes. Weights are determined using the Census of Manufacturers, customs statistics, and statistics compiled by official organisations and cooperating associations. Weights for the Corporate Service Price Index are determined using input-output tables published by the Ministry of Internal Affairs and Communications.

Summary

Of the five selected nations, three are using probability sampling methods and two are using judgement sampling. All have some sort of regular sample rotation to keep their sample up to date, ranging from one to seven years.

top