13045_om_not-searchable
/en/priser-og-prisindekser/statistikker/repi/kvartal
13045_om
statistikk
2008-01-17T10:00:00.000Z
Prices and price indices;Establishments, enterprises and accounts
en
false

Industrial cleaning services, price indexQ3 2007

Content

About the statistics

Definitions

Name and topic

Name: Industrial cleaning services, price index
Topic: Prices and price indices

Responsible division

Division for Construction and Service

Definitions of the main concepts and variables

The price measured in this index, is invoiced price on cleaning contracts, exclusive VAT. It is fixed long-term contracts of traditional cleaning services. The contracts must be representative concerning the turnover for each establishment. We also ask the establishments' to divide a price change, if one have occurred, between changes in the contract and a pure price change.

The turnover used to weigh the index is collected either in the structural business statistics, VAT register or the questionnaire. VAT is not included in the statistics.

Standard classifications

The Standard Industrial Classification used in Statistics Norway (SN2002), is based on the EU standard NACE Rev. 1.1. You can find this standard here: NACE-standards . Industrial cleaning services have code 74.7.

Classification of products by activity is closely connected to NACE Rev 1.1. The standard treat typical services or products within each industry group. The index treat the product group: Traditional cleaning services (74.70.13) .

Administrative information

Regional level

National

Frequency and timeliness

Frequency: Quarterly

Timelines: The statistics are published within 12 weeks after the end of the quarter

International reporting

The statistics will be reported to EUROSTAT

Microdata

Data are stored as text files on UNIX.

Background

Background and purpose

The purpose of the index is to measure the development in prices for industrial cleaning services. Further the index will be used to deflate the National Accounts. The statistics comply the short-term statistics regulation.

The statistics were published for the first time in 2007, with figures from the 4th quarter of 2004.

Users and applications

The index is used to analyse and survey the development in prices and costs in the industry, and in the National Accounts. Others with an interest in the market for industrial cleaning services also use it, for example institutions within research and development and media.

Coherence with other statistics

The index is used in National Accounts. A value index is also produced for this industry group. The structural business statistics you can find at: Structural business statistics . There is a difference in detailing level between preliminary and final figures.

Legal authority

The Statistics Act, Section 2-1,2-2 and 2-3

EEA reference

Regulation (EC) No 1158/2005 of the European Parliament and the Council of July 6 2005 amending Council Regulation (EC) No 1165/98 concerning short-term statistics.

Production

Population

The population is all establishments in Norway classified in the industry group Industrial cleaning services (SN2002:74.7). The statistical unit in the survey is establishments (local kind-of-activity units). The industry group consist of different cleaning services. The statistics are limited to cover Traditional cleaning services.

Data sources and sampling

Statistics Norway's Business Register is used to draw the sample. Price data are collected via a separate questionnaire-based survey. Figures on the establishments' turnover are collected from the structural business statistics, the Vat register and from questionnaires. The reason we use different sources is to reduce the respondent burden.

From the population we draw a sample containing about 90 establishments. We divide the establishments into size groups, determined by the number of employees. We use a PPS-sample (Probability Proportional to Size), i.e. the sample will be dominated by large establishments. Finally, the establishments themselves chooses contracts that are representative for their activity.

Collection of data, editing and estimations

The survey is based on a quarterly questionnaire combined with information from the structural business statistics and Vat register. We offer both a postal and electronic questionnaire. The questionnaire is posted by the end of the quarter, with a three weeks deadline. Respondents that have not returned the questionnaire by the deadline, are given a postal reminder together with a one-week extension of the deadline. If the questionnaire still is not returned, the respondent is given a fine and a one-week final extension of the deadline.

The postal questionnaires are read optically, and the electronic questionnaires are automatically loaded into the revision application. Every questionnaire are afterwards controlled and revised, both automatically and manually. If price changes seems unlikely large, the respondents are contacted.

For each establishment in the sample, price change per contract is calculated. The next step is to calculate sub-indices for every size group, and finally a main index for the industry group. The weights used in the calculation, is turnover in the industry group. We use a combination of turnover by establishment and turnover by establishment and type of cleaning service. The establishments are stratified into five size groups by the number of employees. The index calculated is a fixed-weights index, more precisely a Young Index with fixed weights. The weights will updated annually in connection with renewal of the sample. We find the price development by comparing the prices in the present quarter to the prices in the previous quarter. This enables us to include new contracts continuously without estimating and imputing fictive base period prices. We chain the index every quarter to create the published index.

Confidentiality

It is impossible to identify sensitive information from the statistics.

The use of collected data from respondents will be in accordance with the demands in the Statistics Act provisions. The information will be stored properly and kept confidential.

Comparability over time and space

The time series goes back to the 4th quarter of 2004. There has been no changes in the calculation method.

Accuracy and reliability

Sources of error and uncertainty

Errors might be caused by not correctly answered questionnaires. There are two main sources to measurement errors. First, the respondents states the same price even though there has been an actual price change from previous quarter. Second, the respondents does not report the reason for a price change, report the wrong reason for the price change or they distribute the price change wrong. If a measurement error is suspected, the respondent is contacted. Data are also controlled when the questionnaires are optically read. In general, the production system ensures that the data is controlled on both detailed and aggregated level.

Non-response will always be a factor in questionnaire-based surveys. Some respondents do not return the questionnaires, or the questionnaires are rejected because they are incompletely filled out. About 95 percent of the questionnaires are returned to Statistics Norway.

The sample of 90 establishments constitutes approximately 4 percent of the industry group's population, and approximately 50 percent of the turnover. Sampling errors are not calculated since this measurement of uncertainty is not relevant for the index. A price index must be seen upon as an estimate on the general price development, within the area of its scope. This is also valid if we collect prices for every contract from every establishment in the population. Hence, we will always have uncertainty since the index is an estimate. Furthermore, we can not calculate the correct sampling error when we use two-stepped sampling with representative contracts in the last step.

In calculation of the price index we have chosen a stochastic approach. In light of this we calculate a model error. Meaning, we measure the variation in price relatives that are used in the calculation of averages leading to elementary indices. The model error is mainly influenced by variation in the prices, and the number of price observations in an elementary index. To calculate a model error for a elementary index, we need at least two price relatives. We calculate the model error on every level of the aggregation of the index.

Given the model we have chosen for this index and the underlying assumptions, we have calculated a confidence interval. The interval gives a estimate for what we believe the index number will lie within, with a 95 per cent probability. We calculate the confidence interval by withdraw or add two times the standard deviation to the index number. This gives us the lower and upper limit shown in the table below.

Period

Index

Standard deviation

Lower limit

Upper limit

1st quarter 2006

100.9

0.35

100.2

101.7

2nd quarter 2006

104.3

0.27

103.7

104.8

3rd quarter 2006

105.2

0.27

104.6

105.7

4th quarter 2006

106.2

0.31

105.6

106.9