Source data collection programs
- Data source: The index is based on a survey which covers enterprises with 10 or more employees in mining and manufacturing.
- Nature of indicators used: 75 percent of the basic series consist of deflated delivery data, 20 percent measure output in physical quantities and 5 percent consist of hours worked. Delivery volumes are based on delivery values, deflated by the producer prices index.
- Nature of weights: Value added in the previous year.
- Sources of weights: Based on an annual survey of industrial production in manufacturing and the National Accounts.
- Period of current index weights: Previous year.
- Frequency of weight updates: Every year.
- Item selection: Not applicable.
- Unit selection: Data on delivery is collected by a monthly survey. Businesses with at least 10 employees are selected from the Business Register of Statistics Sweden according to their main activity. The survey includes all businesses with employment greater than 500. Below this threshold, stratified random sampling is used.
- Sample sizes: Around 2200 kind-of-activity units in the delivery-survey.
- Data collection methods: Data on delivery is collected by a monthly mail questionnaire. Information on physical output is collected from businesses and trade associations, and hours worked is collected by a monthly mail questionnaire to about 8000 businesses with at least 5 employees.
Source data statistical techniques
- Computation of lowest level indices: The lowest level indices are obtained at NACE Rev 1 2-, 3- and 4-digit level depending on the information available.
- Aggregation: Laspeyres chain index.
- Linking of re-weighted index to historical index: New weights are used to re-calculate the series as far back as is necessary, usually about five years.
- Reference period: 2015 = 100.
- Procedures for non-response: Rates of response over 90 percent weighted. Imputations for non-respondents are made by various means based on previous replies and/or on the replies of other businesses.
- Missing data are imputed by various means.
Other statistical procedures
- Introducing new units: New units are not introduced during the sample year. A new stratified random sample is drawn annually in March.
- Verification of data: Accuracy and reliability of data is verified on business level, as well as on the aggregated level.
- Verification of processing: No special procedure exists at present.
- Seasonally adjusted indices: Seasonally adjusted data (using TRAMO/SEATS) are published at the 2-, 3- or 4-digit level depending on the business category.
- Consistency with other indicators of industrial production: The production index is compared with indices of turnover and new orders on a monthly basis.