European Conference on Quality in Survey Statistics

1
European Conference on Quality in Survey
Statistics
2
Developing and implementing a new SBS at
Statistics Sweden

• Marcus Johansson
• marcus.johansson_at_scb.se
• Stefan Berg
• stefan.berg_at_scb.se

3
Structural Business Statistics

• The objective of the Structural Business
  Statistics (SBS) is to ascertain the structure,
  activity, competitiveness and performance of
  enterprises.
• The most important variables are
• Production value
• Value added
• Intermediate consumption
• Turnover
• Total assets
• Number of employees

4
Background

• 1997 use of administrative register starts
• Enterprises gt 50 employees are surveyed through
  questionnaires
• Enterprises lt 50 employees are surveyed
  through registers
• 2001 a revision of the 1997 method starts

5
The purpose of the project

• Reduce the respondent burden
• Improve contents and reliability
• Make the data collection more effective
• Better focus on the largest enterprises
• Electronic tool for data collection
• Extend the content for the service sector

6
From tax data (SRU) to SBS data

• Data transfer from the Swedish Tax Agency twice a
  year (August and December/January)
• The files consists of corporations and private
  firms
• Data is transferred to SBS population
  (non-financial enterprises)

7
Contents of the SRU

• About 150 variables, whereof 50 are calculated
• Income statement
• Incomes (turnover, other operative incomes,
  income from interest etc)
• Costs (personnel costs, depreciations, cost of
  interest etc)
• Profit/loss (operating profit, profit after
  financial items, net profit of the year)
• Balance sheet
• Assets (tangible assets, financial assets,
  inventories etc)
• Equity and liabilities (share capital, untaxed
  reserves, long-term liabilities etc)
• Investments (investments in land, buildings,
  shares etc)

8
Treatment of the SRU

• Automatic corrections (enterprises that do not
  sum up after automatic corrections are not
  approved and will be imputed)
• Manual corrections
• - Investments in tangible assets
• - Investments in shares
• - Changes in equity
• - Salaries (per employee)
• - Large changes in stocks
• - Large changes in pension provisions
• - Large increases in turnover (more than
  900-1100 times)
• - Strange relations between Commodities and Raw
  materials

9
Swedish SBS
10
Two data collection periods

• Spring collection (Complete form)
• Income statements
• Balance sheet
• Investments
• Share and participations
• Forms based on annual accounts
• Autumn collection (Income and cost form,
  Investment form, Shares and participation form)
• Pre-printed main posts from the SRU
• Forms based on tax data
• Electronic web form
• Variables linked to a general chart of accounts
  (BAS)

11
Survey designs - an overview

12
Pratical issues in realtions with ?ps-sampling

• Chose an appropriate fixed size ?ps-sampling
  scheme
• What size measure should we use?
• How to handle small inclusion probabilities?

13
The SpecRR sample for 2004
14
Future improvments

• Improvments and simplifications of the
  questionnaire
• Improve the coordination with other surveys
• Regional breakdowns