Further data linking with WERS 2004

Document Sample
Further data linking with WERS 2004 Powered By Docstoc
					   Further data linking with
        WERS 2004

             John Forth
WERS 2004 Information & Advice Service



                               National Institute
                               of Economic and
                               Social Research
Rationale for data linking
• Rationale 1:
  – Extend the information set for each workplace or
    employee
• Examples:
  – Augment workplace records with data on financial
    performance
  – Augment workplace records with data about wider
    enterprise
  – Augment workplace or employee records with contextual
    data about the industry sector or occupation
Rationale for data linking
• Rationale 2:
  – Reduce measurement error in a particular item
• Examples:
  – Compare WERS data with external source of information
    on legal status or foreign ownership
  – Compare WERS financial performance data with other
    sources
Rationale for data linking
• Rationale 3:
  – Extend the information set for another dataset
• Example:
  – Add workplace data from WERS to a separate workplace
    or employee-level dataset
Issues in data linking
• Confidentiality
• Extent and reliability of links
• Measurement error
• Loss of information


See also: Chesher and Nesheim (2006) DTI
  Occasional Paper No. 3.
Existing instance of linked data
ONS Annual Respondent Database (ARD2) matched
onto WERS 2004
– Data on financial performance of business units (from ABI)
– 847 (48%) of XS workplaces in trading sector
     (ASTATUS1<=8) have an ABI 2004 return in ARD2
– Offers data on value-added, profitability, labour costs, capex
– Low survey and item non-response
– Prospect of data for multiple years
– Issue: many returns (~80%) concern whole enterprise
– One solution: combine with FPQ data
Existing instance of linked data
ONS Annual Respondent Database (ARD2) matched
onto WERS 2004
– 1998-2004 Panel Survey: 166 (~25%) of trading sector w/ps
  have financial data for both years
– Issue: fewer than 30 have site-level data for both years; no
  WERS Financial Performance Questionnaire
– General issue: available only in ONS Virtual Micro-data Lab


Further info: Forth and McNabb (2007) WIAS
 Technical Paper No.1
Work in progress
ONS Business Structure Database matched onto
 WERS 2004
  – Version of the IDBR prepared for research use
  – Provides info on legal status and country of ownership
  – Provides information on enterprises’ internal structure
  – Provides info on demographic events: birth, death,
    merger, takeover
  – BSD record available for 2,143 (93%) of XS workplaces
  – One potential use: identify siblings for 714 (45%) of the
    XS workplaces that belong to multi-plant enterprises
Work in progress
WERS 2004 matched onto the ONS Annual Survey of
 Earnings and Hours (ASHE)
  – Annual survey of 165,000 employees
  – Detailed information about hours and earnings (basic,
    overtime, incentive payments, pension arrangements)
  – WERS can provide more information about the employer
  – Around 5,000 ASHE records (3%) link to XS workplaces
  – In most cases (95%), there are 2+ ASHE records per
    workplace
  – One potential use: examine union impact on composition
    of earnings
Further opportunities at ONS
• Surveys deriving from IDBR (to be investigated):
  – Community Innovation Survey (CIS4): data on
    product/process innovation and R&D from 16,000
    enterprises over period 2002-2004
  – E-commerce Inquiry: annual survey of 10,000 enterprises
    covering access to ICT and e-commerce activity
  – Annual Inquiry into Foreign Direct Investment: annual
    survey of 16,000 enterprises covering foreign ownership
    and financial flows
Matching on summary data items
Possible options (dependent on demand):
• Industry:
  – Levels: down to SIC 2003 Class / Sub-class level
  – Examples: aggregate productivity, profitability, labour
    costs
• Occupation:
  – Levels: down to SOC 2000 Minor Group
  – Examples: composition by gender, ethnicity, age;
    average qualifications; average wages
Matching on summary data items
Possible options (dependent on demand):
• Geography:
  – Levels: any postcode-based classification that does not
    compromise the anonymity of respondents. See National
    Statistics Postcode Directory for options.
  – Examples: unemployment and vacancy rates by TTWA;
    profile of labour force in TTWA; urban/rural indicator.
Comments and suggestions

   WERS Information and Advice Service

      URL: http://www.wers2004.info

      Email: wers2004@niesr.ac.uk

     Telephone: +44(0) 20 7654 1933