Glen Watson Director Statistical Outputs Group ONS

1
Glen WatsonDirectorStatistical Outputs GroupONS
The increasing value of administrative records to
create statistics
2
SCOPE OF TALK

• 1. Features of Admin data Survey data?
• 2. What we collect/use already in ONS
• 3. Why is admin data strategically important to
  ONS?
• 4. International Experience
• 5. What data wed like to access and why
• 6. What were doing about it
• 7. How statistical independence helps
• 8. What it all means for users
• 9.Conclusions

3
1 ADMINS vs SURVEY DATA
4
2 ONS use of admin data - examples

• Registration data (births, deaths, marriages,
  civil partnerships)
• IDBR
• Wealth Assets
• Census LS and claimant count linkage
• Neighbourhood Statistics

5
3 Why is admin data so important?

• Save money on surveys and the Census
• Reduce compliance costs

6
Compliance costs

• Compliance QPT
• Where
• Q no. of questionnaires despatched
• P rate of pay for respondent
• T time taken to complete questionnaire

7
Cost of complying with ONS business surveys (m)
2010 Target -19
2015 Target -25
8
3 Why is admin data so important?

• Save money on surveys and the Census
• Reduce compliance costs
• To meet new demands (sub groups, small areas,
  longitudinal)
• Improve quality/coherence
• Declining survey response rates

9
Survey response rates ()
10
not just me that says so

• 1999 - Social Exclusion Unit PAT 18 - Better
  Access to Information
• 2000 - PMSU Adding it up Analysis Modelling
  in Central Government
• 2002 - PMSU Privacy Data Sharing Way Forward
  for Public Services
• 2003 - DfES Every Child Matters
• 2004 - ODonnell Financing Britains Future -
  Review of the Revenue Depts
• 2004 Allsopp Review of Statistics for Economic
  Policy Making
• 2005 - Hampton Reducing Administrative Burdens
• 2005 - Cabinet Office Transformational Govt
• 2006 - House of Commons TSC Independence for
  Statistics
• 2006 - MISC 31 Vision to improve public
  services and save money, must get better at
  using/reusing data. Must be proportionate and
  protect peoples rights.
• UN Fundamental Principles of Official Statistics

11
4 - International Experience

• Most economic competitors have statistics
  legislation that addresses data sharing
• Europe France(1951), Italy(1989), Norway(1989),
  Austria(2000)Finland(1992), Eire(1993),Czech(1995)
  , Germany, Netherlands.
• USA(2002) and Canada (1970 to 1981), Australia..
• Canada has reduced the burden on business by up
  to 25, Australia by 40
• Reforms to structural business statistics XBRL
• Register based population/census systems

12
Census traditional or register based?
13
5. What wed like to access and why

• Business and economic statistics
• Corporation tax returns , VAT returns, PAYE, Self
  Assessment tax returns, company accounts, scanner
  data.
• Population migration statistics future Census
• NINOs, Child Benefit, PLASC, WPLS, e-borders.
• Social, economic and neighbourhood analysis
• Work Pensions Longitudinal Study, Lifetime
  Labour Market Database, Council Tax.

14
(And its not just about ONS wants)

• DWP WPLS / Census linkage
• DfES / HMRC earnings data linkage
• DfES / ONS life events linkage

15
6 - What are we doing about it?

• Failed attempts to secure legislation
• The Business Case
• Admin sources strategy
• Pilots
• Working with HMT on independence

16
The Business Case
Benefits Additional analysis, reduced cost,
reduced compliance
Options 1. Full access rights 2. Enabling
legislation, with individual statutory
instruments to follow 3. Name specific data
sources in primary legislation
Recognition that it needs to be two way
Safeguards
17
A bad week in the office?
18
Pilots - Companies House data

• FAME database (Bureau Van Dijk)
• 2.5 million public/private UK/Eire companies,
  almost all data from Companies House
• Linking with ABI2 records, focus on lt250
• 64 matched
• 71 of these had turnover within 5

19
Pilots - Companies House data
20
Pilots - Using HMRC data
21
7 - Statistics Board Bill

• Must preserve access to admin data ONS already
  uses
• Likely to need to provide access to registration
  data if GRO moves
• Will probably take the form of enabling
  legislation
• with secondary legislation on a case by case
  basis
• Likely to provide provisions for two way sharing

22
8 What it all means for users

• How can users influence ONS priorities?
• Access to data archive, VML, Census SARs etc ?
• Availability of admin data to external
  researchers ?

23
9 - Conclusions

• Already make use of admin data
• Clear benefits for making more use
• UK behind most major countries
• Hopeful that independence will be the catalyst
• Confidentiality safeguards needed
• Needs careful planning and proper resourcing