Summary of Changes to Main Data Sources

1 While research and consultation are invaluable, official data also have an essential role to play in assessing the impact of the National Minimum Wage. In this appendix, we review the most important changes made since the 2005 Report to the data sources used in our analyses of employment and earnings. These are: the Labour Force Survey (LFS), the Office for National Statistics (ONS) employee jobs series and the Annual Survey of Hours and Earnings (ASHE).

Employment

2 There are two main sources of employment information, the LFS and the ONS employee jobs series. The two sources measure different things and give different results: the LFS captures the number of people in employment, while the employee jobs series measures the number of jobs in the economy.

Labour Force Survey

3 The LFS is a quarterly survey of about 60,000 UK households and is the official data source used to measure the number of people in employment and unemployed. The data set is a rich source of information on personal and socio-economic characteristics including information on an individual's current and previous jobs, income, gender, region, industry, occupation, ethnicity and disability.

4 There are two main sources of LFS data:

• those published on the ONS website, which are weighted to the latest population estimates. These are generally seasonally adjusted and published monthly in the ONS Labour Market Statistics Integrated First Release; and
• the LFS Microdata, which are weighted to population estimates in Spring 2003. These data produce estimates of the UK population that are lower than the First Release and are not seasonally adjusted.

5 Up until January 2006 the LFS Microdata were published in seasonal quarters ­ Spring (March­May), Summer (June­August), Autumn (September­November) and Winter (December­February). Since then, they have been released on a calendar quarter basis ­ Q1 (January­March), Q2 (April­June), Q3 (July­September) and Q4 (October­December) ­ in response to Eurostat requirements for consistent series across the European Union. Unfortunately, the ONS has only produced a limited back series on a calendar quarter basis.

6 In this report, analyses of aggregate employment, unemployment, hours worked, redundancies and productivity are based on the monthly and quarterly (calendar) data published on the ONS website. For detailed analyses of the labour market by age, ethnic status and disability, we have used the LFS Microdata based on seasonal quarters from Spring 1998 to Summer 2004 and calendar quarters from Q4 2004 to Q3 2006, with a break between the two series.

Employee Jobs

7 Figures for employee jobs are derived from the Short Term Employer Surveys, which collect information on the number of employees from a sample of 9,000 businesses in production industries each month and 30,000 businesses in the service and distribution industries each quarter. The strength of the ONS employee jobs series lies in the rich and timely industrial breakdown that it provides. However, figures are only available for Great Britain and are not seasonally adjusted.

8 The employee jobs series is published quarterly in March, June, September and December and is benchmarked annually to the latest results from the previous Annual Business Enquiry (ABI). However, in 2006 the usual December ABI revisions to the data were not published. As a result, the employee jobs data used in this report, which cover the period up to September 2006, are still benchmarked to the 2004 ABI.

9 Since September 2005, social work activities with accommodation (Standard Industrial Classification (SIC) 8531) and social work activities without accommodation (SIC 8532) have been merged and it is no longer possible to separately identify these two sectors. Our employee jobs analyses in this report are therefore based on the amalgamated sector, social work activities (SIC 853).

Earnings

10 There are two main sources of earnings data that we use in this report ­ ASHE and LFS. ASHE is regarded by the ONS as the best source of information on earnings but for analyses of disability and ethnicity we have used the LFS.

Annual Survey of Hours and Earnings

11 The main source of structural earnings data in the UK is ASHE, a survey of employees completed by employers and conducted in April each year. It is based on a 1 per cent sample of employees in pay-as-you-earn income tax schemes. The self-employed are excluded.

12 ASHE contains information on the levels, distribution and make-up of earnings as well as on hours, gender, age, geography, occupation and industry. However, there is no information on employees' ethnic backgrounds nor does it collect information about disability.

13 In 2004, ASHE replaced the New Earnings Survey (NES) as the principle source of structural information on earnings. Compared with its predecessor NES, ASHE has improved coverage (especially of the low-paid); is weighted to UK population totals; and includes imputation for some missing data (item non-response). In addition, a new questionnaire was introduced in 2005 which included improvements to the collection of data relating to allowances, pensions and incentive pay.

¹ Output Areas are the smallest geography at which demographic data are released from the 2001 Census. They replace the previously used postcode-based enumeration districts as the smallest building block for larger geographies.

14 Further changes were made to ASHE in 2006. First, the geographic basis of the survey moved to output areas instead of enumeration districts¹. The change has a negligible impact at the level of government office region but is more significant for lower level geographies. Second, ASHE weights were revised to take account of the move from a seasonal to a calendar quarter basis in the LFS. Third, the VAT-only supplementary survey was dropped as it had not significantly improved the coverage of the low-paid.

15 The discontinuities produced by these changes to the ONS earnings series result in three earnings data sets:

• The original New Earnings Survey (NES) covers the period from 1970 to 2003;
• ASHE without supplementary information covers the period 1997­2004. This series is basically NES data with imputation and LFS weighting; and
• ASHE with supplementary information covers the period 2004­2006. The ONS regards this series as giving the best estimates of low pay. Thus, we use this series where possible. It also replaces the previously used low pay central estimate ­ a combination of LFS and NES/ASHE without supplementary information.

16 It should also be noted that our analyses of ASHE use low-pay weights, which take into account absence from work and those not on adult rates. With the exception of the low-pay estimates, our analyses differ from those available on the ONS website because the ONS uses weights based on those adults whose pay has not been affected by absence.

Labour Force Survey

17 Given that ASHE does not provide information on disability or ethnic background, the LFS is our main source of information on the earnings of workers with disabilities and minority ethnic groups. But data on pay and hours in the LFS are less reliable than in ASHE, especially when provided by proxy respondents. Individuals tend to record more hours worked, possibly leading the derived hourly earnings variable to underestimate hourly pay. However, the ONS has done some work recently to improve the methodology to derive low-pay estimates from the LFS and this new methodology yields similar earnings results to ASHE. See Hayes, Ormerod and Ritchie (2007), Ormerod and Ritchie (2007a, 2007b) and Griffiths, Ormerod and Ritchie (2007) for further details.