Estimates of the number of people in low paid jobs, and specifically those paid less than the national minimum wage, are published by the Office for National Statistics (ONS) annually in October.

It is important to note that these estimates do not measure non-compliance with the national minimum wage legislation. This is because the surveys used to provide the estimates do not indicate whether an individual is eligible for the minimum wage. For example, some people, such as apprentices or new trainees, are exempt from the minimum wage or are entitled to lower rates. If employees receive free accommodation, employers are entitled to offset hourly rates.

The estimates are based on analyses from two surveys: the New Earnings Survey (NES) and the Labour Force Survey (LFS). Separate, but similar, low pay estimates are derived from each of these two surveys. The central point between the two represents the best estimate of the number of low-paid jobs, and this is what ONS publishes.

In terms of measuring low pay, both the NES and LFS have limitations. ONS introduced an improved methodology in 2002 which corrects for most of the deficiencies. The data for previous years were re-calculated on the new basis.

The NES information should be accurate as it is taken from payroll records. However, the survey has limited coverage of those workers earning below the Inland Revenue’s PAYE threshold. In addition, the routine survey estimates are not grossed up to population level, causing concern regarding the effect of non-response on low pay estimates. The NES data have therefore been grossed up to population level to reduce the effects of non-response and under-representation of the low-paid in the survey.

The LFS has more complete coverage of jobs, especially of casual and low-paid jobs, but the individual data on earnings and hours are less precise, particularly when supplied by proxy respondents. To mitigate this, the LFS estimate has been based on hourly rate information, where available, and on an imputed hourly rate for those respondents who are unable to supply an hourly rate.

As part of the redevelopment of methodology for the NES, in 2004 ONS will also be reviewing the methodology for calculating low pay estimates.