Skip to main content
Publication, Part of

GP Earnings and Expenses Estimates 2018/19

Official statistics,
Publication Date:
Geographic Coverage:
United Kingdom
Geographical Granularity:
Country, Regions
Date Range:
01 Apr 2018 to 31 Mar 2019


Data sources

GP populations for the earnings and expenses estimates are taken from several sources:

  • the workforce Minimum Data Set (wMDS) for GPs in England. The wMDS became the primary date source for GP data in England in September 2015; prior to this point, data were provided by the National Health Authority Information System (NHAIS). Completeness of this data set is improving from an initial figure of 88.1 per cent in 2015 (refer to the General Practice Workforce report for the percentage for the year in question).
  • the National Health Authority Information System (NHAIS) for GP payments for GPs in Wales
  • from NHS National Services Scotland Information Services Division
  • the Northern Ireland Business Services Organisation

Earnings and expenses data come from Self Assessment tax returns held and analysed by HM Revenue and Customs statisticians.

Earnings and expenses estimates for contractor GPs are based on their medical income from self-employment sources. Income from employment sources is not included in averages for contractor GPs. For salaried GPs, estimates are based on all income from employment sources and medical income from self-employment sources.

GPs can perform both NHS and private work which can be done both inside and outside the practice, including the NHS Out of Hours services. GPs will usually submit a Self Assessment tax return which contains information on all of their self-employment earnings, including both NHS and private earnings while practising as a GP. Therefore, the estimates include earnings and expenses relating to both NHS and private work. It is not possible to distinguish between and disaggregate NHS and private earnings using this data source.

GP data are matched with anonymised data from Self Assessment tax returns to produce the figures underlying the tables and findings in the report. These earnings and expenses estimates are weighted up to the estimated GP population, and then sent to NHS Digital in the form of aggregate non-disclosive summary statistics for inclusion in the report.

The data set, the process of merging the data set with tax data, and the analyses performed by HMRC statisticians on behalf of NHS Digital are described in further detail in this document. 

Deriving the GP population for analysis

All full time and part time UK contractor and salaried GPs (working in the NHS under a General Medical Services (GMS) or Personal Medical Services (PMS) contract) for whom information is available, and who have an accounting year ending in the final quarter of the fiscal year (i.e. between 1 January and 5 April) are included in the sample. GPs who work solely as locums or freelancers are not included.

GPs who are solely Alternative Provider Medical Services (APMS) contracted are not included. However if a GP holds both a GMS/PMS contract and an APMS  contract, their earnings and expenses from any contract may be included. GPs with both GMS and PMS contracts are removed unless their primary contract can be identified.

The data set includes GPs working under a variety of GP type classifications, which denote their relationship to the contract held with a particular Primary Care Organisation (PCO). PCOs hold a contract with GPs who deliver an agreed level of general practice services. A contractor GP is a practitioner who entered into a contract with a PCO, either as a single-hander or in partnership, to provide primary care services. A contractor GP may employ salaried GPs. A salaried GP could also be directly employed by the PCO. The cost of employing a salaried GP could form part of the employee expenses of contractor GPs. If the salaried GP was employed directly by a PCO then the contractor GP(s) within the practice in which the salaried GP works will not have incurred the expense.

Results are presented by three GP types – contractor, salaried and combined. A GP listed in a data set at least once as a contractor, regardless of any other arrangements under other contracts, is identified as a contractor. Figures for combined GPs are a weighted average based on employment and self-employment income for salaried GPs and self-employment income for contractor GPs.

The four GP data sets are validated and duplicate records are removed. During processing, additional fields are derived such as age band (derived from date of birth), practice rurality classification (using organisation postcode) and GP practice size banding (derived from the practice’s summary count of registered patients). The data are combined into a single file and sent to HMRC for processing.

HMRC analysis

The GP data set is sent to the Knowledge, Analysis and Intelligence (KAI) division at HMRC, in order to produce weighted and aggregate non-disclosive summary statistics on earnings and expenses estimates for GPs.

Only those GPs with accounting years ending in the fourth quarter of financial year (i.e. 1 January to 5 April) are included.

The tax data covers income from all NHS and private work.

During validation activity, HMRC statisticians apply exclusion criteria to the data set in order to derive the sample upon which to perform their earnings and expenses analyses as follows:

  • GP not found or no Self-Assessment tax return when data set created
  • GP with accounting period not in the relevant year or not 12 months (as part-year GPs are likely to earn less than the full-year cohort, and their income could distort the figures)
  • GP with inconsistent/incomplete earnings and expenses information
  • GP with non-medical income reported
  • Contractor GPs with employment income but no self-employment income
  • Salaried GPs with no employment income
  • GP with no accounting period in quarter four

These exclusions may not be exhaustive and the sample may include a small number of GPs with non-medical income. While this would not affect high-level results, figures for groups with low sample and population counts may be affected by any extreme values.

As figures are estimates and to acknowledge a degree of sampling error, population counts are rounded to the nearest 50 and monetary values are rounded to the nearest £100.

To maintain taxpayer confidentiality, HMRC suppresses results for any analyses that would produce results for subgroups with low sample numbers and footnotes are provided as necessary.

HMRC provide standard error information for taxable income which is used to calculate statistical significance at the 5% level in the Main Findings and where those values are shown throughout the report.

Stratification of the population and weighting of the results

As this analysis is based upon a sample, stratification is used to enable weighting corrections within strata to account for GPs who are not part of the sample. This ensures that the effect of bias is minimised in the final findings.

The contractor GP population is allocated to one of 12 strata according to country, contract and dispensing/non-dispensing status. Table 1 shows the stratification variables.

Table 1: Stratification variables for GP Earnings and Expenses Estimates, Contractor GPs

Dispensers Non-Dispensers
GMS, England GMS, England
PMS, England PMS, England
GMS, Scotland GMS, Scotland
PMS, Scotland PMS, Scotland
GMS, Wales GMS, Wales
GMS, Northern Ireland GMS, Northern Ireland

The salaried GP population is allocated to one of eight strata according to age and sex. Table 2 shows the stratification variables.

Table 2: Stratification variables for GP Earnings and Expenses Estimates, Salaried GPs

Male Female
<35 <35
35-40 35-40
41-50 41-50
>50 >50

Earnings and expenses estimates are based on a sample and are weighted according to the contractor/salaried GP population. One set of weighting factors are derived, based on strata, and the same set of weights are applied throughout for all analyses. This reduces complexity and potential risk of error.

As results are estimates based on samples that are weighted to the full contractor/salaried populations, they are subject to sampling error. This is because using information from or about a sample of a population, regardless of the steps taken to mitigate the effects, can never be as accurate as using the entire population. Differences between groups and subgroups of GPs or between reporting years may not be statistically significant and where significance has been tested, this is noted in the Main Findings.

Results for combined GPs (contractor and salaried) are calculated by producing a weighted average based on employment and self-employment income for salaried GPs and self-employment income for contractor GPs.

Expenses to earnings ratio (EER) figures presented are calculated by dividing average expenses by average gross income for the grouping being considered. This is not the same as the weighted average EER for that grouping.

Technical Adjustments for Pension Contributions

Employer Superannuation Contributions for Contractor GPs


Prior to the introduction of the nGMS contract in 2004/05, PCOs paid the employer’s superannuation contributions of GPs’ pensions schemes directly to the NHS Pensions Agency. The money did not appear in the practice or individual tax returns and consequently did not form part of income before tax.

However, from April 2004 onwards, under the nGMS contract for contractor GPs, the employer superannuation contribution to the GPs’ pension scheme was included in the global sum payment made to practices and GPs became responsible for the payment of both their employee and employer superannuation contributions. Therefore, employer contributions should have been included in income reported on tax returns and tax relief claimed. This means an estimate has to be made of employer contributions, in order to:

  • remove the monies from the income before tax figures
  • make figures comparable with those from previous years (under the old contract)
  • be a more valid representation of the average gross earnings and average income before tax of GPs

The procedure for the payment of GP pension contributions begins at the start of each financial year and involves GPs producing an estimate of their income before tax for the forthcoming year, and therefore an estimate of what their pension contributions should be. This estimated pension contribution is deducted from their global sum payment by their PCO, and at the end of the financial year the actual contributions due are calculated by the PCO on the basis of a certificate completed by the GP after they have submitted their tax return. The GP either receives a refund if contributions had been overestimated or has to pay shortfall contributions.
There is a time lag involved which means refunds or top-up payments are made in arrears. For example, a GP with a shortfall of contributions for 2018/19 did not pay these (and claim tax relief on payments) until at least tax year 2019/20 and possibly even 2020/21.


In order to estimate the amount of employer superannuation contributions to be deducted, HMRC income before tax figures from previous years are compared with data on superannuable income from pensions agencies for all countries. Subsequently, estimates can be made on what proportion of income is NHS income, and therefore how much of the total income to adjust.

Due to the time lag in receiving pensions data, the two sets of data are not directly comparable. Therefore, an average percentage of NHS income is taken over the latest five years of pensions data available which also takes into account year-on-year fluctuations. In the event of not receiving the latest year of pensions data in time for the analysis to take place, an average of the latest available five years of pensions data is taken.

The proportion of income estimated to be NHS income is then adjusted by the NHS pension employer contribution rate relevant to the reporting year, and to the country of the GP whose income before tax is being adjusted.

In 2018/19, those employer contribution rates were: 14.38% in England and Wales, 14.9% in Scotland and 16.3% in Northern Ireland. These are the percentages used for ‘Employer contributions percentage’ in the methodology below.

The methodology is:

% NHS Income = Average NHS Superannuable Income per country   x 100  = n1

                                         Average EEQ Net Income Before Tax

(Average % NHS Income figure calculated over four or five years as described).

The resultant percentage figure feeds into the equation below to give the final figure from which the employer’s pension contribution should be deducted.

NHS superannuable income before tax for each individual GP (y1) =

 n1    x  GPs EEQ income before tax (h1)                                                                                


Employer contributions to be deducted (e1) are:

Employer contributions percentage   x y1 = e1


Adjusted EEQ income before tax = h1 – e1

The adjustment is applied to those GPs who declared an amount in the tax relief box ‘payments to your employer’s scheme which were not deducted from your pay before tax’ on their tax return. All GPs that recorded an amount in this tax relief box are members of the NHS pension scheme (GPs with other personal pensions claim tax relief using other boxes on the form).

Pensions data is not used to determine membership as it is very out of date, and there is a risk that the latest information is not received in time for the analysis to take place. Also, the pensions data does not cover the entire NHS Pensions membership.

The methodology for adjusting for employer superannuation contributions changed for the 2009/10 analysis. The previous methodology is described in Annex B of previous editions of GP Earnings and Expenses.

Employee Superannuation Contributions for Salaried GPs

In order to put the salaried results on the same basis as the contractor GP results, an adjustment is required to add the employee contributions (plus Additional Voluntary Contributions (AVCs)) to the employment income of salaried GPs.

The adjustment is made by adding a percentage to the income before tax for each GP which takes into account their employee pension contributions.

Since 2015, these contributions have been based upon earnings tiers determined by the GP’s full time equivalent earnings.

The percentages by which salaried GPs’ earnings are adjusted are calculated for each country using the tier for the average income for salaried GPs in the previous year.

The pensionable earnings for each tier vary according to year and by country. For England, Wales and Northern Ireland, the pensionable pay band for whole time equivalent GPs, used to determine pension contribution rates from 2014/15 have been as follows:

Tier 2014/15 Pensionable WTE earnings up to Contribution rate for Scheme year 2014/15    Pensionable WTE earnings band up to Contribution rate for Scheme years 2015/16 to 2018/19
1 £15,431.99 5.0%   £15,431.99 5.0%
2 £21,387.99 5.6%   £21,477.99 5.6%
3 £26,823.99 7.1%   £26,823.99 7.1%
4 £49,472.99 9.3%   £47,845.99 9.3%
5 £70,630.99 12.5%   £70,630.99 12.5%
6 £111,376.99 13.5%   £111,376.99 13.5%
7 £111,377+ 14.5%   £111,377+ 14.5%

In 2017/18:

  • the average income before tax of salaried GPs in England was £58,424.47
  • salaried GPs in Wales earned an average of £52,118.57
  • salaried GPs in Northern Ireland earned an average of £56,688.91

This means salaried GPs in England, Wales and Northern Ireland were all in tier 5 with a pension contribution rate of 12.5%.

The pensionable pay bands for whole time equivalent GPs in Scotland which determine pension contribution rates from 2014/15 to 2017/18 were:

Tier Pensionable WTE earnings band up to Contribution rate for Scheme year 2014/15 Contribution rate for Scheme year 2015/16   Pensionable WTE earnings band up to Contribution rate for Scheme year 2016/17   Pensionable WTE earnings band up to Contribution rate for Scheme year 2017/18 Pensionable WTE earnings band up to Contribution rate for Scheme year 2017/18
1 £15,828.99 5.0% 5.2%   £16,128.99 5.2%   £16,528.99 5.2% £16,928.99 5.2%
2 £21,601.99 5.6% 5.8%   £21,817.99 5.8%   £22,217.99 5.8% £22,439.99 5.8%
3 £27,089.99 7.1% 7.3%   £27,360.99 7.3%   £27,634.99 7.3% £27,910.99 7.3%
4 £49,967.99 9.3% 9.5%   £50,466.99 9.5%   £50,971.99 9.5% £51,481.99 9.5%
5 £71,337.99 12.5% 12.7%   £72,050.99 12.7%   £72,770.99 12.7% £73,498.99 12.7%
6 £111,376.99 13.5% 13.7%   £112,490.99 13.7%   £113,625.99 13.7% £114,760.99 13.7%
7 £111,377+ 14.5% 14.7%   £112,491+ 14.7%   £113,626+ 14.7% £114,761+ 14.7%

In 2017/18, salaried GPs in Scotland earned an average of £62,857.14, putting them in tier 5 with a pension contribution rate of 12.7%.

Therefore, the percentages added to the employment income of salaried GPs for employee contributions for the 2018/19 report was 12.5% for England, Wales, and Northern Ireland and 12.7% for GPs in Scotland.

In addition to the employee contributions, and in keeping with the practice of previous years, an additional 0.53% has continued to be added for salaried GPs in all four countries as an estimate for AVC contributions. This 0.53% is based on information originally received from the West Yorkshire Central Services Agency.

Consequently, a total of 13.03% was added to the income before tax of salaried GPs in England, Wales and Northern Ireland and 13.23% for salaried GPs in Scotland.

The notion of AVCs has changed in recent years, with options closed and less lucrative and therefore the average AVC amount is likely to be smaller than 0.53%, but as contemporaneous data is not available, it has not been possible to calculate a more up-to-date estimate.

The following assumptions are made in order to make the adjustment:

  • All salaried GPs (less than 60 years old) are part of the NHS Pension Scheme
  • All employment income before tax of salaried GPs is NHS income (and therefore pensionable)
  • The proportion of AVCs paid by salaried GPs in the West Yorkshire Central Services Agency when the methodology was originally approved is representative of the UK

The adjustment made to account for salaried GPs’ employee pension contributions applies only to their income from employment; any income from salaried GPs’ self-employment is not affected. Since the earnings and expenses estimates consider only self-employment for contractor GPs, this adjustment is not required for contractor GPs. 

Where a salaried GP earns both employment and self-employment income, the percentage is only added to the employment income before an average of self-employment and employment income is calculated for that individual.

Means, Medians and Quartiles

The mean is the average and can be defined as the sum of a list of values divided by the number of values in that list.

The median is the "middle" value in an ordered list of values, it is a point that splits the values in two, half above this point and half below.

Quartiles are points in an ordered list of values that has been split into four parts, each comprised of an equal number of values, the 1st quartile is the same as the 25th percentile, the 2nd quartile is the same as the median, etc. half of the values fall between the 1st and 3rd quartiles.

The position of the mean relative to the median can sometimes reveal information about the existence and/or extent of extreme values.


To add context to the analyses, earnings and expenses estimates are broken down in the Excel Time Series according to whether a practice is defined as being ‘rural’ or ‘urban’. Since 2008/09 the rural/urban classification has been based solely on the postcode of the practice rather than taking into consideration the postcodes of the registered patient populations which may be more variable.

NHS England and NHS Improvement regional splits

The structure of the NHS in England changed on 1 April 2013. The country and regional sections of the earnings and expenses estimates report that previously included Strategic Health Authorities (SHAs) now contain NHS England and NHS Improvement Regions, formerly known as Commissioning Regions. To ensure that the figures are as meaningful as possible, in view of the fact that earnings and expenses estimates are published approximately 18 months in arrears, the definitions of the regions for which results are presented in the Excel Time Series are effective as at 31 March in the publication year.

Last edited: 9 September 2020 1:35 pm