Bob Bennett

Distinguishing someone who is self-employed from another who is waged seems at first obvious: one works for themselves, the other works for others. However, the distinction was often far from simple in the past, and today has been politically contested.  

The differences between employment and self-employment are important for demographic analysis, potentially affecting risk and legally who is responsible for hazards at work. It is important for economic analysis because it reflects different levels of agency or freedom and the dynamics of many businesses. It may be a critical influence on income: a very few who are self-employed may get astonishingly rich. It is also important for understanding motivation, possible levels of self-fulfilment, and how different people respond to changing demands and opportunities in an economy. This all has important historical and policy implications. 

Photo by Dominik Bednarz on Unsplash.

Modern discussions often focus on agency work and zero-hour contracts: is someone self-employed, or are they de facto employed by a business contracting them and giving minimal security? Should this be regulated by government? How should trade unions be involved? How should laws balance between people’s wish for flexible working, and for employers’ response to workforce needs that can fluctuate widely from day to day or week to week? 

The Employment Rights Bill considered by the UK parliament in 2025 is attempting to define and enforce the differences between self-employment and being waged. Its discussion in Parliament and outside demonstrates the current difficulties of definitions, but these are of long standing. 

When did the distinction become important? 

Historically there was freedom between contractors and employers to determine how they would be classified – as self-employed or employees. Arrangements were flexible and often ambiguous. The main force that began to change things was income tax.   

In Britain, income taxes on profits of trade and farming began in 1798, but only became continuous in 1842. Mostly those liable were clearly defined by the nature of their trade, and while the tax rate remained low it was not worth much effort to challenge it. Difficulties began when taxes on wages, which were minor at the outset, were widened in the late 19th century. It then became increasingly necessary to distinguish between waged and self-employed. However, this was initially fairly simple because only ‘salaries’ were taxed (generally affecting only higher earners). 

Arthur Whinnom, Pay Day. Woodhorn Museum.

Major difficulties emerged during and after WW1, when tax rates rose rapidly and the rate of tax on profits became considerably lower than that on wages. When an additional tax on payroll (NICs) was expanded after 1948, this taxed employees even higher than the self-employed. The spiral intensified as governments often favoured increasing NICs on payrolls rather than income tax rates as they were less politically visible. These changes resulted in a 100-year long shift of incentives for people to define their income as self-employment wherever possible.  

During the 1980s and 1990s, the numbers of self-employed increased for structural reasons, but a major element of growth was purely driven by ‘tax efficiency’ – that is, people who claimed they were self-employed when the justification was at best marginal or even bogus. 

Inevitably, tax collectors were keen to limit this effect. The most concerted attack was the IR35 rule introduced in 2000. This set a specific test: that having a regular single or main employer meant you were an employee, even if you were a one-person company. To be self-employed you had to be working on different contracts for different people, with independent negotiation of your prices and costs, whether you were a company or not. This test was up to individuals and had little effect. But in 2017 the test was shifted to public sector employers to determine liability. Employers were also made responsible for collection and any errors in definition (as for other employees through PAYE). This responsibility was extended to private sector employers in 2021.  

There was a ‘flight to safety’. A small reduction occurred of self-employed numbers in 2018, but there were large reductions from 2022. Many of the remaining own account proprietors contracted to large organisations moved to ‘umbrella’ agencies that administer their incomes. In the Figure 1 (below) it is difficult to separate this effect from the Covid pandemic, but most of the 2021-2 reduction is estimated to be the effect of the change in IR35 rule. Highly-paid celebrities who had claimed to be self-employed though working for a media company or the BBC made a lot of noise about their loss of ‘self-employment’. However, at the other extreme, many poorly rewarded agency staff were incorrectly classified as employed. The 2025 Employment Rights Bill attempts to tighten these definitions further. 

Figure 1. Self-employment in England and Wales 1994-2024. Source: LFS. [Discontinuities in data definitions over early years have been smoothed].

The ability for online platforms using Apps to connect suppliers of services with consumers has challenged some of the old regulations. Celebrated recent legal cases in Britain have had to define self-employment in light of these new circumstances. However, although the technology may be new, the decisions are almost identical to long-standing distinctions. 

For example, Uber in London was found by the Supreme Court in 2021 to be an employer. Their cab drivers were not self-employed; the contract for hire was between the passenger and company, not between the passenger and the cabbie despite them ordering through an App.  

The fine distinctions are clear from Deliveroo, where the Supreme Court found it was not an employer in 2023. In that case, the rider had freedom to accept or reject jobs and could do jobs for others. This may change under the 2025 legislation in parliament. 

Deliveroo rider taking the lane in Bristol.

Historical examples

The building industry is well-known for the difficulties of definitions. In the past, as now, the smallest and most numerous group of self-employed were jobbing builders: own account and small partnerships that employed only themselves or a small workforce. Most were single tradesmen such as bricklayers, joiners and carpenters or masons. Many larger firms employed these individuals and small partnerships on large jobs. Were the small sole proprietors then employed, or self-employed? 

Firms varied. The largest contractors for railways, canals and dock construction usually hired all navvies and most other workers for the life of the project, and had few permanent staff. In contrast, even the largest ‘master builders’ responsible for constructing churches, public buildings and factories mostly had a core permanently employed workforce of skilled craftsmen and foremen. When needed they employed labourers and a few extra self-employed craftsmen as extras. These workers might continue to be self-employed or be put on the employer’s payroll.  

One way to examine definitions is to look at pay ledgers. Detailed accounting in building firms for personnel attributed to each contract was used to calculate tenders, monitor works, and negotiate any extras needed on completed works. Ledgers were often divided into salaries for foremen, clerks, managers and senior staff, and wages for tradesmen. For John Mowlem & Co. in the 1880s, almost all people were on continuous pay ledgers; only lightermen, carters and extras on piece-work were recorded as self-employed.  

Dock Workers, Dundee. University of Dundee, Duncan of Jordanstone College Collection.

Franchises and ‘tied’ activities 

Pubs, inns and beer houses were often owned or tenanted by those who operated them, making them a nationwide source of self-employment, often important for women. Some pubs were tied to breweries who appointed waged managers, but until the late 19th century many tied pubs were tenancies run by the proprietor, with the tie only restricting the drink sold. Most innkeepers were self-employed. 

Some tied pubs were an early form of franchise. Franchises have become a very common business model in modern times. Much depends on the franchise contract as to whether they are sources of self-employment. Most modern fast-food chains and retailers rarely qualify as run by the self-employed. They are de facto branches, though there may be incentives with wages linked to size of turnover. 

Martha Matilda Harper is often credited as creating the first franchise model. Through domestic service she saved enough to open a hairdressers and began selling hair products, especially shampoo. Her business expanded by opening branches where she selected and trained other women using her brand. She had over 500 shops operating across the USA by late the 19th century; but their staff were de facto her employees. 

Isaac Singer’s Sewing Machine Company licenced the rights to sell its machines through agents who were charged a fee. Initially operating in the USA, the business expanded rapidly and by 1867 opened a factory in Glasgow, expanding to 60,000 sales per year by 1880 using about 300 offices in Britain, with numerous agents selling the machines.  

Workers leaving Singer sewing machine factory on Clydebank.

The Prudential used almost entirely an agent model to sell insurance policies: ‘the man from the Pru’ collected from the door. Founded in 1848 and incorporated in 1861, it had 10,600 agents by 1888 and 18,000 by 1902. Theoretically they could be self-employed, and initially many were also agents for other businesses. But in 1875 almost all were full-time working only for the Pru as de facto employees. The Pru’s district offices were almost entirely concerned with managing the agents. 

Worker-contractors and outwork  

Worker-contractors and outworkers highlight the problematic distinctions between self-employed and workers. Outworkers generally operate at home. Some could exercise a lot of control over the amount of work they took in, whether it was from one or multiple sources, and had some level of control over prices. Under these criteria they could argue they were self-employed. However, independence required not only motivation but realistic opportunities to trade independently. The majority of outworkers had no opportunities to exercise controls, and may not have wished to do so, preferring the security of a fixed price per piece or contract if the employer was viewed as fair. 

One manufacturer locally viewed as fair and having no history of disputes was Daniel Gurteen & Sons of Haverhill, Suffolk, a clothing, towelling, rug and hosiery manufacturer. In 1888 they had a major works with some 1,500 workers, two-thirds female. Additionally another 1,500 or so did work at their homes in the villages around. They had a fixed rate and their work was checked and managed. ‘Old men with donkey carts took out and fetched back the work’. Not only did these locals have little motive to trade independently, but Gurteens in any case monopolised most contracting opportunities.

Self-employment in 19th century Britain was not just about the trade or the person’s motives and desires for independence. It was also about where they were, and the scope their town offered to run an independent business. 

National statistics

In the 1891 Census, for the first time, everyone who was economically active was supposed to identify as either a worker, employer, or working on own account. Combining the last two of these categories should give us a count of all ‘self-employed’; everyone else who was economically active was a worker. 

Unfortunately, various defects in the wording and design of the 1891 census, and also its administration, make the resulting information difficult to use. The question was re-designed in 1901. The fundamentals of the census question then continued until the modern censuses, though in almost all censuses the challenges of definition of self-employment make the responses problematic in detail.  

The 1901 census question.

It is notable that the census did not define an employer of domestic servants as an employer: these employers were not counted in census number, yet domestics were counted as employees – resulting in difficulties of balancing estimates of national self-employment rates. 

As a result of this and other difficulties of census design, most modern economic analysis has used the Labour Force Survey (LFS) in preference to the census. This was introduced in the 1970s. However, even this has suffered major problems, and since the 2010s has had severely declining response rates. The Bank of England and others needing an accurate view of self-employment in the economy have been highly critical. In the late 2020s it is planned to introduce a new version of the LFS. Defining self-employment remains a continuing problem! 

Further reading 

Open access 

• UK Government Employment Rights Bill: factsheets https://www.gov.uk/government/publications/employment-rights-bill-factsheets 
• The British Business Census of Entrepreneurs maps self-employed women and men 1851-1911 and gives access to the full data by area. https://www.bbce.uk/ 
• Drivers of entrepreneurship and small business: project and working papers using the UK census 1851-1911. https://www.campop.geog.cam.ac.uk/research/projects/driversofentrepreneurship/ 
• Bennett, R. J., Smith, H., and van Lieshout, C., Employers and the self-employed in the censuses 1851-1911 (2017). https://doi.org/10.17863/CAM.9640 
• Bennett, R. J., Employers and self-employed in the censuses 1921-2011 (2020). https://www.repository.cam.ac.uk/handle/1810/300054 
• Bennett, R. J., Montebruno, P., Smith, H., and van Lieshout, C., Entrepreneurial discrete choice: Modelling decisions between self-employment, employer and worker status (2019) https://doi.org/10.17863/CAM.37312 
• Bennett, R. J., Montebruno, P., van Lieshout, C., and Smith, H., ‘Business entry and exit: Career changes of proprietors in England and Wales 1851-81 using record-linkage’, Social Science History (2022).  https://www.doi.org/10.1017/ssh.2021.48 

Paywall 

•  Bennett, R. J., Smith, H., van Lieshout, C., Montebruno, P., Newton, G., The Age of Entrepreneurship: Business proprietors, self-employment and corporations since 1851 (Routledge, 2019). https://doi.org/10.4324/9781315160375