Regis Tremblay

Writing about work: who does it, on what terms, and how the claims made about it compare with what has been measured.

Who works where ยท 1.3

Self-employment, and the four things it means

Self-employment, and the four things it means. What is actually the case, and how it compares with what is repeated.

For a practical software perspective on this topic, this Monitask guide provides a concise companion guide.

Self-employment is a single line in the statistics covering at least four arrangements that have almost nothing in common.

The four

The independent professional or tradesperson. Multiple clients, control over the work, able to turn business down. The picture the word conjures, and the smallest of the four in most countries.

The owner-manager. Runs a business, may employ others, and is self-employed in the sense that nobody else employs them. Their situation resembles a small firm rather than a worker.

The dependent contractor. One client, who sets the hours, the rates and the manner of the work, and who would be called an employer if the arrangement were named honestly. Legally self-employed and economically not.

The platform worker. Work allocated by an application, terms set unilaterally, no negotiation. Whether this is a distinct category or a variety of the third is a live question in several jurisdictions.

The definitional problem that moves the number

Whether an owner-manager of an incorporated business counts as self-employed or as an employee of their own company differs between statistical systems, and it is a large group.

Some countries publish both figures, some publish one, and comparisons across borders frequently place one country's narrow measure beside another's broad one. A reported rise in self-employment can be a change in incorporation incentives in the tax code and nothing else.

What the classification decides

In most jurisdictions, whether the minimum wage applies. Whether there is sick pay, holiday pay, notice, or protection against unfair dismissal. How tax and social contributions are calculated, and by whom. And whether the person may bring a claim at all.

Those are large differences and they are decided by a test, applied after the fact, on the substance of the relationship rather than on what the contract calls it.

How the tests work, in outline

They ask about control, about substitution, about who bears financial risk, about whether the person is integrated into the organisation, and about mutuality of obligation. The exact list and the weighting differ by jurisdiction and some places have moved to simpler statutory tests.

The important structural point is that the label on the contract is not decisive anywhere. A person described in writing as a contractor may be found to be an employee, and the finding is retrospective.

Nothing here is legal advice, the tests differ between countries and between states, and they have been changing quickly.

Why the ambiguity persists

Because it is expensive to resolve. Establishing status requires a claim, a claim requires somebody willing to bring one, and the person best placed to bring it is usually the one who most needs the work to continue.

So the ambiguity is not an oversight. It is a stable arrangement in which the risk of being wrong sits with the party least able to carry it.

Reading a self-employment figure

Ask which of the four it contains. Ask whether incorporated owner-managers are in or out. Ask whether it counts main activity or any activity. Three questions, and a headline number rarely survives all three unchanged.

The direction that is real

Across several countries the dependent end of the range has grown faster than the independent end. That finding is more robust than any particular figure, because it holds across definitions rather than depending on one.

The second job

A growing share of self-employment is a second activity alongside employment. Whether such a person appears in the self-employment figures depends on whether the survey records main activity only, which most do.

So a real increase in supplementary self-employment can be almost invisible in the headline series, while showing clearly in tax data. The two sources tell different true stories.

What a good statistic here would look like

Self-employment split by number of clients, by whether the person can send a substitute, and by whether they set their own rates. Those three would separate the four arrangements almost perfectly and none of them is routinely collected.

Some surveys have added questions along these lines and the results are informative wherever they exist. It is the clearest case in this part of a measurement that would be cheap and is not made.

What has been changing in law

Several jurisdictions have moved from multi-factor tests towards statutory presumptions, under which a worker is treated as an employee unless the engager demonstrates otherwise. The direction is broadly consistent and the detail is not, and it is changing quickly enough that anything written here about a specific rule would be out of date before it was read.

What this rests on

  1. Labour force surveys publish self-employment with varying treatment of incorporated owner-managers; the treatment is documented in each survey's methodology.
  2. Employment status tests are set by statute and case law and differ by jurisdiction. This entry describes their general shape and names no specific test as applicable to any reader.
  3. The observation that dependent forms have grown faster than independent ones appears across several national studies using different definitions.

For broader context, consult OECD self-employment data.

Also in who works where