Like any good villain in a horror movie, some policy ideas are both terrible and refuse to die. Think of Arnold Schwarzenegger in The Terminator. One example is “mandatory ethnicity pay gap reporting”.

Studies show mandatory diversity schemes tend to fail

This is the idea that large employers be compelled to publish the difference in average incomes between ethnic groups. Such measures already exist to publish data on differences between the sexes.

Mandatory ethnicity pay gap reporting was government policy in 2018 but that changed in response to the Sewell report, which recommended it remain voluntary. But that has not deterred the House of Commons Women and Equalities Committee. It said in February there was “no excuse” for a lack of ethnicity pay gap statistics.

Then last week, it issued a statement saying the government displays a “lack of will or care” to “reduce disparities between ethnic groups”. Its chair, Tory MP Caroline Nokes said: “Introducing mandatory ethnicity pay gap reporting for larger businesses would set the ball rolling, reducing inequalities between different ethnic groups.”

The committee claimed there are “clear incentives” for the policy, and “research estimates that addressing race inequality in the UK labour market could boost the UK economy by £24 billion a year”. Because companies already report their gender pay gaps, they are “already well resourced” to gather data on ethnicity.

There is no evidence that ethnicity pay gap reporting will contribute towards reducing the ethnicity pay gap, since it has never been undertaken, anywhere. We can however look at the impact of mandatory gender pay gap reporting to see if it works.

Since 2017 large employers have published data on differences in pay between the sexes, with the pay gap defined as the difference in average pay between women and men, as a percentage of men’s. There has been little change; the average pay gap hovers around 12 per cent. This hardly seems like “setting the ball rolling”.

Studies show mandatory diversity schemes tend to fail, since they antagonise people.

The most recent release from the Office of National Statistics (ONS) states “the ethnicity pay gap between white and ethnic minority employees has narrowed to its smallest level since 2012”. It stands at 2 per cent.

The same data suggest that “systemic racism” has little to do with it. They show Indian and Chinese groups earn more than the white British. Black Caribbean women earn more than white British women, while black African women earn less. There are no detrimental pay gaps for any ethnic minority groups, between the ages of 16 to 29.

Differences can often be attributed to other variables — for example the pay gap among British-born Bangladeshis compared to white British, drops from 21 per cent to 8 per cent after controlling for other demographic factors

This is not consistent with the idea of discrimination as the key explanation. If so, then there is no moral argument for government intervention.

As Nokes’ committee has observed, there are problems with ethnicity data. Ethnicity is not a binary category like gender. The ONS recognises nineteen ethnic groups, meaning a company in London will produce potentially eighteen pay gaps, all of which will likely tell no consistent story.

Mandatory ethnic pay gap reporting rests on the invincible fallacy

Some companies in places like Middlesborough will be returning statistics based on just a handful of non-white staff, leading to highly volatile estimates.

So, who benefits from this? This will be a costly bureaucratic exercise, running into the thousands, as well as fuelling unwarranted reputational damage and groundless litigation. It is illiberal in that employers have to prove themselves innocent, albeit with data that are not necessarily evidence.

Such data will provide HR, as well as the diversity and inclusion sector, with a powerful moral argument for expanding their budgets and control, without a clear idea of possible solutions.

The claim that eliminating the ethnicity pay gap, along with all other differences between groups, would enrich us by £24 billion, is pure pie in the sky. It stems from the MacGregor-Smith Review and is calculated by assuming that ethnic minority groups can be made to spread out across the economy in the same way as the ethnic majority, across all sectors and hierarchies. 

Immigration is an arduous process for many. Around half of our ethnic minority groups are immigrants, often recent. Immigration usually entails clustering in niche sectors before spreading out. There is no reason to expect equality of outcomes, nor does it follow there is a moral responsibility on government to engineer it, nor any evidence for its competence in such matters.

Many argue that statistical differences between groups are evidence of something called “racial inequality” and that it is incumbent on all to bring about “equality”.

On the other hand, you have Thomas Sowell who says statistical disparity between groups is the norm, always and everywhere, since groups differ in many ways. He writes of the “invincible fallacy”, namely the idea that differences are automatically taken as evidence of either one group oppressing the other, or the fault of “the system”. It requires no evidence and always returns, no matter how many times people show it to be wrong. 

Mandatory ethnic pay gap reporting is a bad idea, resting on the invincible fallacy. It is time to say “hasta la vista, baby” to it, but sadly, I suspect, it will be back.