The new migration system isn't working

New data from Freedom of Information requests also shows the power of selectivity to affect migrants' earnings

Before the election I made a number of Freedom of Information requests which I hadn’t had time to write up yet.

In previous posts I have explored new(ish) data from HMRC Real Time Information, which looks at the number of employments, and the distribution of earnings for people from different countries in the UK.

I got some more up to date data out of HMRC, and also some new cuts of the data - enabling us to look at the effects of both age and the public sector / private sector split. You can download the new FOI data below:

Employments And Earnings By Age Nationality And Percentile

695KB ∙ XLSX file

Download

Download

Employments And Earnings By Nationality And Public Private Split

161KB ∙ XLSX file

Download

Download

Employments

As regular readers will know, we don’t have a breakdown of recent net migration by country of origin (only grouped into EU vs non-EU). We do have data on the number of visas issued by country, but that doesn’t tell us the net flow.¹

So the data on employments by nationality² gives us a further (imperfect) way to understand migration trends.

It is far from perfect, as it only covers employment (not those who aren’t working) and doesn’t cover self-employment. The data by country of nationality also measures the number of employments not the number of employees (people can have more than one employment³). People who arrived in the UK under age 16 will count in the UK total. And obviously it doesn’t measure the grey economy and people not visible to HMRC.

Still, the employments data shows us the extraordinary changes since the introduction of the new migration system in January 2021: a system which is more restrictive towards EU migration, but much less restrictive towards migration from the rest of the world.

This is the effect of the system itself, plus supplements to it like the post-study work visa and social care visa, which have reduced selectivity. The less selective routes available to adult dependents of people on work and study main visas have also been much more heavily used in recent years. The effect has been to take net migration to record highs overall, while making it less selective for non-EU nationals.

This piece uses the FOI data above to look in more detail at the composition of that new wave of migration.

Although the new migration system kicked in in January 2021, the effects of the Covid-19 pandemic meant employment was temporarily depressed at that point, so comparisons with the pre-pandemic period are probably more sensible. I’ve used December 2019 to December 2023 comparisons here.

Compared to pre-pandemic there are:

• 1.481 million more employments

• 1.465 million more accounted for by people from outside the EU

• 257,000 more employments for UK nationals…

• …offset by 242,000 fewer employments for EU nationals

Within that non-EU total, the biggest growth in employments in absolute terms were among nationals of India (+488,000), Nigeria (+279,000), Pakistan (+101,000), and Ghana (+55,000). There were also 175,600 more employments held by people from unspecified other non-EU countries.

Obviously, there is no lump of labour, and whatever the other pros and cons of migration, there is no fixed number of jobs that migrants are competing for. But it is pretty striking that the UK economy has created more additional employments for nationals of both India and Nigeria as single countries than for UK nationals over this period.

If we were to look only at private sector jobs the contrast would be even greater - the number of private sector employments for UK nationals was marginally down (29,000) over the period, so all of the growth in private sector employments over this period was coming from a 1.2 million increase in employments for non-EU nationals. In the private sector 20% of employments were held by non-UK nationals in December 2023, somewhat higher than in the public sector.

Looking at the proportional changes, there were about 10% fewer employments for EU nationals working in the UK, and 69% more people from outside the UK than pre-pandemic. Compared to pre-pandemic there are three times more employments held by Nigerian nationals and more than twice as many held by Indian nationals, but 15% fewer held by Polish people and 22% fewer held by Czechs.⁴

You could have a go at working back from net changes in employments to rough net migration numbers (for adults) - but you would have to assume that employment rates by nationality haven’t changed since the 2021 census (of England and Wales)

But employment rates have changed, and probably not uniformly by nationality, so we will probably just have to wait for ONS to come up with figures for population by nationality.

The biggest proportional growth in employments has been from countries which have a mix of mid-table employment rates (India, Nigeria) and low employment rates (Pakistan, Bangladesh) while the declines have been concentrated among countries with high employment rates (Poland, Lithuania).

Earnings

I think the most interesting thing in the new data is what it shows happening to the average earnings of the nationalities that have seen either more liberal or tighter rules.

For India and Nigeria, the two largest growth groups, the average earnings of Indian and Nigerian nationals relative to UK nationals has sharply declined since the new and less selective system was introduced in 2021.

And these numbers are comparing non-UK nationals to an UK average which includes many teenagers and pensioners, who have much lower earnings.

If we look at the earnings of people of the same age the effect is even more striking: in fact, we see a cash decline in the median earnings of Indian and Nigerian nationals aged 22 to 40 - meaning a larger decline in real terms.

A less selective approach has massively dragged down the average, from young working age people from India and Nigeria earning 15 and 10% more than UK nationals of the same age before the pandemic, to earning less now. The two charts below show the same data twice, first in cash terms and then as an index:

Given that a large proportion of the higher-earning Indian and Nigerian nationals who were here before the new system was introduced will still be here⁵, this implies that the new arrivals since the new system was introduced have significantly lower earnings, and are dragging the average down. To have moved the average from above to below the UK median the wages of the new arrivals must be below the new averages for those groups⁶.

This is not to say that they are bad people, or doing a bad thing: indeed by definition all of the people in this data are working.

But as I have argued before, migration has fixed costs for existing residents in terms of diluting our stock of capital, infrastructure and housing: stocks of capital which can only grow so fast, and have built up over long periods.

For migration to benefit existing residents economically, we would ideally see new migrants being mainly very high skill and high earning. But the new system seems to be taking us away from that, with newer migrants among the biggest growth groups earning relatively less than previous migrants from the same places.

It is still the case that there is a massive range of average earnings between people from different nationalities, with people from richer countries (USA, Australia, Canada, NZ and Western Europe) generally earning more than people from poorer countries.

And within each of these nationality groups there’s obviously a range of earnings.

If we want to make the UK the grammar school of the western world, we want to reduce the number of people who come and either don’t work, or work on low wages, and so are less likely to be net contributors to the public finances.

The last government promised that the new migration system would attract the ‘best and brightest’. But instead, we seem to be going in the wrong direction.

India and Nigeria on their own account for half of the net increase in employment of non-nationals, so the lower earnings for newer migrants under the new system matters for the total effect. Do we see the same for other countries?

The picture is mixed, but broadly speaking the trend is towards maintenence or improvement in the relative earnings of EU nationals (for whom the system has become more selective - in blue below) and relative reductions in the earnings of non-EU nationals (for whom it has become less selective - in red below).

To understand what the the numbers here mean - Austrian nationals went from 131.5% to 134.9% of the UK average, so +3.5, while Bulgarians went from 86.6% to 82.5%, so -4.1%.

The change has come between the end of 2021 and the end of 2023, and has been driven by private sector wages.

Looking at a weighted average of medians for the EU⁷ and non-EU, a private sector wage premium was building up for non-EU nationals which has disappeared under the new less selective system since the end of 2021.

The data shows pretty powerfully how a more or less selective system can influence the composition of migrants, to increase or lower their average earnings: Non-EU employments have grown sharply, while wages have relatively regressed, EU employments have fallen a bit, while their median earnings have relatively improved.

While there is still a wage premium for non-EU nationals in the public sector⁸ (compared to UK-nationals in the public sector), this doesn’t change the overall picture much - there is relative regression under the new system for the non-EU migrants who make up the bulk of arrivals under the new system.

It is interesting to look at this in the context of earlier work by the Migration Observatory, who had detected relative improvement between 2016 and 2021 - however, the more recent trends for 2022 and 2023 look less favourable. We will need to see if they continue.

There is a lot of variation of course. Among the nationalities that have seen the largest increases in the number of employments, some countries like Zimbabwe and Iran have also seen wages fall relatively backwards, while others like Pakistan and Ghana have not, and indeed Bangladeshi nationals have seen earnings rise towards the national average, albeit from a low base. But the overall effect is dominated by the relative decline in earnings among the largest groups like Indian and Nigerian nationals.

Conclusion

The new migration system has brought about huge changes. Despite the promises (and a reasonable expectation among voters) that it would be more selective, it doesn’t seem to be, and indeed recent data shows earnings of the fastest growing groups of non-UK nationals relatively decline.

It is obviously possible that people’s earnings will catch up over time - though there is also evidence against this proposition, including work by Prof. Brian Bell on downgrading, which found that workers who initially came in on lower wages in the noughties, “remain toward the bottom of the wage distribution as their time in the UK increases.”

Putting together the picture on employment and earnings shows the potential for a much more selective migration system. The census found people born overseas have lower employment rates on average than people born in the UK, in some cases much lower, while earnings data shows some of the same nationality groups with low employment rates also see low earnings if they are in work.

There are lots of countries where both sets of data are not available but, as the back-of-an-envelope chart below shows, in general countries with higher earnings are more likely to also have higher employment rates⁹.

Given the fixed costs in terms of capital, housing and infrastructure pressures, we need to select for migrants who are significantly higher-earning than existing residents in order to improve the net impact on the existing population.

Rebooting the system to shift the balance towards higher skilled / higher earning migrants and groups with higher employment rates (as I proposed in a CPS report) could provide a boost to the public finances and the economy more generally.

1. I think ONS are planning to publish some net migration data for the largest 10 countries of origin in November, and we may get statistics on the population by nationality some time in 2025.

2. Nationality is defined as that reported by individuals when they register for a National Insurance number through the adult National Insurance number registration process. This process records one nationality per individual at the point of registration and is not updated if an individual subsequently changes nationality or citizenship.

3. In December 2023 there were 32 million employments and 30.3 million employees.

4. I have taken Ukraine out of the analysis of % changes in employments as it messes up the scale.

5. For example, of non-EU nationals who arrived in the HMRC data in or before 2014 three quarters were still on payroll in 2022.

6. I am grateful to David Algonquin for pointing out this trend.

7. Excluding Luxembourg and Malta, for which data is lacking. I appreciate that a weighted average of medians is not ideal, and we would rather have a median for the whole, but it gives you a rough idea.

8. For most nationality groups public sector median pay was significantly higher than private in December 2023. The only nationalities where median private pay was higher were the USA, France and Australia.

9. This is a very rough back-of-envelope chart with lots of problems: as well as the different age ranges the Census employment data is for 2021, for England and Wales, and is based on country of birth. The RTI earnings data is for December 2023, the UK and is based on what nationality you had when you first got a NINO.