By Paul Ashton, Neil MacKinnon and Patrick Minford
Revised: September, 2016
Noone is arguing, from Leave or anywhere else, that we should restrict skilled immigration: the Leave proposal of introducing a ‘green card’ or skill-based points system would deliberately leave overall skilled immigration untouched, while tightening up on uncontrolled unskilled immigration from the EU, treating it as we now do unskilled immigration from the non-EU. Thus the net benefits from skilled immigration would remain the same and it would be unskilled EU immigration that would fall.
Hence the key question about immigration is what are the costs of unskilled immigration, which would be the target of the green card system. This has been widely neglected in the referendum debate.
We show that under the UK’s welfare system an unskilled migrant worker on the minimum wage costs the taxpayer in the region of between £57 (for a single person) and £29000 a year (for a worker with a family).
Estimates on the data available suggest that with the current population of 3 million EU migrants the cost to the average UK worker of supporting EU unskilled migrants could be around £2 a week; each adult unskilled EU immigrant costs the UK taxpayer on average £3500 per annum. For workers in an area with a dense migrant population such as Leicester the cost to local taxpayers could rise to £6 a week. There is no mechanism for spreading the cost burden evenly across the country because welfare resources are mainly provided locally.
However the key point is that within the EU there is no upper limit to the inflow of unskilled migrant workers from the rest of the EU. This cost could escalate without limit therefore. Every million extra unskilled adult immigrants we admit in the future adds these costs over again.
Everyone understands the politics of this problem. However, Economists for Brexit believe that the economic losses involved have not yet been properly appreciated. The basic point is that modern developed economies with advanced and costly welfare states provide benefits well in excess of the low tax take from unskilled workers; indeed these workers may receive more in tax credits than they pay in taxes and thus contribute negatively to the welfare benefits in kind that the state provides to them.
The green card system proposed by the Leave campaign meets the skill needs of the economy while properly limiting the drain on the economy from unskilled immigration.
The economics of unskilled immigration
Immigration is viewed as primarily a political issue. But it is indeed a major economic issue. There are measurable effects of immigration that indicate on average it is probably good for the economy. This has been disputed by Rowthorn (2015); he argues there that when you take into account effects on the NHS, education, housing and congestion, it is a doubtful calculation and the net effects may well be negative. There is more work to be done on this issue.
But we should realise that it is not the relevant question. The key question is the effects of unskilled immigration. Noone is arguing, from Leave or anywhere else, that we should restrict skilled immigration: the Leave proposal of the green card would deliberately leave overall skilled immigration untouched, while tightening up on uncontrolled unskilled immigration from the EU, treating it as we now do unskilled immigration from the non-EU. Thus the net benefits from skilled immigration would remain the same and it would be unskilled EU immigration that would fall.
On unskilled immigration the analysis is rather clear: at low (e.g. minimum) wages these immigrants pay little tax and probably are net recipients from HMRC via tax credits. In addition, they get benefits in kind from the NHS, free education and social housing. The net balance of the benefits they bring is plainly negative.
The Table following shows calculations for two types of unskilled immigrant household:
Notice these two calculations show the net cost to UK finances which is the cost to existing UK citizens; this comes about because the worker brings in a ‘product’ assumed equal to his pay so that this is a net of zero to the rest of the economy (the employer gets the product and pays exactly for it). The net cost then depends on the cost to the public finances. As can be seen this is highly negative. *
However, inside the economy some other things happen: welfare services net of taxes are transferred from residents to the immigrants (our ‘net cost’), and resident workers in services have lower wages which is a transfer to their richer employing households. The last is not a cost to residents as a whole but is to the communities where immigrants locate and compete. The former makes the immigrants better off at the expense of the residents.
We now move on to calculating what in total the cost of the existing unskilled migrant stock might be; it is necessarily somewhat imprecise because we simply do not have detailed enough data on migrant wages, household size, and occupational mix for EU unskilled migrants. What we have are snapshots of the foreign-born generally, from the Census, the Population Survey and the Labour Force Survey. The total stock of EU migrants (including children) was 3 million at the start of 2015. According to the LFS around 40% of all migrants are in unskilled occupations, and we use this percentage for EU-born migrants, giving 1.2 million as the population of unskilled migrants and their dependents.
The following two charts show that the EU migrant population has a different age structure from the general UK population: there are in particular fewer children. The proportion of those under 25 is 24% against 31% for the UK population as a whole.
We now take the family structure of unskilled UK households (defined as those with household income below 60% of the average equivalised income) and adjust it to reflect this lower proportion of young people, reducing the proportion of children to the same 0.77 of the UK. The result can be seen in the following Table where we show the UK unskilled household shares and the adjusted shares for unskilled EU immigrants.
We now calculate as above the cost for each of the EU unskilled immigrant households the total cost to the taxpayer.
Finally, theory says that unskilled wages for other workers will be pushed down closer to the minimum wage; a recent Bank of England report (Nickell and Saleheen, 2015) also estimated that for every 10% rise in the immigrant proportion wages for semi/unskilled services would drop 2%. From the point of view of national welfare this is irrelevant, what matters is the net gain to the UK as above. From the point of view of the local population it is pretty important.
So what we have here is:
1) from the national economic viewpoint we may be paying about £3.5 billion to support unskilled EU immigrants, approximately £3500 p.a. per adult immigrant.
2) from the local populace viewpoint it is a proportionately bigger cost per resident- one that we are unwilling or unable to compensate them for
How big is the cost for the local community? It is much bigger proportionately than it is for the nation. This is because in practice the local community carries all the costs of the benefits in kind- health, education and housing- since no extra supply is made available by the economy as a whole to meet the extra local needs. This in short is what all the fuss is about on the topic of immigration. Local people are hosting large immigrant populations and get no compensation for the extra costs they are bearing.
Just to illustrate how big this difference is, consider first the cost to the existing population of unskilled immigrants, as a cost per head: this is £3.5 billion per annum/31 million employed population= £113 per annum per worker or £2 a week. As a percent of household disposable income per head (about £26000) this represents 0.4%.
Now consider how much it is per head of the working population where immigrants settle. Below is a Table of regional population density of immigrants. Regionally the share varies from a low of 1.6% in the NE to a high 12.5% in the SE. London has a very high share but it also will have the mass of the skilled EU immigrant population which will on balance be high net taxpayers; this no doubt accounts for the observed lack of general concern about immigration in London.
In the following Table we show both regional and some selected high local area densities.
Table of immigrant shares of local population:
Regions Share (%)
East Midlands 5.8
West Midlands 7.6
N Ireland 1.5
Local authority areas
Forest Heath 23.0
So now consider the burden per head of local population; it is the average burden per head (i.e. for the total population) times the local immigrant density relative to the average density. For the region with the densest immigrant population, say Leicester, this burden per head rises to £287 per annum or about £6 a week. This is without counting any effect on local wages which we have no reliable estimates for. But notice this is equivalent to around 1% of average UK household disposable income per head.
We are dealing here with a burden for the country as a whole but one that is distributed most unevenly around the country. Those bearing the biggest burden, for which they are in no way compensated, naturally feel put upon. In the immigration hotspot areas the costs borne are higher than the 1% or so addition to the national cost of living. These areas are areas of lower than average disposable income and so these percentages understate the actual fall in these local living standards.
In this short note we have reviewed what evidence there is about the costs of unskilled immigration from the EU: it seems that the average cost to existing UK residents of an extra unskilled adult immigrant is £3500 a year. For the country as a whole there is a cost of existing unskilled immigrants (roughly 1 million adults) equal to a 0.4% rise in the cost of living. For individual local authority areas this can rise to 1% on the average UK cost of living. For poor people in these areas with incomes well below this average the burden can be a much higher percentage. Every extra 1 million adult unskilled migrants adds these costs over again.
The key point is that unskilled migrant workers do create a cost to existing UK citizens- they do not ‘pay their way’. What this means is that future unskilled migrants will add to this burden to a similar extent and that it is therefore entirely rational for the existing UK population to limit their numbers on the basis of reasonable economic calculations such as are implied by a green card system. This is not a device that will reduce growth or unnaturally prevent our enrichment by limiting free market forces; on the contrary a free market solution would involve charging these unskilled immigrants for entry according to this excess cost that they levy on us.
For well-off members of the elite these costs may not be apparent, since they show up in declining health, education and housing standards for poor communities which are not easily measured. For those who directly experience these declines however their losses are directly observed; the fact that those in power do not appreciate them makes the situation still harder to bear.
This loss to UK residents is a pure economic loss; there is no reason why the taxpayer in general would wish to pay it, or more accurately would wish to compensate the communities that do pay it by bearing the direct welfare state costs. This is clear enough from the fact that there is no practical means, with UK budgets under pressure and difficulties of measuring such declines, by which these communities are likely to be compensated. The remedy has to lie in avoiding this cost by restricting unskilled immigration through a green card system, as is widely practised by developed countries with a welfare state, such as the US, Australia and New Zealand. As has been clearly shown this remedy is not on offer within the EU. It can only be activated by leaving the EU. This is why it was a major issue in the UK Brexit referendum.
*The calculation made here assumes that the rest of the economy remains fixed, and is a ‘partial equilibrium’ calculation. However, if we were to allow for the full effect on the economy (‘general equilibrium’), the calculation would look little different. We have no model sufficiently detailed to calculate exactly this total effect of adding this small group of immigrants to the population. But we know the general behaviour of the economy from immigration: wages will fall so that the extra labour supply is absorbed, total demand will rise by the extra value supplied, and any required extra capital will be supplied by savers at the marginal cost of capital and it will generate an equal marginal product. At the end of the process the extra output will be paid for by the extra cost in (immigrant) labour and (home) capital; if wages fall, this will create a transfer between groups of workers- basically richer workers will enjoy higher real wages, while poorer workers will enjoy lower real wages. This last is not a cost for the economy, as it is just a transfer within the economy, but it is for poorer workers. If there were no cost to the taxpayer, then existing UK residents would be indifferent to the extra unskilled workers. It is the taxpayer cost that represents the loss to them. This is what we focus on.
One possible general equilibrium scenario would be to think of unskilled immigrants as purely providing services such as gardening or chauffeuring. The incomers offer these services at a lower wage and so induce people who would not otherwise to hire a gardener or chauffeur. We wind up with them being hired at low wages, producing extra service output, and driving down the wages of competing resident service providers: the buyers have extra product but spend less on other output, the immigrants extra wages which they spend on the extra output released by the buyers, so that there is extra GDP and extra employment equal to the immigrants now hired. This leaves existing residents having the same GDP per capita, while the immigrants have now the extra wages=GDP per capita which just compensates them for their effort. These aspects alone leave both residents and immigrants as well off as before.
Sources and calculations:
Pay from the National Living Wage rate (outside London).
Income tax, NIC, and benefits for specific household types confirmed with government-approved calculators for 2016. N.B. housing costs and council tax based on poorer areas in the South East outside London.
Education costs per pupil from NAO (2015), adjusted downwards to account for slightly lower age distribution of children in our families from the average.
Average health costs per person derived from dividing The King’s Fund (2016) estimate of the total budget for the NHS in England for 2015–16 of £116.4 billion by the population of England.
Calculation of 40% unskilledbased on distribution across types of employment- Migration Observatory (MO) briefing note. ‘Characteristics and outcomes of migrants in the UK labour market’. Men: 37%; women 45%.
Calculations of share of migrant population in UK local areas from Migration Observatory map
Shares of foreign-born population by region: from MO briefing note ‘Migrants in the UK- an overview’
Migration Observatory website: www.migrationobservatory.ox.ac.uk
King’s Fund (2016), http://www.kingsfund.org.uk/projects/nhs-in-a-nutshell/nhs-budget (accessed 2 August 2016)
NAO (2015), "Funding for disadvantaged pupils", 30 June 2015 NAO
Nickell, Stephen and Saleheen, Jumana (2015) The impact of immigration on occupational wages: evidence from Britain, Bank of England Staff Working Paper No. 574
Rowthorn, Robert (2015) The Costs and Benefits of Large-scale Immigration: Exploring the economic and demographic consequences for the UK, Civitas, London, December 2015.