Winner of the New Statesman SPERI Prize in Political Economy 2016

Tuesday 12 December 2017

Immigration and Real Wages: Reality and Perceptions


Robert Skidelsky’s recent piece applies a textbook model to the question of immigration, which implies real wages do fall for some time. The story goes like this. There is sudden increase in labour supply because of a wave of immigration. Immigrants compete for jobs, which forces down real wages, encouraging firms to switch to more labour intensive forms of production (automatic car wash to hand car wash), which reduces labour productivity. However as firms respond to higher profitability and invest, everything including real wages and productivity goes back to where it was, the only difference being that the economy is now larger. So immigration leads to lower real wages for as long as it takes firms to invest.

Did the large increase in EU immigration that started in 2004 have this effect? There is no noticeable change in the rate of increase in nominal or real aggregate wages until after September 2007, which was when the Northern Rock crisis happened. Everything from 2008 onwards is dominated by the recession. But we have emerged from the recession with an economy with low productivity and low real wages, and as yet firms are not investing. Is this the model working, and has the second stage involving more investment somehow stalled for some reason? Do we need to cut immigration so we can have increasing real wages again?

There are many problems with this textbook model. First, there was a large increase in work related immigration from non-EU sources in the second half of the 1990s, but no associated decline in real wage growth. Second, the model treats immigration as an exogenous shock, by which I mean immigrants just turn up and start competing for jobs. If that was the case, we would expect to see a temporary increase in unemployment associated with immigration inflows. You do not see that in the late 1990s, or immediately after the start of EU migration in 2004. And third, econometric evidence suggests immigration has no noticeable impact on real wages (see also hereResults are similar in the US).

There is a simple explanation for all these difficulties: we are using the wrong model. An alternative model is one where firms find it advantageous to produce in the UK, but have problems getting workers, a problem that is solved by recruiting from overseas. It is very telling that under Theresa May the Home Office commissioned many studies designed to show how immigration was holding down real wages, and all were suppressed because they could find no evidence. In this production led model, there is no reason for real wages to fall. The additional employment, investment and output go hand in hand.

What about the experience of low real wages after the Global Financial Crisis (GFC)? The first point to make is that there is nothing about real wages that a combination of productivity growth, higher indirect taxes (in 2011) and exchange rate depreciations cannot explain. We have not seen a significant shift in the distribution of income towards profits. So any immigration effect would have to come through via lower productivity. As I explained here, you can tell a story where poor UK productivity performance since the GFC is the result of (a) a global productivity slowdown (b) poor UK investment in R&D and new technology (c) unexpected negative shocks: first the recession, then the lack of recovery (austerity) and finally Brexit.

I would say at the moment the evidence strongly suggests that real wages are not noticeably reduced as a result of immigration, but I remain open to seeing contrary evidence. We should also note the undisputed fact that immigration provides more resources to the public sector, so workers are better off via this route. But if this is the case, why on earth do the public believe the opposite so strongly?


To many, in all social classes, it is obvious that immigration reduces real wages. The logic appears watertight. If more people with skill X come into an area, that reduces what the existing workers with skill X can charge. For the self-employed plumber, it means more competition for a fixed number of plumbing jobs, so quotes for any job will have to fall. For workers being paid a wage, it means they can be more easily replaced if they or their union asks for a higher wages.

The same thing happens in reverse when people point to Brexit leading to shortages in labour in some sectors because immigrants are no longer coming. Are, they say, that will raise wages in that sector, which is bound to make the remaining workers in that sector better off.

Telling people that econometric evidence suggests that there is no evidence that immigration lowers real wages cuts little ice on its own, because the logic above seems so clear. There is also no point in telling them that immigrants will increase demand, because it is obvious that immigrant plumbers will do their own plumbing, or immigrant workers producing X will demand far less X than they produce.

The mistake people are making is right at the beginning of this thought experiment. By talking about real wages, people immediately think about themselves as workers. As a result, they imagine a world where there are more immigrants doing the kind of work they do. And, understandably, they keep everything else constant. But immigration controls are never about one group of workers, or the immigrants that have just appeared in the neighbourhood. They are about workers across the economy, influencing not just their nominal wage but the overall level of demand in the economy. 

How do we get over the failure to generalise? One way is to ask people to imagine the opposite of most people’s implicit thought experiment. Ask what would happen if there was more immigration in every trade except their own. They should realise that immigrants would require the services that this person provides, so the demand for their trade would increase, allowing them to increase their quotes and raise their standard of living.

Another mistake some people make is to assume that because many EU immigrants are from poorer East European countries, and immigration happens because of differences in standards of living, we have to get some kind of equalisation of living standards as a result of immigration. To see why this is the wrong way of thinking about it, imagine if workers were forced to move the other way, from rich to poor countries. Would this raise real wages in poor countries? Of course not: they would be paid the same as domestic workers producing the goods the poor country does.

A far better way of thinking about EU immigration is that it is allowing production in the UK to happen which would otherwise go abroad. Of course there may be a small number of cases where overseas workers are exploited by restricting their ability to leave their employment, but domestic laws and resources should exist to stop that happening.

One way proponents of less immigration play on these misleading perceptions is to talk about just controlling unskilled immigration. That is clever, because it diverts people away from thinking in terms of a labour shortage model. The best response to that is to suggest various occupations that are not classed as skilled, and ask people whether they really want less care workers or nurses and so on.

But what if people still feel that immigration must reduce wages in some way. Perhaps just the possibility of potential immigration restrains workers for asking for pay increases. The final mistake that people make so often is to assume that lower nominal wages means lower real wages. For that to happen prices have to be unresponsive to lower wages. And that in turn would mean the share of profits in national income would have steadily risen since we had large migrant inflows, which it has not. The main determinate of real wage growth in the UK has and always will be productivity growth.

Falling nominal wage growth would still have an important effect, however. It would keep inflation low, which would allow policymakers to let unemployment fall relative to previous trends while keeping inflation on target (it would allow the NAIRU to fall). And that has to be a good thing.


Although the evidence overwhelmingly suggests that immigration does not reduce real wages, it is fairly easy to see why people would think otherwise. That contrast is ripe for exploitation. How it was exploited, and how it is turned a previously outward looking, tolerant nation into one that tries to deport someone who have legally lived here for over 50 years, will be the subject of a second post.


  1. It's possible in sectors like say hairdressing wages come down and possibly among other self employed. In computing contract rates as many abuse tier 2 visas some say and BA does.

    Point being can models cope? Are contrary examples pure anecdotes? I certainly think people pushing any view here should acknowledge difficulty of modeling or making hard conclusions.

    Is the data even available? I know someone pointed out that vast increase in self employed had barely increased income by the sector. However was that merely because some shifted from benefits to getting odd tenner and benefits.

    Anyway enough of my ignorant rambling.

  2. replace "Too many" with "To many"

  3. Your alternative model only feels like half an alternative to this non-economist. The NHS can't move its production overseas, and neither can Pret-A-Manger or the local theatre. So lots of the immigration and pressure on wages is not obviously affected by those companies you mention that have the option of moving production overseas.

    1. Point taken, but the NHS is also an example of where the first model does not fit either. In that sense it is also an industry where labour is being drawn from overseas to fill jobs, rather than competing for existing jobs.

    2. The point people are trying to make though is that there is an disincentive in industries - including the NHS - to create NEW jobs if there is an infinitely elastic supply of such labour from overseas. Expansionary macro policy will also not help if there is such an infinite supply of labour: which is why much of Britain did not benefit from the Great Moderation.


  4. Have you tried talking to any employers about the well-educated and highly able job applicants who are ready and willing to work in the UK.
    Pop into a couple of Oxford restaurants and hotels and ask the staff how they came to be in England, and talk to the managers as well.
    Then ask the employers if they feel any need to increase their wage bill.

  5. If the British state paid a parent working 16 hours a week £3,000 (or more)for £500 worth of work - which it does - might that explain our low productivity? Gordon Brown's system of in-work benefits, which has been mightily difficult to reverse out of, is a massive subsidy to every "entrepreneur" whose business is unable to pay wages which would cover the costs. I can have you serve coffee in central London for £7.50 an hour even if your rewards are £50 an hour, and it's this wage...sorry, sorry, sorry...income which drew you half way across Europe.

    Might the £100bn we spend subsidising businesses which can't sell their output at a price which covers their costs perhaps explain the productivity (non) puzzle?

    1. The government are currently cutting in-work benefits in real terms. As yet no sign of rising real wages.

    2. In-work welfare is a powerful explanation for the UK's low productivity, just as massive non-wage labour costs and difficulties in getting rid of workers (so machines rather than proletarians) is a good explanation of higher French productivity. In-work welfare is "progressive", so we have a terrible collective blind-spot regarding its role in creating battalions of low-productivity workers.

  6. Hi Simon,
    I’m really enjoying the Blog.
    I'm a former student of yours who is now working at a micro-economic consultancy. I've been asked to look at 'Brexit risks' for a major UK supermarket. One of the things I've been looking at is the sourcing of fruit and veg from the UK.
    One of the risks I've been considering is if the UK doesn't retain access to seasonal labour from the EU what happens to the price of produce traditionally sourced from the UK where labour is a large proportion of total costs. So far I've been modelling this as a substantial increase in labour costs due to increased wages - consistent with what you have described as the textbook model of immigration.
    Are you saying that I should factor in that wages in general will be lower because of the loss of demand from these seasonal workers (and EU migrants in general) so wages in this sector don't substantially increase? I can see that would be a countervailing effect, but in a sector where labour is 90%+ from the EU it seems unlikely that an end to free movement would have no impact on wages.
    Separately, I can see the argument that free movement is allowing production in the UK to happen which would otherwise go abroad. I think in the medium term that is what the end of free movement would mean for British fruit and veg farming.
    Any thoughts would be very welcome!

    1. It sounds like you have it right. There may be a short term increase in wages just to harvest existing crops, but otherwise the model is like the one I describe where production is mobile and will go overseas. UK Fruit + Veg production will be replaced with less labour intensive crops.

    2. One thing I would like to know is how did these labour-intensive agricultural industries survive before we had large scale immigration (which started during the Blair years). Or are these new industries that came about with large EU immigration?


  7. Thanks, good thought experiments. "Ask what would happen if there was more immigration in every trade except their own. They should realise that immigrants would require the services that this person provides, so the demand for their trade would increase, allowing them to increase their quotes and raise their standard of living." I still have questions. I am an immigrant but tend to agree that immigrants are wage reducing on a net basis, for the locals.

    Most people agree immigrants add to GDP. But it is the question of distribution of that GDP. Indeed if an immigrant adds to sectors other than his own, a local doing the same job (and may have lost his job to the immigrant) too would have added to other sectors and had a similar multiplier ratio (if not greater as immigrants are known to save and remit money back home)? So the question is does the immigrant add soo much more than the local in terms of multiplier to compensate for that loss of job of the local. Now is the time you say, but locals do not necessarily lose jobs. I agree the job loss is not 1 for 1, but some number less than 1. In that because of the immigrant, there is a positive non-zero job creation in the economy. But is this extent of job creation enough to compensate for the loss of job of the local?

    Keep in mind that the native too would have created jobs via his employment. So it is the incremental contribution of the immigrant we are talking about, not absolute job creation or absolute multiplier.
    Many studies are focused on GDP per capita (or wages), and not GDP per capita (or wages) separated out for locals and immigrants. As in, it is insufficient to show real wages have increased during the same period when immigration was high. The important question that people are wondering is whether real wages *for locals* be higher more than it would if immigration weren't there - the counterfactual.

    I also do not buy the argument that there is not enough skill locally and therefore you hire foreigners. If immigration reduces and there is demand for a skill, then real wages have to rise to make it more attractive for locals to learn and get into those areas. And educational institutions and govt will pay attention to getting people skilled in those areas.

    I will go through the links you cited here. The ones I have seen as a layman show correlation ergo causation, or post hoc ergo propter hoc biases. A link to a good theoretical model focussed on modeling wages for locals and immigrants separately and quantify the difference in wages for locals with immigration vs without immigration would be useful. Would you know of such a paper? I don't expect empirical studies alone to satisfy my suspicion of the counterfactual as it is hard to answer counterfactuals using historical data.

  8. This misses a number of things.

    Training is expensive, so much better for business to bring in people already trained. So the comparison isn't necessarily shelf-stacker with immigration vs shelf-stacker without immigration, it also has to consider what happens to that shelf stacker with less immigration; perhaps they become a nurse through taking up the additional training places that have been created since the referendum, a clear early win for Brexit (and yes I know those places could have been created anyway, but Leavers like me will claim that one for Brexit).

    It isn't about bulk immigration vs no immigration. The immigration that really annoyed people was bringing in Eastern Europeans to work at minimum wage in large mail-order establishments in former mining areas. Does making the UK the best place in Europe for low-wage activity make the UK/ better off? I would contend that low-wage work does not pay for itself once social welfare costs and housing costs have been taken into account

    Finally there is housing. We have been importing a city the size of Newcastle just about every year for ten years. We haven't been building a city the size of Newcastle every year. So we now have a housing crisis. Did the studies take into account the increase in housing costs? If not then it is not a proper comparison.

    I content that Brexit was about more than just in or out of the EU. It was about stopping a politics that sees the working people of the UK as a problem and immigration and top-down managerialism as the solution, and replacing it with a politics that makes working people the solution to the political problems through being ambitious for working people to reach their potential.

  9. It seems obvious to me as a non-economist but with a few relatives here from Eastern Europe that all of them need somewhere to live from day 1, so until they have been here long enough (in aggregate) to build the hard infrastructure to support them (which depends on all sorts of things, including the availability of land and other natural resources), there will be greater competition for homes and pressure on prices (particularly, based on my experience) at the lower end of the market. Common sense? I'll look forward to an explanation of why I'm wrong.

  10. ... and another thing. The immigration analysis seems to underestimate the shear scale of what is happening. The ONS figures for 2015 are that over 25% of all babies born in the UK were born to mothers born overseas. So 1 in 4 midwives is for immigrants, and in 2020 1 in 4 primary teachers is for immigrants. We don't carry that kind of surplus, so we have to import the social services professionals to look after the children of immigrants.

    The European commission estimate an increase of 25% from 63 million to around 80 million between 2013 and 2050/60 largely through immigration. This isn't a bit of immigration, this is importing an entire medium-sized european nation. The notion that to make this country better we should import another country to live and work alongside us is bizarre. Why don't we just make this country better?

  11. I thought Skidelsky's piece was particularly repugnant because it effectively concluded that the net economic effect of migration might be negligible, then puts his thumb on the anti-immigration scale with a - completely unargued - endorsement of David Goodhart's ethnic communitarianism ("It seems to me that anyone who thinks about such matters is bound to agree with Goodhart").

  12. My line that Brexit is a movement without a leader and Trump is a leader without a movement seems to be holding up well.

    Just reading through the victory for the Democrats in Alabama makes me think these arguments are ripe for being remade in Britain against the forces of Farage, Bannon, and Nouveau Riche-Mogg.

  13. I'm struggling to reconcile your arguments with the classic story about why the UK was the site for the industrial revolution. The classic historical story is that high wages and labour shortages forced mechanisation. Investment in mechanisation leads to profits (Kalecki's profit equation) and productivity growth. By contrast, substituting low wage workers for machines (eg hand car wash in place of automatic car wash), both reduces overall economy wide profits and reduces productivity.

    1. I think that our blogger's argument needs to be read precisely, and that part of his argument is that the *average* of all wages was not impacted by mass immigration. His story is perfectly compatible with hand-car washers earning £4/h cash-in-and and university vice-chancellor compensation tripling in a decade or two.

      «substituting low wage workers for machines [ ... ]both reduces overall economy wide profits and reduces productivity.»

      Low wage workers may reduce overall per-person value-added, but "economy wide profits" can increase or stay the same if there are more workers and/or the share of value-added for profits increases, and productivity *per pound of wage* (rather than hour of work) can increase.

  14. The paragraph starting with "Too many...", me thinks should begin with "To many...".

    1. Methinks me thinks should be methinks.

  15. «the evidence overwhelmingly suggests that immigration does not reduce real wages»

    I always admire the precision with which words can be used to misunderstand issues "reduce", "aggregate", "real", "wages" and so many other precise ways to avoid the issue.

    The question is whether in the absence of mass immigration from poorer countries the standards of living of workers on median wages would be lower, the same or higher than today. Would baristas, cleaners and builders in London and the south-east live 4-8 to a room and barely able to afford the bus fare to their workplace?

    My guess is that our blogger knows very well that immigration can impact the distribution of wages more than the average wage level, or the distribution between wages and profits (cleverly omitting rents), and that wages are not total compensation including T&Cs and that living standards also depend on cost of housing. Indeed he makes a very good example that shows how that is likely:

    «Ask what would happen if there was more immigration in every trade except their own. They should realise that immigrants would require the services that this person provides, so the demand for their trade would increase, allowing them to increase their quotes and raise their standard of living.»

    Of course that assumes, as or blogger is intelligent enough to be aware, that the ratio is everybody else to one, as in the example. What if the opposite happens? Consider that case of Dr. Dentist and Mr. Builder, the only inhabitants of Boris Island, and the two cases: one more dentist immigrates, one more builder immigrates. Obviously in the second case Mr. Dentist can charge a lot more as demand for his work has doubled, and when he want to build a conservatory, he will receive much reduced quotes as both builders will reduce his custom. And what if The inhabitants are Mr. Dentist, Mr. Landlord, and Mr. Builder, and there are 3 homes and Mr. Dentist is an owner-occupier, Mr Landlord owns two homes, and Mr. Builders rents one from Mr. Landlord? How likely is is that among the immigrants there is a new Mr. Landlord carrying in his suitcase two new homes to Boris Island?

    Our blogger has naively made his argument based on the mistake that the distribution of occupations of immigrants is exactly the same as that of existing residents: so for every landlord, dentist, builder that is resident exactly the same number of landlords, dentists, builders immigrates.
    His argument is essentially that if an economy expands equally and perfectly symmetrically by 10% in every way, from number of houses to number of dentists to number of builders to supply of raw materials, then that will by itself not impact on the distribution of income or absolute wage levels.

    Our blogger is fond of cleverly constructed hypotheticals:

    «immigration happens because of differences in standards of living, we have to get some kind of equalisation of living standards as a result of immigration. To see why this is the wrong way of thinking about it, imagine if workers were forced to move the other way, from rich to poor countries. Would this raise real wages in poor countries? Of course not: they would be paid the same as domestic workers producing the goods the poor country does.»

    I note the sudden switch from "equalisation of living standards" to "raise real wages", and "paid the same as domestic workers" instead of "lower wages just like domestic workers". Also note the clever assumption that workers from the richer economy were of the same skill level of the existing residents, that is their higher income in the richer country was totally unjustified by higher productivity.

  16. Overall I think the cleverest sophistries in this post are:

    «An alternative model is one where firms find it advantageous to produce in the UK, but have problems getting workers, a problem that is solved by recruiting from overseas. ... A far better way of thinking about EU immigration is that it is allowing production in the UK to happen which would otherwise go abroad.»

    The key assumptions here are that in the early 2000s the success of 20 years of thatcherism had driven the UK into the initial stages of a catastrophic wage-price spiral, and from Liverpool to Scarborough desperate employers were headhunting each other's workers by offering ballooning wages; and this happened because the main problem of the global economy was an excess of demand, such that if production did not happen in the UK surely it would have happened offshore instead; just like Germany and northern Italy and France in the 1960s the UK economy has been thus able to absorb a large wave of immigration effortlessly in the 2000s thanks to a combination of optimal global conditions.

    My memory of the early 2000s is instead that they weren't quite like the "Trente Glorieuses": the collapse of the dotcom boom plus the WTO entry of China created a situation where global demand for finished products was shrinking and the relevant global supply of labour had dramatically expanded, and there was instead in the UK a large shortage of supply of housing (in the south) and jobs (in the north), not of resident labour overall.

    My memory is that at the time there were strong arguments that in the face of falling global demand and increased exports from China a large influx of cheaper workers and lower pension contributions for everybody was needed to make UK industry more internationally competitive, and that at the time there was huge property price inflation but not wage inflation.

    In the 2000s there was a confluence of pretty dramatic global economy changes: dotcom collapse, resulting colossal credit expansion, China WTO accession, eastern European EU accession, dramatic increase in non-EU immigration to the UK, enormous asset price bubble, resulting collapse in leverage-driven financial markets. Plus gigantic foreign wars... My impression is that it is difficult to separate out influences, and that our blogger's sophistries seem to me strongly biased.

    But overall my impression is that the disappointment of the "left behind" is real, but it is mostly due to the China WTO accession resulting in the transfer of large chunks of some first-world economies to coastal China, and only secondarily to a flood of much cheaper workers from outside the EU (and even less so for those from inside the EU). For an intriguing "detail" first look at the obvious knee in Chinese electricity consumption in the early 2000s, and trade balance, and the amazing increase in UK private debt:

    Compare with the 20% fall in electricity consumption in the UK north-east and north-west since 2005:

    And compare also with this graph of total elecrity consumption 1990-2015 for major EU countries:

    Which shows ever increasing consumption roughly in line with "GDP" (whatever "improved" number that is) and then a stall and sharp decline well before the 2008 crisis. What's more amazing is the sharp decline in electricity consumption in countries with huge immigration inflows like UK and Italy.

  17. I suspect there is a further cognitive problem that makes changing such perceptions objectively difficult. The story suggested in the post essentially says that the downward pressure on real wages from immigrant competition is compensated by equally strong or even stronger upward pressure from the overall rise of demand. The perceptual problem is that it seems objectively hard to generalize and see that upward part: the downward part of the story can well be supported by personal anecdotes, while the (strength of the) upward part only becomes evident upon examination of macro-data.

    Many people may have a personal story where their acquaintance was driven out of an occupation or forced to accept lower wages "because" of new availability of immigrant labor for the boss. (I am thinking about precisely such a story I heard - in Germany, where I live, not the UK, but I imagine it wouldn't be rare to hear similar stories in the UK.) It may even be that the presence of immigrant workers is not all there is to the actual decisions by the boss, and that the same worker may have lost even without any immigrants in the picture, but it is always attractive to humans to draw a causal conclusion.

    In contrast to that, the upward part of the story only works cumulatively over the economy, and can hardly produce similarly vivid anecdotes. As long as that continues, the perceptions in favor of the position that immigrants cut down real wages will be hard to shift.

    Should we despair? I think not, but making the argument is not easy: because of the vividness differential, even the people convinced in one conversation may very well return to their earlier perceptions driven by the vividness of anecdotes. But we know that it is in principle possible to convince large groups of people of the existence of phenomena without natural vivid examples: much of the economic right's propaganda for the people works on that level.

    So perhaps the focus ought to be on explaining over and over again how the generalizing part of the argument works, while being prepared that it would take a lot of time and effort before the broad public will internalize the vision outlined in the post.

  18. Simon, I think the important point from Skidelsky is the dangers of a Dependence on high immigration. As Rawthorn says it becomes a treadmill. These are similar arguments to those against a high dependence on foreign capital and foreign goods. It makes it difficult to establish new industries and skill and maintain workers if there is a constant source of supply from outside. You must be also careful in saying that unemployment/undemployment rates in Britain are low. They are high, especially among the young, and have been for a very long time. And in some parts of Britain it has been, again for a very long time, particularly bad.

    For sure Britain's fundamental problems that predate even Thatcher are not mainly due to high immigration. However an over-dependence on it - as well as liberal attitudes to capital including foreign capital (in order, for example to 'protect the City') have been contributing to a toxic economic and social problem.


  19. "A far better way of thinking about EU immigration is that it is allowing production in the UK to happen which would otherwise go abroad." How does that hypothesis apply to the higher education sector in the UK in which 29.7% of academic staff are non-UK (17.3% from EU).

    "The main determinate of real wage growth in the UK has and always will be productivity growth." But isn't the availability of labour an enormous disincentive to firms to innovate and therefore improve productivity? There were 17.462 million unemployed people in the EU as at April 2018. The minimum wage in the UK ranges from £3.70-£7.83 (people usually only quote the 25 and over rate, forgetting the much lower rates for younger workers).


Unfortunately because of spam with embedded links (which then flag up warnings about the whole site on some browsers), I have to personally moderate all comments. As a result, your comment may not appear for some time. In addition, I cannot publish comments with links to websites because it takes too much time to check whether these sites are legitimate.