Migrant workers: Taking our jobs - or not?

File 745Three reports about the impact of immigration on the UK labour market, all apparently contradicting one another, have come out in less than a week.

The first of the reports – from MigrationWatch – was published on Monday 9th of January 2012, and argues that it is “more than a coincidence” that the increase in the employment of East European workers corresponds with increasing youth unemployment in the UK. The report was covered in some media as showing that East European immigration has been a cause of rising youth unemployment (Daily Express, Daily Mail).

The next report, published on the morning of Tuesday 10th of January 2012, by the National Institute for Economic and Social Research (NIESR), concluded that there is no evidence of any impact of immigration on unemployment in the UK. Again, this result was widely reported in the Tuesday papers (Daily Telegraph, Financial Times).

Then, later on Tuesday 10th, the government’s Migration Advisory Committee (MAC) released a report looking at a broad range of economic and social impacts of immigration in the UK. One of the key findings of the report was that non-EU immigration during 1995-2010 was “associated with” a reduction in the employment of UK-born workers, but that there was no evidence of EU-immigration having an impact on UK-born employment in that period. The MAC finding again received extensive media coverage (The Sun, Daily Mail, BBC) leaving many people understandably confused.

It is not surprising that the publication of three different reports on immigration and unemployment within two days has created space for some news outlets to “pick their own truth” (Daily Mail, Daily Star).

So, what should we make of these different reports, and what do they contribute to our understanding of the impacts of immigration on the labour market (see Migration Observatory briefing on the labour market effects of immigration)?

The MigrationWatch Report

The MigrationWatch report is based around a single chart, which shows changes in youth unemployment and the employment of East European workers during 2004-2011. Both have increased in the period under consideration but not always at the same time. On its own, the analysis is clearly insufficient to draw any conclusions about the impact of EU immigration on youth unemployment. The chart simply shows two variables increasing over time without any further analysis.

But does this undermine the MigrationWatch argument that there may be a causal link between immigration and youth unemployment? No – but the MigrationWatch report contains no analysis to address the question.

The NIESR and MAC reports

When assessing and comparing the NIESR and MAC research, it is important to keep in mind four key differences between the two reports.

First, the two studies are aiming to explain changes in different measures of (un)employment. The NIESR study estimates the impact of immigration on the claimant count rate (i.e. on the number of people applying for jobseekers’ allowance divided by the working age population), while the MAC report analyses impacts on the employment rate (i.e. on the share of employment in the working-age population)

Second, the two reports assess the impacts of immigration on different groups. The NIESR study analyses effects on the claimant count rate of all people in the UK (i.e. including migrants), whereas the MAC study analyses the impacts of immigration on the employment rate of the UK-born population. The impacts of immigration on UK-born and foreign-born residents in the UK may well be different.

Third, the studies use different data. The NIESR study uses National Insurance Number registrations, while the MAC research uses labour force survey data.

Fourth, the time periods under consideration are different: NIESR analyses 2002-2011, while MAC studies the period 1975-2010 including sub-periods (1975-1994 and 1995-2010). These differences are summarised in the table below.

 NIESRMAC
What is the study trying to explain?Claimant unemployment rateEmployment rate
Impacts on whom?All people in the UK (i.e. including migrants)UK-born people (i.e. excluding foreign-born persons)
What measure of immigration?National Insurance Number registrations of non-British citizensLabour Force Survey data on foreign-born persons
What time period?2002-20111975-2010, with separate analysis for 1975-1994 and 1995-2010 sub-periods

So, given that the two studies were looking at different data to answer different questions, different results do not necessarily imply a contradiction.

It is also clear that each report has its own assumptions and methodological limitations which are important to keep in mind when interpreting the results.

For example, while the national insurance numbers used by NIESR allow a more detailed geographical disaggregation than data from the Labour Force Survey, national insurance numbers capture only the inflow of new migrants and do not measure the total stock, i.e. number of migrants in a given area.

Furthermore, receiving a National Insurance Number is not proof that somebody stays in an area, or even in the UK, which means that if a migrant were to receive a NINo in London, and then move to Manchester for a new job then this set of data would not count them as a migrant in the place where they were working.

A limitation of the MAC analysis is that the conclusion about the ‘associated displacement’ of British workers – which generated lots of headlines – is based on an estimate which is sensitive to changes in the specification of the statistical model used.

The MAC makes clear that its conclusions point to a “tentative association” rather than definite proof of a causal link between non-EU immigration and a reduction in the employment of the UK-born.

Another limitation of the MAC research is that the calculation of the stock of migrants is based on data from the Labour Force Survey, which as a survey, is subject to measurement errors.

The examples above are just some of the limitations of these reports. All social science research necessarily involves a range of assumptions and limitations. The limitations of the MAC and NIESR studies do not mean that any of them are “wrong”.

But what they do mean is that none of the reports can provide definite proof for the presence or absence of a causal link between immigration and employment or unemployment in the UK.

The reality is that to assess conclusively the impact of immigration on the labour market we need to know the “counterfactual”, i.e. what would have happened to (un)employment and wages had immigration not taken place.

In practice, without a clear counterfactual world in which to evaluate the UK labour market without migrants, it is not possible to definitely assess the impact of immigration on labour markets. Research uses various different methods for constructing that counterfactual world, but the approaches are all imperfect and subject to methodological caveats.

While this creates a headache for news headline writers, and lots of space for politically motivated misuse of research results, the latest contributions to the research on the labour market effects of immigration in the UK strongly suggest that there can be no “absolute” answers to this question. Anybody who argues that it is “obvious” or “clear” that immigration does or does not create unemployment in the UK needs to think again.

Related material

The blogs below provide more discussion, and different interpretations, of the new research of the impact of immigration on unemployment in the UK: