SWOB 3: Are Immigrants Taking All the New Jobs?

It is still all too often claimed that analysis of the ONS employment data indicates that immigrants are accounting for the majority of new employment, despite the ONS now issuing caveats on their employment numbers.

So where does this perception come from? The confusion seems to be based on the difference between net and gross job growth. The aggregate job change data often cited are net changes - the result of the difference between all jobs created  (or people hired) and those that break up in any time period . All those hirings and firings are the gross flows. There are many more hires in any year than the net change in employment. 

The LFS suggests there were around 4 million people hired in 2015. But there were almost as many people moving out of employment. So the net result of all these flows was that over the year,  net employment grew by 440,000.  

Net employment growth does indeed appear to be skewed toward migrants. For example the latest ONS Labour Market Statistics suggests that between 2014 & 2015 net employment grew by 120,000 for UK nationals and 320,000 for non-UK nationals. On this basis it would seem that immigrants account for around 3/4 of all job growth

Now look at the graph below. It shows the shares of immigrants in the working age population alongside the share of immigrants in all hires (defined here as individuals in work for < 3 months). According to the LFS - the same source as the net employment change numbers - less than 20% of all new jobs/hires were accounted for by immigrants in 2015, broadly in line with the share of immigrants in the population  -as it has been for the last 30 years.

So it is quite possible to have migrant hiring in line with their population share and at the same time  have migrants accounting for a majority of net employment change - if the population of migrants is rising faster rate than the population of UK-born individuals. The stylised example below shows how this could arise

In year 1 migrants account for 10% of the employed population (100/1000) and there is also a 10%  hiring rate ie every year the workforce hires 10% (and fires 10%) of the workforce. Migrants comprise 10% of all these hires  1/10 - proportional to the share in the population

Now suppose in year 2 the migrant population increases by 20. These new arrivals need food and clothing and so raise the overall level of demand. This generates additional vacancies and so there are an extra 1 worker hired  (from 10 to 11) and job loss among migrants drops to 1. The net change in employment is +10. Migrants now account for 11% of total employment  (20/110)  and 11% of hiring (11/101). The UK-born population stays the same. There are the same number of UK workers hired and lose jobs (90) every year. 

This pattern continues in subsequent years. In each year the number of migrants hired is proportional to the share of migrants in the employed population. By year 5, NET employment  has grown from 1000 to 1040 and all of this is accounted for by a rise in the number of migrants in work  (100 to 140)

BUT  the share of migrants hired in this period is 

11 + 12 + 13 + 14/(90+90+90+90+11 + 12 + 13 + 14 ) = 12.2%

ie the average migrant share in the population over this period.  

To repeat this is a stylised example, but the central message is that it would be wrong to conclude from analysis of the net change in employment that migrants take all new jobs. Rather the net change is mostly a reflection of the changing populations in the two groups, as it would be if any group (red-heads, ice cream eaters etc) grew relative to others.

Jonathan Wadsworth
Centre for Economic Performance
and Royal Holloway College