Why there is no simple relationship between immigration and the rate of unemployment

Background to this post

This post arises from an ongoing Twitter conversation with a rather unpleasant tweeter who was trolling @DAaronovitch on the subject of immigrants causing unemployment and who seems to have shifted his attentions to me. I’m not sure this particular individual is amenable to rational persuasion – his tweets are sprinkled with insults – but the position he seems to hold – that jobs in a particular country are a bit like spaces in a car park – is common enough, and I think is worth challenging in the hope that some people may revise their opinions.

Full Employment

First of all a definition: "full employment" does not, as some people assume, mean that everyone has a job. This terminology is used to describe a state of affairs where the number of vacancies is on a par with the number of unemployed. Of course neither the number of unemployed nor the number of vacancies are things that can be measured particularly accurately, but it was pretty clear that during the Thatcher years - when we had about three million officially unemployed and about a hundred thousand official vacancies - that we did not have full employment. During the 1960s and again for a time in the 1990s we did.

So why do we still have any unemployment during periods of full employment?

Well there are essentially two forces at work here: "frictional" unemployment and "structural" unemployment.

Frictional unemployment occurs simply because the world of work is fluid and dynamic. Jobs are being destroyed and created all the time and people leave jobs and begin new ones. At any one moment in a country of sixty million or so, there are, at any one moment, certain to be significant numbers of people "between jobs" - even if everyone concerned is destined to begin a new job in the near future.

Structural unemployment is a more serious problem. This occurs when vacancies require skills that are in short supply or workers who are in short supply in the area of the country where they are needed. Thus, even during full employment, you may have (say) lots of unemployed bricklayers in the North alongside a shortage of hair stylists in the South.

What can we do about structural and frictional unemployment?

Well frictional unemployment is not necessarily such a bad thing. It provides a pool of workers who can be recruited quickly and contributes to a certain degree of fluidity and flexibility in the labour market, but, to the extent it is a problem, efficient dissemination of information is the key. Structural unemployment can be addressed - though not always successfully - by retraining and help in relocating. Such measures cannot, however, address the underlying problem during periods of less than full employment - a point that often seems lost on politicians and journalists:

Fallacy #1: The problem of unemployment can be solved by education or training

During the abovementioned Thatcher years I remember regularly shouting abuse at my wireless set as commentators repeatedly advanced the argument that, in a situation where there were about three million officially unemployed and a hundred thousand official vacancies, those with the best qualifications and the best interview skills were the most likely to get any jobs that were going (that bit’s true so far) therefore (this is where the reasoning went wrong) all we had to do was train everyone to be good at interviews and to raise educational standards and everyone would get a job. The fallacy behind this argument is so obvious that there does not seem to be any need to point it out, but I still hear versions of it repeated even to this day.

So why don't we always have full employment?

The key fact here is that employees, as well as being suppliers of labour, are consumers of goods and services. Ignoring the effect of imports and exports (or, alternatively, assuming that there is an equal balance of imports and exports) if the demand for goods and services from the population as a whole is roughly equal to the ability of the entire workforce to produce those goods and services we will have full employment; if the demand for goods and services from the population as a whole is greater than the ability of the entire workforce to produce those goods and services we will have an overall shortage of labour; and if the demand for goods and services from the population as a whole is lower than the ability of the entire workforce to produce those goods and services we will have net unemployment - ie unemployment that would still be there even if we could fill every vacancy.

So why don’t we always have enough demand for labour

This is the point where those of different political persuasions tend to part company. It could be argued (and indeed I should argue) that high unemployment was a deliberate policy for a while under Margaret Thatcher since this policy tipped the balance of power against working people and (especially) their trade unions. But let’s assume most governments of whatever political persuasion will usually seek to maximize employment. Left wingers tend to argue that the problem is that pay is too low. “Increase pay” they argue “and demand for labour will increase”. Right wingers tend to argue that the problem is that pay is too high. “Decrease pay, and thus the cost of labour,” they argue “and demand for labour will increase”.

I don’t wish to enter into that debate here, but given what has been said thus far, you can see where both sides are “coming from”.

But what about immigration?

We need to start using some figures now, but I’m going to make up some nice round numbers – it’s the principle I want to get across, not the real UK numbers. What I present here is a series of thought experiments:

Let’s assume there’s a country, not unlike the UK, with 60 million in habitants and a workforce (ie all the people in work plus all the people seeking work) of 20 million.


Immigration to a country with full employment and no unemployment and no vacancies

If we assume that all 20 million of the workforce are in paid employment, we can easily calculate that each inhabitant of this country generates an average demand (to produce the goods and services he/she consumes) of 20m/60m = ⅓ of a worker.

Or, to put it the other way around, each worker produces (on average) goods and services for 60m/20m = 3 consumers.

Now let’s assume that a million people extra people arrive from a country (not unlike, say Poland) and let us further assume (for the moment) that those people have similar profile of demand for goods and services as the “indigenous” population and a similar profile of working ability – ie a third of them are of working age and intend to seek work in the UK; the rest are children or other dependents.

What would happen?

The population of our imaginary country would increase to 61 million. These 61 million people would require the services of 20 and ⅓ million workers. The demand for labour would therefore increase to 20 and ⅓ million and ⅓ million extra jobs would be created to supply the new demand. This would soak up the ⅓ million immigrants who were looking for work and there would be once again full employment and 0% unemployment.

Old population:	60000000
New population:	61000000
Workforce required to supply new level of demand from increased population:	20333333
Old workforce:	20000000
New workforce:	20333333
Old total in work:	20000000
New total in work:	20333333
Old total unemployed:	0
New total unemployed:	0
Old unemployment rate:	0%
New unemployment rate:	0%

Immigration to a country with unemployment and no vacancies

So let’s now consider what happens if we don’t have 100% of the workforce employed. Let’s assume that 18 million of the workforce are in paid employment and 2 million (ie 10%) are unemployed.
Again we can calculate that each inhabitant of this country generates an average demand for 18m/60m = 0.3 of a worker or that each worker is produces (on average) goods and services for 60m/18m = 3 and ⅓ consumers.

Now again let’s assume that a million extra people arrive with a similar profile of demand for goods and services and jobs. The Population of our imaginary country will increase to 61 million. These 61 million people will require the services of 18.3 million workers.
Unfortunately, since 20333333 workers are now chasing 18300000 million jobs, there will not be quite enough jobs created this time round to soak up all the new immigrant workers. Unemployment will rise (in absolute terms) by 33333, the unemployment rate will, however, remain at 100x2033333/20333333 = 10% - just as it was before the immigrants arrived.

Old population:	60000000
New population:	61000000
Workforce required to supply new level of demand from increased population:	18300000
Old workforce:	20000000
New workforce:	20333333
Old total in work:	18000000
New total in work:	18300000
Old total unemployed:	2000000
New total unemployed:	2033333
Old unemployment rate:	10%
New unemployment rate:	10%

Immigration to a country with unemployment and no vacancies where the demand profile of the immigrants does not match the demand profile of the host country

A reasonable objection to the thought experiments presented so far would be to point out that the demand for goods and services among the new immigrants may be lower than that in the host community (the new immigrants may, for example, save more or send money back home) and the demand for jobs among the new immigrants may be higher than that in the host community (the new immigrants may, for example, have a significantly higher proportion of people of working age).

So what happens to the figures when we consider such factors?

Again for simplicity, let’s consider what happens if a typical immigrant consumes only half what a typical non-migrant consumes and is twice as likely to be looking for work (and all other assumptions remain the same as in the previous example).

Old population:	60000000
New population:	61000000
Workforce required to supply new level of demand from increased population:	18150000
Old workforce:	20000000
New workforce:	20666666
Old total in work:	18000000
New total in work:	18150000
Old total unemployed:	2000000
New total unemployed:	2516666
Old unemployment rate:	10%
New unemployment rate:	12%

So in this highly improbable scenario (in which the demand for goods and services halves in a population where twice as many are seeking jobs) there is a slight increase in the rate of unemployment of 2%. If you think this sounds high, then you need to consider that during periods of economic crisis, the rate of unemployment may increase 2 or even 3 fold. But we are also leaving a number of factors out of the picture:

Firstly (partially lifting my self-imposed ban on considering imports and exports) money sent abroad is likely to result in increased demand for imports from the county to which the money is sent and some of that demand is likely to be filled by the country that has received the immigrants.

Moreover, if such monies are held in host country savings accounts pending (or instead of) being repatriated, then the host country banks will have more money to lend to host country businesses and this again will lead to the economy growing more quickly and more jobs and more demand being created as a result.

One possible reason for lower demand for goods and services which has not been considered is the fact that immigrants tend to be paid less for their work and therefore have lower incomes. This fact is, however, balanced by the fact that such workers are cheaper to employ and jobs are more likely to be created for cheaper workers. (Someone with a chip in their bath enamel may call in a plumber to replace the bath if he or she can get a plumber to do this for a reasonable price. If the price of a plumber is prohibitively high, the potential customer may just touch up the blemish with gloss paint and leave his/her money earning interest in the bank.)

All these considerations are likely to mitigate (potentially to a considerable extent) the 2% rate increase suggested by this thought experiment. But a far more important factor that needs to be considered here is the fact (excluded thus far) that, whether we have full employment or high rates of unemployment, there are always significant numbers of vacancies in the economy – for the reasons described above.


Immigration to a country with unemployment and vacancies where the profile of the immigrants does not match the profile of the host country

So bringing in progressively more features into this imaginary example (and thereby making it progressively more realistic but also more complicated) let’s now assume that there are - despite the 2 million unemployed – 1 million vacancies. Let’s ignore frictional unemployment and assume that this is all due to structural unemployment – shortages of the right skills at the right prices in the right places.

Migrants tend to travel to other countries looking for work precisely because they think they have skills to offer, that there will be a call for, in those destination countries. They will also (almost by definition) be prepared to travel to places where labour is required. Given what has been said about structural unemployment we only need to assume that the new immigrants (with their highly improbable demand profile) take up 450000 of the one million existing vacancies which the host population have failed to fill, and the overall unemployment rate will remain unchanged.

Old population:	60000000
New population:	61000000
Workforce required to supply new level of demand from increased population:	18150000
Old workforce:	20000000
New workforce:	20666666
Old total in work:	18000000
New total in work:	18600000
Old total unemployed:	2000000
New total unemployed:	2066666
Old vacancies	1000000
New vacancies	550000
Old unemployment rate:	10%
New unemployment rate:	10%

Conclusions

Of course none of these thought experiments prove anything for certain. It is, for example, logically possible that immigrants always burn 80% of the money they earn, or always so demoralize the host community that everyone else suddenly halves their consumption rates.
There are also more plausible things we’ve left out of the thought experiments - such as birth and death rates amongst the host community and the new immigrants and the timescales over which immigrants arrive and leave and over which jobs are created and destroyed. There is the fact that changes in productivity may allow an economy to produce more (or indeed less) per worker in response to changes in demand - though, again perhaps counter intuitively, there is no simple relationship between productivity and unemployment either. We might also wish to explore a greater range of hypothetical figures.

We have seen that there are highly contrived and improbable sets of circumstances that could (in principle) lead to small increases in unemployment rates following immigration but, in general (as I think the above thought experiments convincingly demonstrate), there is no simple relationship between changes in population size and changes in the rate of unemployment and many people's intuitions here are highly misleading. After all, if there were such a relationship there would be a consistent correlation between population growth/decline and employment rates around the world. We could have never have had full employment during the baby boomer years here in the UK, and Germany would always have higher unemployment rates than (say) Italy.

Clearly the solution to world poverty is not for all the poor people to move to the rich countries. Clearly a population (whether arrived at through high birth-rates or through migration) can be too large for the resources (and ultimately the land mass) of a particular country to support. There are all kinds of issues to do with migration that people might have views on. (I certainly have views: I was an immigrant to another country and am married to an immigrant to this country.) But the notion that the influx of immigrants to a country inevitably causes concomitant rises in unemployment rates is economic illiteracy.