MigrationWatch: Redux

So here's how the MigrationWatch briefing paper 'The Impact of Immigration on GDP per head' is nothing more than anti-immigrant propaganda. Most of this will have appeared in previous posts, but I've been able to think things through a little better a bit since I'm not working things out as I go along. This one was supposed to be edited out a bit, but it didn't work out that way so you'll have to click the 'Read more!' link to, er, read more. Instead of going through the paper section by section this time, I'll be looking at it problem by problem, giving a comprehensive cut out and keep guide. If you can get the scissors to work on your monitor.

The entire premise of the paper

Of course the extra contribution by migrants to GDP per head of the entire country per week is not going to be incredibly high. MigrationWatch appear to have spent a lot of time addressing a strawman, since none of the reports it mentions actually claim that it would be. MW claim that this is because they have all 'missed' something, but it's far more likely that they haven't worked out the contribution to GDP per head because it would be bloody stupid to expect it to be high in the first place.

This doesn't mean that MW are right though. It just means that the contribution of almost any part of society would look tiny when you do what MW have done for migrants'. Remember, they first calculated how much more immigrants produce per year than the general population. They then took that difference and divided it by sixty million. They then took that figure and divided it again by fifty. Surprise surprise, it ended up being not a very big number. For it to have ended up any other way, these people would have had to have migrated from the planet Krypton!

How much would the extra contribution to GDP by firefighters be if you divided it by sixty million and then again by fifty? Shop assistants? People with green eyes? Oxford Professors of Demographics? Would there even be any extra contribution? If not, does that mean that the general population don't benefit from their presence, or are there in fact more things people can contribute to than GDP per head?

And let's not forget that one major pro-immigration argument is that immigrants disproportionately do low-paid work that nobody else wants. So why would anyone interpret that argument as saying they'd contribute vast amounts more to the economy than anyone else?

Claims that lots of studies show the migrant contribution to GDP is shockingly low

In fact, the paper only tries to calculate the total contribution of GDP per head in two of the five studies it examines. Of the other three, two look at how much more immigrants pay in taxes than claim in benefits, and one looks at the contribution of migrants from the new Eastern European countries. These things are different.

Using different measuring methods to make it look as though studies show similar findings when they don't

This is perhaps the most sneaky tactic in the whole paper. MW claim that official Government figures show the extra contribution to GDP by immigrants is 0.01%, and that another study, the 'National Institute Economic Review' actually shows a slightly negative contribution. The thing is, MW use different methods of measuring both immigration and GDP to make it look as though the two sets of figures are in rough agreement.

For the Government figures, MW use net migration to measure how much the population has risen as the result of immigration. This is the right way to do it, since it takes into account how many people have left the country in the same period. But with the National Institute figures, MW uses total migration, claiming that migration has increased the population by 2.249m between 1998 and 2005, when in fact population increased by just 1.349m as the result of immigration in this period. That would make extra migrant contribution average out at about 0.1% per year since 1998. Ten times higher than the Government figures.

It's also the Government's practice to use a different method to calculate GDP. Given that the Government's figure apparently came from the Government, it's reasonable to assume they'd use their own method in their calculations. But MW don't use this method with the National Institute figures. It does show in the footnotes that migrant contribution would have been 4.5% over eight years had they used this method. So using consistent methods to calculate both immigration and GDP would have shown an average positive contribution of about 0.27% per year. That'll be twenty seven times the Government's number then.

Using the completely wrong figure for the official Government estimate

There's a much bigger problem in the inconsistency between the Government and National Institute figures, and that's that the Government don't actually produce figures to show immigrant contribution to GDP at all. MW saw a figure in a Westminster Hall debate and assumed that it was the figure for how much immigrants contribute, when in fact it wasn't. It was just a ballpark estimate for how much had been contributed by migrants from new EU countries since May 2004. It didn't look at the total number of immigrants at all, but just one group.

This is not just a little mistake. The Government numbers were the ones MW were desparate to get in the press. They're the ones everyone remembers, and the ones that made the headlines. And they're wrong. Wrong wrong wrong. Without this figure, the whole house of cards comes tumbling down.

Making hasty assumptions about figures

Using the wrong Government figure is just the most glaring example of how MW assume their conclusion in their calculations. Here, they took a figure and assumed it was something it wasn't - even though it would have disagreed with the other number they used to support it by being between ten and tweny seven times lower - and tweaked the other number so it looked as though it agreed.

There are more though. Most come in looking at the last study, 'The Item Club Report'. Now, I haven't seen this study properly, so I can't comment on all of it, but there are some very hasty assumptions that MW make that cast some doubt on their findings.

MW assume that everyone on the Workers' Registration Scheme works no overtime, has never been promoted and have not moved to a higher paid job since they registered.

They also assume that fifty percent of new EU migrants will return home. There's no way of knowing how many have returned from the Accession Monitoring Report they'd have been looking at, so that's just a guess.

Then, they assume that everyone who returned will have no dependants, so that the total number of dependants will be exactly the same as it appears in the Accession Monitoring Report they use, even after half the adult migrants have returned to Easten Europe. This ignores a note that warns that there is likely to be double counting here, as some people registered as dependants are themselves registered for work. So MW are counting some people who shouldn't be counted.

Counting people as immigrants who are not actually immigrants

There are other people counted as immigrants who should not be. The most glaring example is in MW's dismissal of an IPPR study that shows that immigrants pay more tax than they recieve in benefits. To do this, MW say that children of one immigrant and one non-immigrant parent should be counted as half an immigrant. They don't tell us which half arrived from another country, which isn't helpful.

In fact, none of these children are immigrants. Not by the dictionary definition, since none actually came here from another country, and not by the legal definition, since the British Nationality Act 1981 states that anyone with one UK Citizen as a parent is a British Citizen by birth.

MW repeat this tactic to explain away the unexpectedly large number for contribution to GDP the National Institute figure shows when they adjust for the Government method of calculating GDP. We have to count the children born in the UK to immigrant parents as immigrants and we have to assume there are just enough of these children to add the right amount to the population to cancel out the contribution. There are two problems with this.

MW have used the wrong figure for the addition to the population. It might be possible to assume that 2.249m people could have 405,415 children in eight years, but using net migration would mean that we'd have to assume that 2.249m people had 1.3m children in the same period. Which is a bit less likely.

But more importantly, it's unlikely that any of these children are immigrants at all since they've not actually migrated anywhere, and according to the British Nationality Act 1981, only one of those parents need to have ben granted Citizenship or legally settled here for their children to be born a UK Citizen. It's highly unlikely that the Office of National Statistics count people not legally settled in the UK in their migration figures - but even if they did, do you really think there would be enough people in the figures who have not been granted Citizenship or leave to remain to have 1.3 million children between them?

Hiding contentious figures in footnotes

The number above, which shows how the National Institute numbers would look if Government calculations are used is buried in a footnote. This is not because MW feel that calculations using the Government's method on National Institute are only worth mentioning in passing, because there is another figure included in the main body of the report where MW do the same thing. The difference? That one shows a low number, while the one in the footnotes is a high number. Must be just a coincidence.

Attempting to discredit other studies with irrelevant distractions

In addressing the Item Club Report, MW spend a little bit of time seeming to destroy the Item Club's assertion that the largest number of Eastern European migrants are in the Business and Administration sector by pointing out how many earn low wages. The same report MW address to find out how much the migrants are paid includes a league table to show how many workers are registered in each sector. So which sector is at the top of the table? Business and Administration.

Moving the goalposts

MW have dismissed the findings of a number of studies for different reasons, so no study is good enough.

First is a Government study from 2002 that showed that immigrants paid more in tax that year than they claimed in benefits. This would be a reasonable thing to do, since one of the main anti-immigration arguments is that immigrants are a burden on the welfare system. Once it's pointed out that they're not, the study is deemed useless because the budget was in surplus in 2002, so everybody contributed more. But every year since at least 1998 has had a budget surplus, so every study of this kind can be dismissed out of hand.

Then, an IPPR study attempted to correct this by showing that immigrants contributed more than the general population. But this study is not good enough because it doesn't include children of mixed immigrant and non-immigrant parents are not counted as immigrants. That they're not immigrants isn't relevant, apparently.

Next we have the two studies that measure contribution to GDP and show that immigrants produce a bit more than the general population. Never mind that one set of figures is totally poo, and the other one is shown using dodgy calculation methods - a positive contribution is not good enough because it's not amazingly high after you divide the contribution by sixty million and then again by fifty. One of these would show an unusually high contribution if consistent measurement methods were used, but that's not good enough because it doesn't count UK-born children of immigrants as immigrants, even though they've not migrated anywhere and it's highly unlikely that they would be legally considered immigrants.

It's difficult to know what would be good enough for MigrationWatch. Actually, scratch that, it is possible. Nothing would be good enough because MW assume their conclusion and contort the figures until they support that conclusion. Even the figues mentioned by Professor Coleman wouldn't be good enough, since they've already made some similar ones look smaller by using a dodgy immigration measurement.

The most annoying thing is that the papers will uncritically report the figures without even checking and it's been impossible to get anyone anywhere to take notice of the fact that they're based on a totally nonsense figure. Time to start looking about for stupid PC gone mad stories. Hurrah!