Latest ONS migration stats are out. Some thoughts follow. Key headlines: EU imm continues to fall, a trend seen since the ref. But non-EU imm has risen, almost offsetting the fall, so overall imm isn't falling and remains nowhere near (daft) govt target ons.gov.uk/peoplepopulati…
EU imm is now estimated at 219k/yr down from 284k/yr at time of EU ref. Net EU imm has fallen even further because emigration of EU born citizens has risen. Net imm is now at +74k/yr down from +189k/yr at time of the EU referendum. 
In the initial post-ref period decline came in those looking for work. Over last yr or so, big drop is in those with definite jobs. Makes sense - there are lags in recruitment process so impact of EU cits deciding they don't want to take jobs in UK take time to filter thry
While recruitment of EU citizens into jobs has declined sharply, this has been somewhat offset by a sharp rise in recruitment of non-EU citizens into jobs - this is running at +70k/yr, up 20k/yr on time of EU ref and roughly double the typical 2010-13 level
The overall rise in non-EU imm is much larger than this though - it is now running at +248k/yr net, up 50k on time of the EU ref, and the highest level in at least a decade. Not a product of more students - their numbers are stable - so most likely labour market/family reunion.
A few points I'd make re: these figures. Firstly, they show that it is not really plausible to argue that the big positive shift in immigration attitudes is due to a drop in immigration overall. Total migration hasn't fallen much since EU ref, though its composition has changed.
It may be that the sharp fall in EU imm has had some effect - particularly as such migration is less selective by skill, and as a big part of the drop is in people coming to the UK to look for work. But this has been offset by rise in non-EU imm, is latter less influential?
Secondly, it is hard to square these figures with the current govt's continuing claims that delivering net migration in the "tens of thousands" is their policy goal. Any drop in imm since the EU ref is happening in EU imm, which they can't (yet) control.
The forms of immigration which *are* subject to government control are rising - yet the public doesn't seem bothered by such rises. This pattern is consistent with the hypothesis that what matters to voters is the existence of control policies, not the specific numbers admitted.
Thirdly, these figures are hard to square with the idea that voters have a strong "ethnic preference" in migration policy, preferring migrants who are more culturally or ethnically similar to them, or at least hard to square with such a preference offsetting other factors
If the ethnic composition of migration inflows mattered a lot to voters, then the substantial compositional shift from EU to non-EU imm since 2016 should have produced a negative shift in public attitudes and increased public concern about imm. Instead, the opposite has happened
My own view on that FWIW is that an ethnic preference does exist among at least a substantial minority of voters, but it is only one factor among several - and high skills/educational qualifications, in particular, matter more.
I have a paper coming out soon with @jon_mellon showing that this is a general pattern across most of Europe - generally a mild preference for more culturally/ethnically similar immigrants over more distant ones; but a v strong and consistent pref for skilled over unskilled.
@jon_mellon Compositional shift from EU to non-EU migration may also be a compositional shift in skills mix- with non-EU imms having higher skills on average, due to the stricter controls on non-EU imm. The fall in concern as such imm risesis consistent with skills mattering more to voters
Overall, comparing these ONS stats with public opinion trends suggests to me that what matters to voters on immigration policy is more complex and nuanced than some of the narratives about the EU referendum or about immigration politics sometimes suggest
Numbers matter, yes, but often the symbolism of control can matter more than the substance. Cultural threats matter, yes, but they are not a deal breaker for voters who are happy to (e.g.) recruit doctors and nurses into the NHS regardless of where they come from.
