Garvan Walshe is a former National and International Security Policy Adviser to the Conservative Party. He runs TRD Policy.

The restaurant was awful. The food was terrible – and such small portions.

Theresa May’s conference speechwriter must have thought of Woody Allen’s joke when writing:

“Throughout our history, migrants have made a huge contribution to our country – and they will continue to in the future. Those with the skills we need, who want to come here and work hard, will find a welcome. But we will be able to reduce the numbers, as we promised.”

If migrants are so great, and we can pick the right ones, then why reduce the numbers?

If we discount mere xenophobia — people who just don’t like foreigners, coming over here; and brute racism, people who don’t like foreigners even over there — there are two main causes of the public’s feelings about immigration: numbers and legitimacy. “Numbers” means just that: the Malthusian fear of a small island becoming too crowded. But for many the issue has been legitimacy, the idea that our elected representatives have no way of setting policy, and that the people who did set at least half of it, the European Union, were not our elected representatives.

This may lie behind the markedly reduced hostility to immigration since the Brexit vote. In 2016, around 40 per cent of people thought immigration had a positive economic impact. This has risen to 50 per cent in the population as a whole, and even among Leave voters, doubling to around 20 per cent from a very low ten per cent.

If it was ever to make any sense at all, the Government’s net migration target was only about numbers. It could be met as easily by increasing the number of existing residents who leave, as by reducing the number of newcomers. It was devised on the hoof by a shadow minister (who it was appears to be a matter of some dispute; but we know he was white and bald) and survived eight years of government, without a system being put in place to count the numbers leaving accurately.

This change in opinion gives us the chance to get off this numbers hook.

Though opinion has moved in a more positive direction, it is still highly polarised. Is there a way to allay the fears of communities who worry about a further influx, while ensuring that areas that need and want immigration can get it?

One method, being worked out by the Migration Matters Trust, could be to devolve work permits to city regions and national administrations in Cardiff and Edinburgh, as is already done to a certain extent in Canada. That wouldn’t mean Essex posting border guards on M25 off-ramps to keep more free-wheeling London’s immigrants out. Just as most of a bank’s work isn’t conducted by the tellers at the counter, most immigration control happens “behind the border”, by controlling access to welfare, requiring employers to only employ people with permission to work, and so on.

It has a number of advantages over the alternatives. First, it involves less central planning than the current system – and fewer creative attempts to evade target for tens of thousands of people staying more than one year, by proposing, for instance, 11-month seasonal worker schemes, and further expanding the list of “shortage occupations”. By involving regional authorities, it brings decision making closer to the people, and encourages them to have a debate about the quantity of immigration their community needs. Third, it is within the Government’s capacity to administer. Work Permits UK, which issued permits with high standards of efficiency, was only abolished just over a decade ago. Institutional memory of its effective system has been preserved throughout government.

Finally, basing a new immigration system on a salary threshold to distinguish between so-called “high-skilled” and “low-skilled” immigration, as hinted at during the Party conference, raises more questions than it answers. Even if the threshold is kept at £35,000, rather than raised (suggestions that it might be raised to £50,000 appear to be fake news) a level that seems reasonable for London will be difficult for Manchester or Glasgow, where 80 per cent of jobs pay below that figure. And though a test based on salary appears simpler than a detailed list of special occupations, it ignores the important matter of experience. Few young enthusiastic people earn £35,000 straight out of school, let alone university. Regional devolution would allow the cities where young people want to live to attract the people they need.

Failing to meet an immigration target based on an arbitrary number destroyed confidence in the immigration system. Replacing it with central planning takes back control from an open European labour market only to give it to faceless bureaucrats in Whitehall. Fresh thinking is needed, and devolving work permits to city regions would just be the way to obtain consent for an immigration policy that meets Britain’s needs.