This is the full version of the article I wrote last month for the International New York Times on the grammar schools debate in Britain. (I cannot publish my INYT articles on Pandaemonium until a month after it is published in the newspaper.) It was originally published under the headline ‘Why Britain Fails in Class’.

Few people could have predicted the first policy on which Britain’s new prime minister would take a stand. It is none of the issues that have dominated British politics in recent months: Brexit, immigration, terrorism. Rather, Theresa May has decided it is to be grammar schools – state-supported secondary schools that select their pupils through an exam taken at age 11 known as the 11-plus. Once a centerpiece of Britain’s education system, they have largely been phased out over recent decades in favour of a non-selective school system.

The history of grammar schools goes back to the Middle Ages, but the modern version emerged out of the 1944 Education Act, one of a series of laws that shaped social policy in postwar Britain. The act introduced free education up to the age of 15, and set up three kinds of schools: grammar, technical and secondary moderns. Few technical schools were ever built, so children came to be divided between grammar schools, which focused on academic studies, and whose pupils were destined for universities and better professions, and secondary moderns, which were intended for children deemed suitable only for less skilled jobs.

Over time, it became clear that what separated pupils in the two types of schools was not ability, but class. A vast majority of children in grammar schools were from middle-class backgrounds; a vast majority of poor children were condemned to secondary moderns. In 1965, the Labour government began replacing grammar schools and secondary moderns with a nonselective ‘comprehensive’ system, under which all children went to a single type of state school. A handful of Conservative local authorities in England have continued to maintain selection, but the numbers are small. Of more than 3000 secondary schools in England, just 163 are grammars.

The debate about grammar schools has become a proxy for a discussion of class. For many on the left, grammar schools institutionalize class inequality, shutting out the poor and catering to the wealthy. For many on the right, opposition to grammar schools is an expression of class envy and a misguided egalitarian plan to ‘level down’ rather than ‘level up’.

In recent years, however, even the Conservative Party has rowed back on its traditional support for grammar schools, as part of an attempt to ‘modernize’ itself and shed its old image of elitism. So why has Theresa May reignited the debate with a promise that any school can turn itself into a grammar school by selecting its pupils by ability? In her first address from the steps of No 10 Downing Street, as Britain’s new Prime Minister, she promised to tackle the problem of inequality. Supporters of the new grammar schools policy claim this is the first fruit of that promise. But the evidence that grammar schools hinder rather than enable social mobility is even stronger now than it was half a century ago.

In areas that retain grammar schools, fewer than 3 per cent of grammar school pupils are eligible for free school meals (a proxy measure of social deprivation), compared with 18 per cent in non-grammar schools. High-achieving children from poor backgrounds are less likely to be selected for grammar schools than those from prosperous areas with similar abilities. Over all, school results in Kent, one of the few English counties that retains a selective system, are similar to those in the rest of the country, but poor children do worse and rich children do better.

If there is evidence that grammar schools help entrench inequality, there is, however, little evidence that phasing them out has helped improve social mobility. A landmark 2005 study from the London School of Economics, which described social mobility in Britain as ‘low and falling’, showed that two children born, respectively, into poor and prosperous families in 1958 were more equal as adults than two similar children born in 1970. More recently, a 2010 report from the Organization for Economic Cooperation and Development found that on social mobility Britain came near the bottom of the class among rich nations.

The geography of social mobility has also changed. A government report published earlier this year observed that the key division now is between London, where even children from disadvantaged backgrounds do relatively well, and coastal and old industrial towns in the north and east of England, where most of what the report termed ‘social mobility coldspots’ are to be found.

This division maps onto a deepening political faultline. London overwhelmingly voted Remain in the recent referendum on membership in the European Union. Many of those ‘coldspots’ are in areas that voted to Leave; they’re also areas where, unlike London, the populist UK Independence Party enjoys significant support. Issues of social mobility and social disaffection have merged. Against this background, the new grammar schools policy seems less about improving education for the poor than about stemming social disaffection, the political consequences of which are becoming unsettling.

The fact that neither selective nor nonselective school systems have improved social mobility in Britain might suggest that the problem lies in the very idea of using schools to engineer a more equal society. A decent education system can help a few individuals progress beyond the circumstances of their birth, but it is unlikely to change fundamentally the social and economic structures that entrench inequality and restrain social mobility.

In focusing on social mobility, what has gone missing is the idea of education as a good in itself. One of the reasons people regard grammar schools with nostalgia is that they seem to represent a standard of good education. But they do so for only a few.

At the heart of selective schooling is the assumption that pupils from disadvantaged backgrounds are better off getting ‘vocational’ training rather than being intellectually challenged. The trouble is, that sentiment has persisted in the nonselective schools, too. The result is that Britain has ended up with a state system in which every child receives an equally mediocre education.

Expanding selective education will do little to improve the overall quality of British schools. The new grammar school policy addresses neither the problem of stalled social mobility nor the question of what good education means for all children.

The paintings are RS Lowry’s ‘Coming out of school’ and one of Cy Twombly’s ‘blackboard paintings’.