Should public policy be colour-blind? Or must governments and public institutions take account of people’s ethnicity and culture in formulating policy? It is a debate that has been reignited by President Nicolas Sarkozy’s attempt to introduce ethnic monitoring in France.

Unlike in Britain, where public institutions routinely collect information about people’s ethnic origins, it is illegal in France to classify people in this fashion. The foundation stone of the secular French republic is that all citizens should be equal and free from distinctions of race or religion. But senior politicians have begun to recognise that France remains deeply disfigured by racism. To combat this, Sarkozy argues, it is necessary to collect ethnically based data. The British experience suggests that such policies often do more harm than good.

Two assumptions underlie the argument for ethnic monitoring: first, that ethnicity and culture are the most important labels we can place on people; and second, that there is a causal relationship between membership of such a group and disproportional outcomes between groups.

If Bangladeshis are overrepresented in poor housing or if African Caribbeans are underrepresented in higher education, it is viewed as a consequence of belonging to those particular groups. Neither assumption is valid. Minority groups are not homogenous but are as divided by issues of class, gender, age and so on, as the rest of the population. These factors often shape individuals’ lives far more than do race, ethnicity or culture.

Take, for instance, the question of educational attainment in Britain. We all know that Asians excel at school and that African Caribbeans perform worst. Except that they don’t. Pupils of Indian origin tend to do well, but the performance of Bangladeshis and Pakistanis is similar to that of African Caribbeans. Bottom of the class come white working-class boys.

Children of Bangladeshi and Pakistani origin used to be labelled “Asian”. Now they are more likely to be seen as “Muslim”. When they were Asians they were bracketed together with children of Indian origin, and the differences between the groups were largely ignored. Now that they are Muslims, the poor performance of Bangladeshis and Pakistanis has attracted attention, but is often put down to “Islamophobia”.

To improve the educational possibilities of all these groups — Bangladeshis, Pakistanis, African Caribbeans and working-class whites – we have to understand what they have in common – which derives less from their race than their class. Bangladeshis and Pakistanis in this country come largely from poor rural areas, and African Caribbeans are predominantly working class. A system obsessed by ethnic categories, however, is rarely able to make those kinds of connections.

Ethnic monitoring does not just produce misleading data. The process of classification often creates the very problems it is supposed to solve. Local authorities have used ethnic categories not just as a means of collecting data but also as a way of distributing political power — by promoting certain “community leaders” – and of disbursing public funds through ethnically based projects. Once the allocation of power, resources and opportunities become linked to membership of particular groups, then people inevitably begin to identify themselves in terms of those ethnicities, and only those ethnicities.

Take Bradford. The majority of Muslims in the city come from the Mirpur area of Pakistan. But few identified themselves as Muslims – until the local council began rolling out its multicultural policies in the 1980s. Declaring that every community had “an equal right to maintain its own identity, culture, language, religion and customs”, the local authority looked to the mosques to act as the voice of the Muslim community and funded social projects along faith and ethnic lines. The council thereby helped create a Muslim identity that had barely existed before. By 1990 the city’s Mirpuri community boasted 18 mosques. Fourteen of them had been built in the previous decade, in the wake of the council’s multicultural policy. A community that had worn its faith lightly now became defined almost entirely by that faith.

National government has pursued a similar policy. Rather than appealing to Muslims as British citizens, with a variety of views and beliefs, politicians of all hues prefer to see them as people whose primary loyalty is to their faith and who can be engaged only by other Muslims. Should we be surprised then if, as a consequence, many Muslims come to see themselves as semi-detached Britons? Last week the government published Contest 2, its new anti-terrorism strategy. But it has still not understood the extent to which its own multicultural policies have helped fan the flames of Islamic radicalism.

Citizenship has no meaning if different classes of citizens are treated differently, whether through multicultural policies or through racism.

Kenan Malik’s book From Fatwa to Jihad is published by Atlantic Books next month

Times Recommends

• Care in old age: we won’t pay, the State can’t
• Viewers’ choice: our talentless Everyman
• Memo to the poor: go forth and don’t multiply