One of the urgent challenges of our time is to evolve a liberal tradition of anti-racism. At present the anti-racist movement has been captured by people whose policies tend to heighten the importance of race rather than to diminish it. In the liberal tradition, a person's race should be like any other first impression - not something a reasonable person would attach much weight to.

The Blair Government is in the process of copying US anti-discrimination laws just at a time when many Americans are having second thoughts. In particular, the implementation of the 2000 Race Relations (Amendment) Act will institutionalise awareness of race by building it into day-to-day decisions in the workplace and public services. The Government intends to impose duties on public bodies to promote the proportionate representation of racial groups and during this summer and autumn the Commission for Racial Equality (CRE) will devise codes of practice for most of the public sector which will come into force in November 2001. There will be a further 6 months for compliance until May 2002.

The Government's aim is to bring about proportionate representation. According to its February 2001 consultation document, the Government feels it is unsatisfactory in a 'modern, pluralist society' that we have no ethnic minority generals or High Court judges and that less than 2% of senior civil servants and only 3% of teachers are from ethnic minorities, when their representation in the general population is about 7%. It intends to overcome these statistical disparities by imposing racial quotas on the public services, from the police to the fire service through to the hospitals and schools.

The Government denies it is imposing quotas, claiming only that it is imposing race 'targets'. But how will a typical manager in the NHS feel when he is told that his target is to ensure that 7% of staff must be from ethnic minorities? How many managers will assume that nothing will happen if they fail?

These measures will be followed by the outlawing of 'indirect discrimination' under the EU race directive which must be implemented by July 2003. Under this legislation, statistical disparities may provide prima facie evidence of indirect discrimination. In such cases the burden of proof has been reversed so that employers have to prove that there was a good reason for any disadvantage experienced by an employee.

These public policies are based on the most naive of assumptions. Because ethnic minorities make up 7% of the total population it is assumed that they should comprise 7% of any sub-division of the population. This view relies on several unspoken beliefs.

The underlying assumption is that the qualities which make people suitable for an occupation, including teachers or generals, are equally distributed between ethnic groups so that large statistical disparities would not arise without discrimination.

It is further assumed that statistical disparities are measures of such discrimination. In practice it is also maintained that any disproportion which cannot be explained must be caused by discrimination. That is, discrimination has become the residual explanation for disparities which cannot be accounted for in any other way.

A further corollary is the belief that the economic circumstances of individuals are determined to a greater extent by discrimination than by personal choices or cultural preferences. Often, the influence of culture and choice are disregarded as explanations of outcomes.

Statistical disparities must be caused by discrimination

The assumption that statistical disparities must be caused by discrimination would be invalidated if such differences could be caused by other factors. Few of us, for instance, are surprised to learn that relatively few high court judges had unskilled labourers as fathers. Social class, with all that it implies about the ability and motivation of parents, the atmosphere of the home, and the parents' expectations, has been a powerful predictor of occupational outcomes. Statistical differences between groups can also be based on many other factors including age, gender and education, not to mention personal abilities. Discrimination may be the cause of disproportionate representation in some cases, but it is only one among several possible explanations.

Age: The Government wants ethnic minorities to be better represented in 'high-level' jobs, but people who have reached senior positions will tend to be older, whatever their ethnic group. What is the age breakdown of ethnic minority groups in Britain? According to the 1998 General Household Survey, 21% of whites were under 16, whereas 36% of ethnic minority groups were aged under 16. There are also significant variations between ethnic minority groups. Some 41% of Pakistanis were under 16, whereas only 30% of Indians were of that age.⁽¹⁾ In view of these differences, is it rational to assume that these ethnic groups should be expected to fulfil an equal share of their quota of generals and judges?

Newcomer Status: Some ethnic groups include newcomers and any person going to a foreign country may be at a disadvantage simply because everything is new. Their race may have nothing to do with their difficulties. About 51% of the members of ethnic minority groups were born in the UK, compared with 96% of the white population. Again ethnic groups vary: 13% of Indians aged 25 or more were born in the UK compared with 32% of people classified as black.⁽²⁾

Self-employment: Ethnic groups have different attitudes towards self-employment. According to the Labour Force Survey, just under 12% of the total population are self-employed, compared with about 24% of Pakistanis and Bangladeshis, 19% of the Chinese and about 9% of blacks.⁽³⁾ Groups who prefer self-employment will not be able to fulfil their quota of teachers and civil servants.

Command of English: Fluency in English is of particular relevance to the ability of individuals to hold down jobs in teaching or the civil service, perhaps handling phone calls from members of the public. According to the latest survey by the Policy Studies Institute (1994), of those classified as 'African-Asian' 91% had a good command of English, compared with only 75% of Bangladeshis and 76% of the Chinese.

To sum up: even taking into account only the most obvious categories it is plain that there are other explanations for statistical disparities between ethnic groups. Disproportionate representation alone is not evidence of discrimination.

Is discrimination the main cause of economic outcomes for individuals?

If discrimination cannot be taken to account for outcomes, how can they be explained? Do the lifestyles of members of ethnic minorities reflect their personal choices or their shared culture at all? Do personal qualities make any difference? To take a highly visible example, the success of British athletes during the last Olympics suggests that athletic prowess is not equally distributed across ethnic groups.

More important still, how important is culture? Consider attitudes to marriage and children. Decisions to marry or stay single and to have few or many children vary significantly from group to group and make a big difference to representation in the workplace.

Family breakdown: Family breakdown has a major impact on ability to get jobs in the public sector. Regardless of their ethnic group, it would not be easy for a lone parent to hold down a full-time job as a teacher or civil servant. Indeed, lone parents are only expected to work 16 hours a week to qualify for the working families tax credit. According to the 1994 Policy Studies Institute survey, 21% of white families with children were headed by a lone parent, whereas for those classified as Caribbean the proportion was 45% and 'South Asian' 8%.

Family structure: Family structure and the number of dependent children also make a huge difference to availability for work. According to the 1994 PSI survey, only 4% of white families had four or more children, compared with 7% of Caribbean families, 11% of Indian families and 42% of Bangladeshi families.⁽⁴⁾ Asians were also more likely to have complex family structures, one measure of which is the number of households containing three or more adults with or without children. Such household structures make a large difference to the ability and willingness of members to work outside the family home. Some 17% of white households were of this type compared with 49% of those from Pakistan or Bangladesh.⁽⁵⁾

Intensifying Ethnic Group Awareness

Social causation is complex and culture may have a bigger impact on outcomes than race. But there is a still deeper truth, namely that disparities will inevitably be discovered whenever researchers base comparisons (of any substantial sub-group of the population with the whole) on the sort of outward characteristics that can be measured by statistics. We are all highly complex in our individuality and statistical measures only get at part of this complexity. For this reason social scientists who use statistical measures have evolved complex methods of reducing the risk that their measures will be inaccurate. Race is a characteristic of all individuals, but how certain can we be that we have identified something significant about a particular individual when we define his or her race?

As the distinguished economist Amartya Sen has argued, we all have overlapping identities and our race may well be of little significance to us. He is, he says, from India but is also Master of Trinity College, Cambridge, an economist and someone who has lived most of his life in England. Why single out his skin colour as the most important thing about him? We could all point to overlapping identities: wife, mother, good neighbour, a considerate colleague, etc. If we are seeking to define the characteristics of individuals which explain outcomes, such as wealth or type of job, we cannot be sure that race has been the leading or even a major causal factor.

The racial monitoring and the preferential treatment which is supposed to follow will heighten each person's racial awareness. It will encourage people who have sought to live their lives without regard to race to see themselves primarily as members of an ethnic group. This is how preferential policies affected a very successful Puerto Rican businessman of my acquaintance who had reached the upper echelons of a major US corporation. Although he was born in Puerto Rico, he had left at an early age and thought of himself as an American and he could hardly call himself deprived. Then, when his daughter was 18, he found that she could not get into the college of her choice on merit, but that she could gain admission if she applied as a Puerto Rican and took advantage of the quota for ethnic minorities. What should he do? He decided that if the benefit of the racial preference had accrued to him personally he would not take advantage of it, but the benefit accrued to his daughter. Could he deny her a place in a better college? With a heavy heart he decided not.

Many Americans have been through similar thought processes. They had been living according to liberal ideals, trying to succeed through hard work and on their own merits, only to find that their approach was rather like turning down a pay rise. The rules of the game meant that it paid to play the race card.

Indirect discrimination and EU law

So far I have claimed that racial disparities may not tell us very much about the presence or absence of discrimination and that preferential policies increase the importance of race. Nevertheless, the Government is in the process of imposing new laws on the assumption that statistical disparities between ethnic groups are proof of discrimination, when no such conclusion can logically be drawn. Worse still, to enforce the offence of 'indirect discrimination' on such a dubious basis will be to punish people who have done no definable wrong and harmed no one. Yet, under the influence of the European Union this offence is to be extended.

Under the EU Race Directive, 'indirect discrimination' occurs when 'an apparently neutral provision, criterion or practice would put persons of a racial or ethnic origin at a particular disadvantage compared with other persons, unless that provision, criterion or practice is objectively justified by a legitimate aim and the means of achieving that aim are appropriate and necessary.'

It is recognised in EU law that there could be legitimate reasons for disparities, but the burden of proof is on employers to show that a 'particular disadvantage' experienced by a member of an ethnic minority was not the result of discrimination. To face a burden of proof is itself a penalty involving costs in time and money equivalent to a fine or an award of damages. It is justified by the assumption that any unexplained disadvantage must be due to discrimination, but there are many legitimate reasons for disparities which an employer might not be able to explain.

To take only the most obvious of examples, employers whose workforces are less than 7% from ethnic minorities. Most such employers will be unable to offer an explanation because they will be unaware of the reasons which led potential candidates not to apply to work for them. They will simply not know who they were, let alone their reasoning. Some potential employees may have disliked the industry (tobacco or meat processing) or the location of the workplace, perhaps on health grounds (near a motorway or overhead power lines). Such factors do not involve prejudice. They are simply beyond the control of the employer.

To impose penalties for indirect discrimination will result in punishing individuals who have done no direct harm to any other person. Normally any such use of state power would be considered unjust.

Liberal and Illiberal Anti-Racism

So far two claims have been made. First, that statistical disparities do not prove discrimination. Second, that the tendency of recent laws will be to heighten racial awareness. But why is heightened racial identity undesirable?

The vast majority of people in Britain want to treat ethnic minorities fairly and do so in their every day dealings. The liberal approach has been to emphasise equality before the law. Underlying this ideal is the concept of a society of free persons who accept constraints in order to allow opportunities for all. Liberals took it for granted that all are born with different dispositions and talents and that upbringing also provides advantages and disadvantages. The aim of liberals was to ensure fair play for all under the law. No groups should be barred from the opportunity to make the most of their life. During the 19th century the salient issue was religion. Liberals sought the removal of disabilities facing Catholics and non-conformists so that any person could be judged according to his or her personal qualities. They never demanded the proportionate representation of Catholics or Methodists because it seemed to them to be common sense that a society which wanted outlets for talent might not conform to a pre-ordained pattern.

Religion is an important guide to living, yet it was a fundamental tenet of liberalism that it should not automatically bar access to public office. Having discounted a major cultural influence, it followed with all the more force that visible first impressions should be of low significance compared with inner qualities. Individuals should be judged on their merits, not superficial physical characteristics.

How do the Government's proposals measure up to the liberal ideal? Will they lead to a reduction in the importance of race or an increase? Will it be more likely that individuals will be judged on their merits, or less? Unfortunately, campaigns for anti-racism have been captured by people who seek to heighten the importance of race and to enforce preferential policies for minorities at the expense of others.

Civic anti-racism v. Sectional anti-racism

There is more than one way to improve race relations and the one currently favoured by the Government and the CRE is the least compatible with a free and democratic society. Moreover, this is not an issue which can be understood as a reflection of the traditional conflict between the political Left and Right.

The current direction of race relations policy raises some genuine conundrums which are troubling to all shades of liberal thought, from libertarians to egalitarians. We now face a choice between two traditions of anti-racism: one may be called 'civic anti-racism', because it emphasises social solidarity as the basis of public policy; and the other 'sectional anti-racism', because it demands that public policy should serve the interests of designated groups at the expense of others.

The tradition of civic anti-racism emphasises equality. Its underlying ideal is that people of all races, classes and religions should be equal before the law and equally able to participate in making the laws which protect us all. No less important is the solidarity we feel as members of a society which ensures that all individuals are able to make the most of their talents.

However, civic anti-racists do not agree about the extent to which governments can legitimately impose equal outcomes. Classical liberals tend to draw the line at equality of opportunity, whilst egalitarian liberals in the tradition of John Rawls believe that the power of government can be used to transfer from rich to poor so long as fundamental civil and political rights are not undermined. This division of opinion derives from a dispute about the manner in which we are united as a people, with one group attaching greater weight to equality of material conditions than to equality of political and civil rights and their corresponding duties. Both classical liberals and egalitarian liberals, however, rest their view of the good society on each individual's sense of solidarity as a member of a wider, civil society.

The 'sectional anti-racists' do not see a society as made up of individuals who share in common institutions, evolved and maintained for the common good. They see people as, above all, members of groups defined by physical appearance. And they demand that individuals should be treated in a particular way because of their ethnic status.

This focus on group identity is inconsistent with the ideal of egalitarian liberals who emphasise the equality of all groups within the nation. Disputes about public policy towards the poor provide the clearest demonstration of the incompatibility between sectional anti-racism and egalitarian liberalism. In a sentence, if you want to help the poor - help the poor! Giving preferential treatment to ethnic groups will transfer advantages to both rich and poor within such groups. This problem is very relevant to the towns in the North West of England where there have recently been riots. If we assume that there is a case for public investment of some kind it is of the utmost importance that it should be directed towards people of all races who have encountered hardship because of the decline of the textile industry, not merely ethnic minorities.

Why Racial Preferences Do Not Work

Political pressure for governments to assist groups regardless of the needs of individual members is best understood as the natural consequence of the permanent presence in human affairs of self-interest. Liberals always recognised that the use of government power was necessary to maintain freedom but that such power could always be abused by self-serving individuals. Among the main dangers to be avoided was the capture of the state by powerful groups who would seek to disadvantage rivals. During the 17th century the chief danger was from religious groups and the 17th century saw the beginnings of truce between them, under which the power of the state would not be used to persecute religious groups (although Catholics and non-conformists remained under some residual disabilities). In the 18th century the United States built a similar safeguard into its constitution by outlawing all government financial support for religions. Also during the 18th century, liberals battled against other interest groups who sought to abuse the power of government to advance their own interests. Adam Smith's main targets were the commercial monopolies which continue to the present day to be on the lookout for protected status. British car manufacturers, for instance, have long enjoyed protection from Japanese imports to the detriment of car owners.

This longstanding liberal concern to avoid the abuse of government power continues to be relevant. In a democratic age, self-serving groups hoping to abuse state power need to present their case in favourable terms. Protecting British jobs has often been a useful strategy for winning taxpayer subsidies, not least because it has a measure of truth in it. Similarly, the latest ploy, seeking designation as a victim group deserving of preferential treatment, also has appeal because ethnic minorities contain people deserving of assistance. But, American experience is that the benefits of preferential treatment are often harvested by the already-privileged members of protected groups.

The declared intention of US anti-discrimination law was to have a behavioural effect, but the effect has not always been the one intended, chiefly because reformers neglected the reality that self-interest is a constant presence in the human condition.

For example, one of the mistakes made in America was to give preferential treatment to ethnically owned businesses as a strategy for alleviating poverty. Few consider it to have been a success. If anything, 'contract compliance' has benefited the already-successful members of minorities. Under the Small Business Act businesses owned by ethnic minorities were entitled to a proportion of government contracts. However, the American minority businessmen who were awarded these contracts were in no sense 'deprived'. They enjoyed a personal 'net worth' above that of the average American.⁽⁶⁾

Ethnic groups have leaders - perhaps we should say that they have people recognised by governments as the leaders. These individuals have their own agenda and tend to be pre-occupied with issues of concern to the already-privileged members of their group, such as university access or workplace promotions.

Studies of racism in the NHS provide a UK example. Over 30% of NHS hospital doctors are members of ethnic minorities, a substantial over-representation of ethnic minorities in a prestigious occupation. However, critics tend to focus on promotions, demanding that ethnic groups have a proportionate share of lucrative consultant posts. If the result of these criticisms is that a higher proportion of consultants are from ethnic minorities it will help the relatively small number of doctors directly concerned, but it will not improve the conditions of the least well off members of ethnic groups.

To sum up: on American experience, preferential policies for ethnic groups have not diminished social inequality. The benefits tend to accrue to the better off.

Moreover, once the advantages of protected status are demonstrated then more groups tend to demand it. In the US it began with black Americans, then feminists demanded that women be given protected status, then homosexuals, and then disabled people. Under EU law there are now six protected categories: race, gender, age, disability, religion or belief, and sexual orientation. Furthermore there is a tendency for existing groups to expand. Disability has been especially open to expansion, with alcoholics and obese people claiming protected status in America.

As a result, the political process becomes less a discussion of which laws and policies will best serve the common good, and far more about wrangling over which groups should have legal protection or benefit from preferential treatment. A polity riven by factions encourages group polarisation and tends to increase violence.

Anti-discrimination Laws May Make Matters Worse

The presence of self-interest also means that employers do not necessarily react to anti-discrimination law as predicted. They may seek to minimise the impact of the laws on their business with the result that anti-discrimination laws decrease the employment of ethnic minorities, especially those with the least skill. The labour market for people with useful workplace skills tends to be relatively free of prejudice because employers want people who can do the job well. However, many jobs require few skills and it is often difficult to judge during interview whether a particular person will prove suitable. Consequently, it is rather important to an employer that he should be able to dismiss an unsuitable employee after a trial period. If there are two candidates, one black and one white, and it is not possible to decide which is the best candidate, an employer is likely to choose the one who will be easiest to dismiss if things do not work out. This is likely to reduce the openings for unskilled job seekers from ethnic minorities.

Does the theory fit the facts? ILO unemployment in Winter 2000/2001 was 4.7% for whites and 12% for all ethnic minorities. Unemployment, however, is not equal between ethnic groups. There is a higher level of unemployment among ethnic groups with the highest proportion of unskilled members. In 14% of white men of working age had no qualifications, compared with 39% of Bangladeshis, 27% of Pakistanis, 24% of Black Caribbeans, and 15% of Indians.⁽⁷⁾ Ethnic groups with the highest proportion of unqualified men experienced a higher rate of unemployment, but there was not so much unemployment among whites. Among Pakistanis and Bangladeshis ILO unemployment in Winter 2000/2001 was 18%, and among those classified as 'black' it was 15%, whereas for Indians it was 6%.⁽⁸⁾

Much of the unemployment among ethnic groups will be explained by the lack of qualifications alone, but the figures are also consistent with the theory that employers are more reluctant to take on unqualified members of ethnic minorities than to employ unqualified members of unprotected ethnic groups. If true, the outcome of anti-discrimination laws will have been to increase the reluctance to employ the least skilled members of ethnic minorities.

Conclusion

The following claims have been made:

• Statistical disproportions are not proof of discrimination.
• Racial quotas and preferential treatment will tend to heighten the importance of race in social relations, creating racial polarisation and making race relations worse.
• Anti-discrimination laws will also tend to benefit the already privileged members of ethnic groups, leaving the poor unassisted and perhaps even putting the unskilled members of protected groups at a greater disadvantage than members of unprotected groups.

What are the issues for public policy debate? Above all, no one should submit to the atmosphere of intellectual menace created by the race-relations lobby, especially in the wake of the Macpherson report. We need to reaffirm our national commitment to anti-racism, but we should not be misled by the race-relations activists into adopting illiberal policies based on group antagonism. Classical liberals and egalitarian liberals should stand shoulder to shoulder in favour of civic anti-racism, so that the life-chances of all members of society depend on a combination of our personal qualities and the mutual concern of fellow citizens, and not on our membership of politically designated groups.

1. Living in Britain (General Household Survey 1998) London: TSO, 2000, p. 37.

2. Living in Britain (General Household Survey 1998), p. 36.

3. Labour Market Trends, June 2001, p. 297.

4. Tariq Modood et al., Ethnic Minorities in Britain, London: Policy Studies Institute, 1997, p. 41.

5. Ethnic Minorities in Britain, p. 46.

6. Sowell, T., Preferential Policies, New York: Morrow, 1990, p. 158.

7. Labour Force Surveys, 1998-2000.

8. Labour Market Trends, June 2001, p. 297.